Title of Invention

A SYSTEM FOR ENHANCING AUDIO SIGNALS AND A METHOD THEREOF

Abstract 1. A method for enhan-ing transmitted i".idio data, comprising: coding audio data into a digitally formatted signal; enhancing the digitally formatted signal by pre-emphasizing frequencies find dynamics expected \o be losi or distorted, resulting in an enhanced audio .signal; transmitting the enhanced audio signal to a client site; decoding data contained in tat enhanced audio signal after transmission to die client site, resulting in a decoded audio signal; and processing the decoded audio signal to recover frequencies and dynamics •preserved by pre-eraphasis of the frequencies and dynamics expected to be lost or 15 distorted,
Full Text

WO03/10J924 PCT/USOJ/KTSS
ACOUSTICAL VIRTUAL REALITY ENGINE AND ADVANCED TECHKIQUES FOR ENHANCING DELIVERED SOUND
5 CROSS REFERENCE TO RELATED .APPLICATIONS
This application claims priority to United Staes Provisional Application Serial Number 60/335,541, titled "Advanced Technique for Enhancing Delivered Sound/1 filed on 5 June 2002, and to United States Provisional Application Serial
Number , titled "Acoustical Virrual Reality Engine/1 filed on
10 20 May 2003 via United States Express Mail (Express Mail Label No. EV331S71310US).
TECHNICAL FIELD
The present application relates to advanced processing techniques for
15 enhancing delivered audio signals, such as music delivered over limited
bandwidth connections, and more specifically lo processing techniques for
creating a live performance feeling in a listener jistening to a digital sound
recording delivered ire in any source of digital informaion.
20 . BACKGROUND
The rapid spread of the Internet has brought with it a rush to develop newer and more effective means for using its communicative techniques, beyond mere text-based applications. Two new applications that have garnered interest. are audio and video broadcasting. Both of these applications have a common .
25 problem: their utility suffers when the- connection to the Internet is limited in bandwidth. Because of its greater demands on bandwidth, video broadcasting is particularly problematic for the bulk of the Internet end-users (i.e., clients) who use United bandwidth connections.
One common method of delivering audio, such as music, on the Internet is
30 the "downloading" of audio files to the client's computer. Digital audio files are also commonly copied and compressed into MPEG audio, or other formats., onto a compact disc (CD), personal player or a computer hard drive, where they may be listened to in a more favorable or portable listening environment, compared to streaming audio.

V/0 03/104924 ■PCT/US03/177S8
Another common form of Internet-delivered audio is streaming audio. "Su-earnin£!: refers 10 listening while downloading. Generally, the server has a very high bandwid;h ccnnectiort TO the Internet, relative to the client's connection. In tie use of streaming audio for music, an Internet host site (i.e., the "server") 5 provides live music concerts, disc-jockey selected music or archived music to the listening end user (i.e., the "client") via an Internet connection. But due to the typical limited bandwidth, connections of clients, streaming' or downloaded (compressed) music is far from an ideal listening experience, particularly for clients accustomed to CD quality music.
10 The degradation of the listening experience can be traced to two main
sources: the compromises made upon compressed signals to compensate for limiicd bandwidth transmission requirements or reduced file size needs for storage purposes, and poor listening environments of the client. With respect to the latter, Internet-downloading or downloaded music is frequently listened to on speakers
15 attached to the client's computer, and, generally, little attention is paid to providing a good listening environment where the computer is situated. While recent efforts have been directed tc ameliorate the limited channel bandwidth problem, the problem of the poor listening environment has yet to be satisfactorily resolved. Accordingly, it would be advantageous to provide for technological
20 solutions that enhance the environment in which a client will receive and listen to sound signals received over a limited bandwidth connection. Furthermore, il would be advantageous to provide a system that can compensate for the distortion that results from compressing audio files into a smaller file size.
Performed music is composed of an extremely complex dynamic sound
25 field. The constantly changing listening environment of audience members and musicians along with variances in timbre, meter and unpredictable live performance dynamics combine to create s unique and moving musical experience. A live sound field is created when instruments and voices, supported . by environmental acoustics, meet to form a time domain based acoustical event.
30 Each of these elements is in constant dynamic change. Room, modes and nodes vary with listener position; music dynamics change with the artists' moods; even 2 listener's head position varies the experience from moment to moment.
Various schemes have been used by olhe:.; to clarify voice and sclo instruments in digital recordings. The most commen method used in traditional
2

WO 03/10-W24 . rCT..'US03/177SS
enhancement techniques is the addition of harmonic distortion to tlie upper ■ frequency range of the sound wave ("exciier"), But artificially injecting distortion into a stereo sound field creates user fatigue ana discomfort over time. Enhancement processes based on 'exciter" type processing often require a bass 5 boost circuit to compensate for thinness created by over-emphasizing high frequency harmonics.
Another approach deployed in televisions and car stereos for clarity enhancement of a stereo waveform, is the addition of a time delay circuit in the low frequency range along with a time delay circuit in the mid frequency range,
iC where both delays are set to a fixed delay point relative 10 the high frequency range. The purpose of r.his circuit is not acoustical simulation, but speaker normalization and is meant to compensate for impedance in the speaker circuit causing frequency-dependant phase errors in an amplified and acoustically transduced sound wave In this design, the high frequency level is adjusted by a
15 VCA control voltage that is initially set by the user with an ^adjust to taste" level control and is concurrently dynamically adjusted radiometric ally after a calculation of the RMS summed values of the delayed mid- and low- frequency bands. Banded phase-shift techniques emphasize upper-frequency harmonics and add a high frequency "edge" to the harmonic frequencies cf the overall mix, but can
20 mask and reduce the listener's ability to discern the primary fundamental frequencies that give solo instruments and voices depth and fullness, rendering them hollow sounding and not believable. Another problem with this speaker correction method is that it is not useful with all types of transducers, but is only ■ useful with those transducers that exhibit the type of high- and mid- frequency
25 time delay errors that match the time correction circuits within this process.
Another approach used for clarity enhancement of a mix is the addition of a time delay circuit in the low frequency range set to a fonnulaic delay point relative to the high frequency range. Banded phase-shift techniques emphasize upper-frequency harmonics and add a high frequency "edge" to the overall mix,
30 but mask and reduce the listener's ability to discern the primary fundamental frequencies that give solo instruments and voices depth and fullness. The effect of phase-shift techniques, when combined with a compensating bass boast circuit, is the "loudness curve" effect: more bass and treble with de-emphasized solo instrument and voice fundamental frequencies.

WO 03/104924 PCT/XS03/177S8
Compressors and voltage controlled amplifiers (VCAs) have been applied
to more sophisticated versions of these high frequency boosting circuits to adjust
the amount of distortion or phase-shifted material applied to the original sound
wave based on detected signal RMS values.
5 ' While useful as special effects on individual tracks prior to summing ths
track-into a stereo mix, high frequency boost ("exciter") processes are too deleterious to the fundamental frequencies of solo instruments and voice, and to the overall balance of the stereo-sound field, to be used as a professional-quality stereo mastering tool. Additional compression or downsampling of the music
10 waveform can c^use very unpredictable negative effects when distortion or phase-shift signals are added prior to signal density reduction. Loudness curve schemes are effective at low listening levels, yet moderate or high iistening volumes cause the mix to sound harsh and edgy, leading to listener fatigue and dissatisfaction.
It is therefore desirable to provide signal processing methodology
15 technology that accurately creates a live performance feeling in a user listening to a digital recording or other source of digital information, without the undesirable side-effects produced by conventional practices.
SUMMARY OF THE DISCLOSURE
20 An improved audio signal processing method and system is disclosed in
this application. The disclosed method/system is used to enhance the quality of an
audio signal that is about to be compressed and/or has been compressed. The
system uses an array of adjustable digital signal processors (DSPs) that perform
. different functions on the audio signal feed. According to one embodiment, the
25 method/system can be used to "rip1' an audio signal before it is compressed to a smaller format. As described above, compression of the audio signal may be necessary in order to transmit the signal over a limited bandwidth network connection. Compression may also be necessary in order to store copies of an audio signal on media with limited storage space, such as diskettes, CD-ROMs,
30 flash memory, and magnetic drives. Another embodiment of the method/system is used to enhance audio signals after they are decompressed. For example, the method/system may be used with a client-based streaming media receiver to enhance the audio signal after ii is decompressed by a streaming receiver. According to another example, the method ;md system enhances ths audio signal
4

WO 03/1Q-1924 PCT/VS03/1-7SS
as it is read and decompressed from limited storage media. In a preferred embodiment, die disclosed method'system is used £.t both the compression 2nd decompression ends of the audio stream, tt is contemplated., however that the disclosed method/system can be used exclusively at either of the compression cr 5 decompression ends of me audio stream.
Qne application for an upstream (i.e., compression-end) embodiment of the method/system is a "ripping" program that processes the audio signal at speeds faster than real time. This "ripping" program is useful for enhancing an electronic audio file before i: is compressed and stored onto a storage device. Because the
10 "ripping" program operates at speeds faster than real time, the time required to compress the file is greatly reduced. The upstream embodiment of the method/system can also enhance an audio signal before it is transmitted over a limited bandwidth network, such as the Internet. According to this embodiment, the method'system compensates for the distortion that arises from compression
15 ; prior to transmission over the network. Yet another application is a downstream (i.e., decompression-end) embodiment of the disclosed method'system. The downstream embodiment can be used to enhance the audio signal as -it is read and decompressed from the storage media, The downstream embodiment can also be used to enhance z streaming audio signal as it is received by a receiver. Because
20 the disclosed method/system can operate at speed faster than real time, it can effectively enhance the decompressed audio signal with minimal time delay effects.
In accordance with the disclosure of this application, Adaptive Dynamics type processing creates a believable, live sound field that is true to an original
25 actual musical performance through the use of FSM (Flat Spectra Modeling) acoustical environment modeling techniques. The processing techniques described herein can be utilized for the playback of digital music recordings, sound effects, sound tracks, or any digital audio source file, whether the source is a "rear1 recording or machine-generated (e.g., computer game soundtrack or audio
30 effects). Live music emulates life: unpredictable, sparkling, dynamic and ever-changing. The Adaptive Dynamics type processes are a balanced and life-like approach to performance restoration for digital sound. When combined with the recording environment simulation technology described hsrein, the sound waveform is analyzed and modified in the lime and frequency domains
' 5

WO 03/304924 PCT/LS03/177SS
simultaneously, ihcr. an acoustical rendering is generated based on predictive modeling of live performances, When used with artificially generated or '■foJsy:i sound fields - such zs those found in movie sound tracks - or synthesized sound tracks such as those found in games, the use 01 this technology adds a new 5 dimension of realism never before realized.
The disclosed technology creates a believable acoustical virtual reality generated environment which adds both dynamic intensity and overall sonic realism and clarity to the entire waveform through the combination of broadband Adaptive Dynamics type processing and Flat Spectra Modeling. This can be
.10 accomplished through the implementation of a complete 32- and 64- bit virtual-reality acoustics engine, where dialog is articulated, spaces are created and manipulated, and the user has simple and complete control of voice and sound environrr.erit characteristics. Each instrument and voice is focused and clear: even the fundamental frequencies that are the primary basis of each musical note.
15 The Adaptive Dynamics type processing approach of the present invention does not add a harsh edge or merely center on harmonics. The present invention reactivates the clarity and "life" of the entire sound field. Definition and focus are maintained in all audio bands with no undue or unnatural harmonic emphasis in any one band.
20 The Adaptive Dynamics type processes and recording environment
simulation technology involves the cooperation of two core processes: a multiple . path processing of the sound waveform using several filtered bands, and an unfihered band, which are lined up in time; and wall and room simulator functionality. The sound waveform is analyzed and modified in the time and
25 frequency domains simultaneously, then an acoustical rendering is generated based on predictive modeling of live performances, by setting processing parameters in these core processes,
The Adaptive Dynamics type processing creates a time beat which is intended to emulate the unpredictable, dynamic, and ever-changing characteristics
30 of live sound. This is accomplished by the use of multiple filtered bands or sound paths, and an unfiltered band or sound path, which are aligned in time, but which differ in acoustic characteristics. These differences in acoustic characteristics are implemented in one disclosed embodiment by applying different compression parameters (e.g., auack, release, gain ratio and target level) for each of the
6

WO 03/104924 ' PCT/tSOJ/n-SS
multiple filtered b^ds and the uimltered band. For example, the compression applied to the tmfiltered band may be sat to provide a sound that simulates the way :n which sound is emanated from a stage where there is no surrounding environment, while the compression for a midrange band is set to simulate a 5 sound emanating from a more lively environment, such as a scoring stage. These differences cause a- time beat to be created between the sounds being output from these different sound paths, and thereby tend to create hi the listener a perception of a more lively or dynamic performance. This time beat preferably is created without the use of rime delays between the sound paths.
10 Another important feature of the disclosed embodiments is the use of wall
and/or room effects processing following the Adaptive Dynamics type processing to provide a "tail'"' to the sounds. The wall/room effects processing add early, mid mid late reflection components to the sound, and thereby create a virtual sbsll or set of surfaces around the performance. This shell or set of surfaces can be varied
15 according to the environment whjch is desired to be created.
The Adaptive Dynamics type processing when combined with the walls block (early reflections) combined with the room block (late reflections) sen's to simulate a random event like a musical performance coupled with a relatively static system (with some variance due to sound waves impinging on materials)
20 such as an acoustic environment. The combination of the unpredictable even! (through Adaptive Dynamics type processing) combined with the predictable environment (through wall and room reflections) is unique and provides a perception in the listener which analogous to a live music experience. Therefore, the disclosed technology accurately creates a live performance feeling in a ussr
25 listening to a digital music recording, movie or game sound track, or other source.
Another element that could also increase believabiliry in the process as a
proper simulator for a live event would be the addition of a mechanism (such as a
microphone and a speaker) for determining the characteristics of the user's
listening environment which would give the overall process information about
30 listening levels, impulse response of the listening space, and time and frequency information regarding the listening spate and transducers used by the listener. This information; although optional to the proper operation of the disclosed embodiments, could be used as a calibration of the system.
7

W'Q 03/10-1924 PCT/US03/177SS
BRIEF DESCRIPTION OF THE DRAWINGS
A more complete understanding of the present invention may be derived
by referring to the detailed description and claims when considered in conjunction
with the accompanying drawings.
5 FIG. 1 is a flow diagram of an advanced technique for enhancing
compressed audio dziz, in accordance with a preferred embodiment.
FIG. 2A is a block diagram illustrating enhancement processing occurring at a server-side of a network, in accordance with a preferred embodiment.
FIG. 2B is a block diagram illustrating the enhanced processing occurring 10 at a client-side of a network, in accordance with a preferred embodiment.
FIG. 3 is a block diagram illustrating the enhanced processing occurring at the client-side of the network, in accordance with another preferred embodiment.
FIG. 4 is a block diagram illustrating signal processing functions for
enhancing audio signals, in accordance with a preferred embodiment.
15 FIG. 5 is a block diagram illustrating signal processing functions
associated with client-side enhancement of limited bandwidth music, in accordance with a preferred embodiment.
FIG. 6 is a block diagram illustrating signal processing functions for
enhancing audio signals, in accordance with another preferred embodiment.
20 FIG. 7 is a block diagram illustrating signal processing functions for
enhancing audio signals, in accordance with another preferred embodiment
FIG. S is a block diagram illustrating signal processing functions for enhancing audio signals, in accordance with another preferred embodiment.
FIG. 9 is a block diagram illustrating signal processing functions 25 associated with client-side enhancement of limited bandwidth music, in accordance with a preferred embodiment.
FIG. 10 is a schematic representation of an example vocal enhancer element suitable for use with the architecture depicted in FIG. 1.
FIG. 11 is a schematic representation of an example spatial enhancer 30 element suitable for use with the architecture depicted in FIG. 10.
FIG. 12 is a schematic representation of an example Wall Effect element suitable for use with the architecture depicted in FIG. 10.
FIG. 13 is a schematic representation of an example Room Effect element suitable for use wuh the architecture depicted in FIG. 10.
3

WO 03/JO492-I PCT/lJ50o/l"7s8
FIG. 14 is a schematic representation of an example SubSonic Effect element suitable for use with the architecture depicted in FIG. 10.
FIG, 15 is a schematic representation of an example Look-Ahead AGC
element suitable for use "with the architecture depir.ed i:i FIG. 10.
5 FIG. 16A provides an illustrative example of one implementation of the
Adaptive Dynamics type processing block (labeled core process) in FIG. 10.
FIG. 15B is an illustration of the lime response characteristics of the sound paths of FIG. 16A.
10 DETAILED DESCRIPTION OF PREFERJRED EMBODIMENTS
Techniques for enhancing sound delivered to a user via a limited bandwidth transmission system, or from a compressed digital 51e, are disclosed herein. And more particularly, what is disclosed are techniques for client-side enhancement of sound files, which can be delivered as streams or as downloads
15 via the Internet or other means to client devices such as CD, portable players, set-top boxes and the like, and which can he played over a computer-based sound system having limited fidelity and in an environment with ambient noise or other poor acoustical attributes. Also disclosed are techniques for compressing an audio signal at speeds faster than real-time so that the audio signal can be broadcast over
20 a limited bandwidth connection. Other embodiments include client-based applications wherein an audio signal is enhanced after it is decompressed, such as a streaming media receiver or an electronic audio file player (i.e., an MP3 player). : Accordingly, the disclosed method/system can be used in the following ■ applications:
25 • a server-side "ripper" operating a speeds faster than real-time;
• a client-side enhancer device without the need for pre-ripped sound files;
• a broadcast server where audio signals are enhanced in real-time;
• a receiver client where audio signals are enhanced in real-lime;
30 ♦ a server-side ;tripper!"' where compressed files are decoded later at the
client side for further enhancement of quality and clarity; and
9

WO 03/10-J924 PCTA3SOJ/177S3
• a client-server arrangement where the audio signal is enhanced at the
seirer side prior to compression and further enhanced at the chent side
after decompression.
FIG. I is a Low diagram depicting an advanced technique for enhancing
5 audio data, in accordance with a preferred embodiment. At step 102, audio date is
coded in a digitally formatted signal. At this point the digital signal may also be
compressed in preparation for subsequent transmission, Once in a digital format.
z\ step 104, the coded audio signal can be enhanced by using various processing
techniques that emphasize frequencies and dynamics expected to be lost or
10 destroyed during subsequent transmission. Thereafter, at step 106. the enhanced
audio signal is transmitted over a connection, which may be of only low or
■ medium bandwidth, to a network, such as the internet. After reaching a client site,
at step ] OS, the transmitted audio signal is decoded (and also decompressed if
necessary). Finally, at step 110, the now decoded audio signal is subjected to
15 further enhancement processing to recover the frequencies and dynamics expected
to be lost or destroyed during transmission.
FIG. 2A shows the enhancement processing occurring at the server-side of a network (i.e., the Host Site), in accordance with a preferred embodiment. At the host site210r music is selected from a music source 202, such as., fox example, 20 stored files or a live feed. Interposed between the music source 202 and an audio codec 204 is an enhancement processing element 212. The enhancement processing element 212 enhances the audio signal prior to being coded by the transmitting audio codec 204. Enhancement processing is beneficial if the -Streaming server 206 is broadcasting to clients with known and/or similar listening 25 environments. • Also, it is beneficial when the type of music being broadcast is known or determined, or always of a similar type, because the enhancement processing can be adjusted in a way thai maximally bsnafus that particular kind of music.
The transmitting audio codep 204 processes music through an encoder 30 ' (i.e., the transmission half of a codec OUT) that formats and compresses the music in s manner that is adapted for the bandwidth of the client's Internet connection.
A codec is an encoder/decoder system, mat fcr discussion purposes herein, functions as an audio data-compressor (encoder) and an audio/data decompressor
10

WO 03/104924 PCT/US03/I77SS
(decoder). A data cojnpressing'deccmpressiiig codec is also known as a "compander." In this disclosure, "data compression" will refer to any process which reduces -he size of a dam file, while "sound-ievsl compression" will refer to any process which reduces the dynamic range of an audio signal, Some 5 commonly used codecs are Sony STrack, Dolby AC3. and \VMA (U?3).
After applying the transmitting audio codec 204, a streaming server 206 lb en transmits die cat a-compressed and formatted music data to the designated address over ouipu; connection 214 to the Internet. Although this description primarily refers to the streaming and enhancement of-.music, it equally applies to
10 any audio or audio/video material. Further, it should be noted that this system and technique can bs used with a variety of sound transfer protocols, including, for example. Real Audio, M?3, and Windows Media.
As used herein, tlreal-time" means that the listening client hears the music substantially at the same time as the server is processing it within the audio codec.
15 While there may be some delay resulting from the connections to the speakers to be considered "Teal time" it is preferable that there be no substantial buffering of any segment of the music between the music stream at the music source and the speakers where the client is listening, and sequential music segments follow at the speakers. Downloaded files may be stored in their entirety and played at a later
20 time and are preferably compressed in the same way as streaming files, although the compression ratio may be less than the ratio used for real-time streaming.
FIG. 2B shows the enhanced processing occurring at the client-side of a network (i.e., "decoder-side enhancement11) in accordance with a preferred • embodiment. This type of enhancement processing is beneficial in circumstances
25 where there is a wide variety of listening environments and/or music types. Through low or medium bandwidth connection 222, the enhanced, coded signal reaches the client site 230. Specifically, the signal 222 can be provided to a personal computer 244 or another suitable processing platform. In the preferred embodiment, the personal computer 244 includes a modem 242, a processor 2^4
30 associated with the receiving audio codec 246 and an enhancement processing element 252, speaker drivers 24S, and speakers 250. Like the enhancement processing element 212 provided at the server site 210, the enhancement processing elemen: 252 preferably provides for enhancement of a decoded signal, after it has bean decoded by the receiver audio codec 244.
11

"VYO 03/10-I92J PCT/HSW/j77S8
The processor of ths client's receiving codec 246, which is associated with the CPU 244, pirfcmis what is largely the inverse of the server's transniining audio codec 244. Specifically, the receiving codec 246" converts the data stream back to a readily-usable music format, and uncompresses the music ic restore it zs 5 closely as possible ;o its original quality at the music source 202. The receiving, audio codec 244 process may be running in software en the CPU 244, or may be performed in hardware by ihe use of an add-on sound card. Speaker drivers 4S can also be found cr. the sound card or implemented in software. Speakers 250 ir. a typical client's listening environment consist of a pair of poor- to niedium-
10 quality midrange drivers, and may include a woofer and/or sub woofer. Hie client site 230 in which the client and computer are located is the last component of the listening environment: it considerably affects the quality of the perceived sound because of its spectral response, such as resonances, and the ambient noise that it introduces.
15 The transmitting' audio codec 20^ and receiving audio codec 246 are-
designed to produce an output that is substantially similar to the input signal given the bandwidth limitations of the connection between them. The data-compression processes of those codecs (204, 246) may inn-oduce undesirable artifacts and distortions. Those compression procedures are not necessarily
20 modified by the advanced techniques described below.
to the configurations of FIG. 2B (and FIG. 3), the enhancement processing element 252 is preferably software associated with the processor. But other arrangements are also envisioned for alternate embodiments. For example, the processing may take place in a specialized digital signal processor located either
25 locally or on a connected device.
FIG. 3 shows ths enhanced processing occurring at the client-side of the network, in accordance with another preferred embodiment. Distinguishing from the embodiment depicted in FIG. 2B, a microphone 302 is included at the client site 300 in the embodiment depicted in FIG. 3, The microphone 302 is connected

30 via coupling 306 to the enhancement processing element 252 to provide feedback to the element. Based on that feedback, the enhancement processing element 252 is able to provide additional control of the speaker drivers 24S.
Seversi improvements and techniques are utilized to allow for exceptional processing performance with the use of only modest or typical power. One such
12

WO 03/104924 PCT/USQ3/i:788
technique is to do the sound processing using an extended, bit depth to produce ?. large dynamic range in the system, obviating the need .for strong input-limiters ?.nd reducing truncation error noise.
■ The degree to which any type of processing (e.g., mixing of signals, 5 equalizing, compression, etc.) alters the original rfigitd data varies inversely with the bit resolution of the data. For the sake of illustration only, the below describee techniques employ 64-bit sound samples for stages of the data processing, It is contemplated, however, that other sample sizes can be utilized, such as S-b::, ! 6-bit, 24-bit, and 32-bit.
10 FIG. 4 is a block diaeram illustrating sisnal processine functions for
enhancing audio signals, in accordance with a preferred embodiment. In FIG. 4, an audio signal 405 is provided to an artificial intelligence (Al) dynamics compressor 410. The Al dynamics compressor 410 works in tandem with the AI dynamics decompressor 415 through signal line 412 in order to enhance the
15 dynamic range of the incoming audio signal 405 to a desired range. An offset in these two processors 410, 415 creates an overall dynamic expansion of the signs]. After being processed by the AI dynamic compressor 410, the audio signal is processed by two components placed in parallel: a high frequency artifacts masking processor 420; and a clarity processor (mid-range) 425. The high-
20 frequency artifacts masking processor 420, which comprises an adjustable filter and a variable time delay circuit, creates a masking effect for undesirable artifacts and undesirable sound from the incoming audio signa:. The clarity processor 425, which also comprises an adjustable filter with a variable time delay circuit, creates a realignment effect for undesirable mid-range frequencies in the incoming audio
25 signal. After being processed by these two eleraeats, the audio signal is combined by a mixer 427 and fed into a 3D/live enhancer 430, The 3D/live enhancer 430 adds life and stereo perspective to the sound field of the audio signal. The 3D/live enhancer 430 uses three-dimensional modeling to determine the extent of signal processing that occurs. After the audio signal has been processed by the 3D/live
30 enhancer 430, it is processed by the recording environment simulator 435, which adds diffusion, reverb, depth, regeneration, and room decay to the audio signal. The recording environment simulator 435 accomplishes ibese effects without adding resonant modes and nodrs to the virtual recording room. After being processed by the recording environment simulator 435. the audio signal is
13

WO 03/104924 PCT/US03 177S8
processed by a voice eliminator 440, which effectively eliminates vocal track in the audio signal. The function is accomplished because most vocal tracks are centered and are relatively dry in the overall audio signal. After the voice signals have been removed, the audio signal is processed by i. wide stereo enhancer 44?, 5 which adds wider stereo perspective to the sound Held of the audio signal. At C'ds point, the audio Signal is fed into the AI dynamics decompressor 415. where it is processed with artificial intelligence algorithms to ensure that the fall dynamic ranee of the audio signal is restored. After the audio signal is processed by the AI dynamics expansion processor 415, it is then processed by an AI fader and
10 distortion detection processor 450, which adjusts the level (i.e., volume) of the signal until the optimum gain is achieved, The AI fader and distortion detection processor 450 is adapted to dynamically adjust the gain of the audio signal so that a consistent signal level is continuously delivered to the listener. At this point, the processed audio signal 455 may be fed to a driver or set of drivers so that an
15 ■ individual can listen to the signal.
FIG. 5 is a- block diagram illustrating signal processing functions associated with client-sice enhancement of limited bandwidth music, in accordance with a preferred embodiment. While only one channel of processing is shown in TIG. 5, it should be appreciated mat multiple processing channels may
20 be so employed, Further, the below-described decoding and enhancement processes are preferably software routines running on a processor, and therefore references to signal paths refer to common programming techniques of passing data from one routine to another. Thus, consistent with the preferred embodiment, • a signal path or pathway is not intended TO refer to a physical connection;
25 however, distinct connections may be used in alternate embodiments.
The enhancement process starts with the audio signals outputted from the reception codec 246. Initially, the signal is directed through channel input 502 to the limiter 504. The limiter 504 is preferably a standard audio limiter. i.e., a processing function that keeps the louder sections of the sound from
30 overwhelming the downstream processing due to lack of dynamic range. In response to the sound levels, the limiter 50^ makes .gain changes winch may have a coloring effect on me sound, such as "pumping'" and "clipping," Changes in gain, which occur as the result of limiting or decompression, are often noticeable
14

WO 03/10492-1 PCTAJS03/3 77SS
by the listener, and this is referred to as "pumping.13 "Clipping7 occurs when the signal exceeds the maximum possible value available in a system.
The output cf the limiter 504 splits the signal :into four discrete pathways
or bands. They are referred to as the full bandwidth pathway 510, the bass
5 pathway 520, the r.iidrange pathway 540, and the treble pathway 560. Each
pathway is preferably processed independently. The full bandwidth pathway 510
■ is for the full-bandwidth sound to reach the output mixer 578. Li contrast with the
processing of the various filtered bands discussed below, the full band pathway
■■■ 510 is preferably not sound-level decompressed. The bass, midrange, and treble
10 pathways (520. 540, 560) preferably filter the signal into non-overlapping
frequency bands.
It should be appreciated that more or fewer pathways may be employed. For example, there may be an additional pathway for a sub-woofer band and the mid-frequency band may be divided into two separate mid-frequency bands. 15 When the number of frequency bands used in an alternate embodiment is very high, the filtering is preferably provided by an ARBI filter. For example, the limiter 504 may be an AREI filter having three hundred stereo channels for dynamic, parametric filtering (and therefore also require three hundred stereo channels of sound- level decompression and three hundred stereo channels of 20 time-delay alignment).
Prior to processing, the respective inputs of full bandwidth, bass.
midrange, and treble pathways (510, 520, 540, 560), are amplified by amplifiers
506a-d. After processing, the respective outputs'of the full bandwidth, bass.
midrange, and treble pathways (510, 520, 540, 560) are amplified by amplifiers
25 507a-d and then combined at the mixer 57$.
Each frequency band formed by the filters is processed independently by the various processing elements shown in FIG. 5 and described in the subsequent paragraphs.
With the exception of the full band pathway 510, each band includes an
30 equalizer for parametric equalization. Such parametric equalizers are denoted by
reference numbers 522, 542, and 562 for the ba$s: midrange,. and treble pathways
(520, 540, 560), respectively. Each such parametric equalizer (522, 542. 56*2)
provides multiple narrow-band filters, each of which has a control for gam,
' bandwidth or llQ;!; and central frequency. The equalizers (522, 542, 562) may
15

WO 03/10-19:4 PCT/t"Si)3/l77SS
include a Nyquist compensation filtert which reduce:; spurious signals clue to sampling aliasing-
A specific, programmable, sounc-Isvel expansion or compression for each
frequency hand is carried out by dynamic processing elements included in each of
5 the bass, midrange and treble pathways (520, 5^0, 560). Such processing
elements preferably comprise various filters toother with an expander and/or
. compressor. The bass pathway 520 preferably comprises a high-shelf filter 524. a
low pass filter 526, and a high pass filter 528, together with an expander 530 and a
compressor 532. The midrange pathway 540 preferably comprises a high-shelf
10 filter 544 and a bandpass pass filter 546, together with an expander 548 and a
compressor 550. The treble pathway 560 preferably comprises a high-shelf filter
564, a low pass filter 566. and a high pass filter 56S, together with an expander
570. The full bandwidth pathway is preferably limited to a compressor 512. It
should be appreciated that the processing elements used in each pathway will vary
15 depending en the number and type of bands associated wiih the pathway as well
as other design choices.
Each band (including full bandwidth pathway 510) preferably also provides time delay alignment elements to compensate for the different time delays that the foregoing elements may produce or which may have been 20 produced in recording or processing on the server side. Such time delays elements are denoted by reference numerals 514, 534, 552 and 572 for the full bandwidth, bass, midrange, and treble pathways (510, 520, 540, 560), respectively. Typically, the time delay for proper alignment will be on the order of microseconds.
After processing, each band output is connected to a mixer 578. The 25 mixer 578 provides a signal balance among the four pathways (510, 520, 540, 560), and directs the mixed signal to a master equalizer 580.
The master equalizer 580 provides parametric -equalization for the signal
that exits the mixer 578. It provides a final, broad-spectrum shaping of the signal.
The equalized signal is then (optionally) passed through highly equalized resonant
30 ■ filters to reinforce the subwoofer and bass frequencies. Such filters preferably
comprise t high-shelf filter 5S2: a low pass filter 584, and a high pass fiUer 536.
■ A wall simulator 590 can be coupled J:o the high pass filter 5S6. The wall simulator 5?0 uses diffuse-field matrix (DFM) techniques to produce time delays simulating ~-he reflections from an actual stage. Simulation of such a sound-
16

WO 03/1049:4 ■ ' ' :■ PCT/i;5n3/i77SS
reflecting environment can add a liveliness, or reverb quality'jo the music, without
introducing unwanted resonant psfafcs.
•j- i..
Convent;onal.-jipFM techniques use numbc-n^heory algorithms for ncn-
harmonic, nor.- resonant wave reflection. For example, the quadratic residues
5 described in Section 15.S and the primitive roeu described in Section 13.9 of
Number Theory in Science and Communication, by M.R. Schroeder, Springer-
Verlag, Berlin, 1936, second edition can be applied in this context. Those
conventional techniques only, however, provide for long-time reflections that
would simulate the "reverb" of a room, A primitive root calculation, which
10 improves upon the methods taught by Schroeder by applying a diffuse field matrix DFM technique so as to provide for early reflections of the sound, i.e.. reflections within 5 to 30 milliseconds of the direct sound, is preferably employed.
The wall simulator 590 can also help to break-up, re-shape, or remove the unwanted effects of strong periodic processing artifacts or troublesome periodic
15 features, The DFM techniques used in the stage simulator do not use regeneration, i.e., feedback from the output to the input of this processing element. Control parameters of this processing stage include the size and distance from the wall.
The output of the wall simulator 590 is directed to the room simulator 592.
20 The room simulator 592 uses DFM techniques to produce time delays and resonances that are similar to natural room acoustics. The DFM techniques are similar to those used in the wall simulator 590, but use regeneration. The room simulator 592 can add reverb and decay to enhance dry musical material, and' further obscure subtle codec-induced distortions. Other parameters of this
25 processing stage include room size, room aspect ratios, and the -wet/dry mix. Another use of the room simulator 592 is to compensate for poor room acoustics in the listener's listening environment- The same DFM techniques used for adding natural room or stage acoustics to a dry signal, as described above, can also be used to de-emphasize resonances or filtering in the listener's room, and to
30 provide for a reduction in the room's perceived ambient noise level. For this purpose, "he listener's room acoustics are obtained by the use of a microphone placed near the listener's usual listening location, and functionally connected to the C?V. as shown in FIG. I. DFM techniques *re preferably used only in the
1"

WO 03/104924 '""" PCT/US03/1778S
wall simulator 590 and the room simulator 592, where only the room simulator 592 uses regenerative components.
Various filters may be applied based on the qualities of the client site or
listening room, which may be measured and compensated for by the room
5 simulator 592. One titter may compensate for the acoustics of the Listening room,
which is based on a Transform function, R( much of the room has soft surfaces, such as carpet, drapes or cushioned furniture,
then it is likely thv. the room transform R(to) will fall-off at high frequencies.
However if the listening room has many hard surfaces, then it is likely that the
10 high-frequency end of the room transform R(co) will not fall-off to such a degree.
The initial step for accomplishing room-resonance compensation is the
determination of the acoustics of the listening room using the microphone 302
(see FIG. 3). The room acoustics are determined by using xht speakers 250 (see
FIG, 3) to produce sound having a known frequency spectrum N0(co), and
15 monitoring the effects of the room acoustics on the sound produced by the
speakers using the microphone. The speakers 250 produce a sound such as "white
noise." which has equal energy at each frequency. The spectrum N,(w) of the
signal transduced by the microphone is then used TO calculate the room transform
. R(ci)) according TO
20 R(fi>)» Ni(cfl) / [N0()],
where both spectra Ni(a>) and N0(o) are measured in decibels on the SPLA scale, and, as above, M(to) is the transform produced by the microphone. Or» if N„(co) is a "flat" white noise spectrum, as in the preferred embodiment, then R(©)-Ni(»)/[kM(a>)], 25 A typical compensating room filter would then be just the inverse of the Toom's spectrum, or
F(fi>) = l/R(fl>), where F(to) is a compensating filter for the listening room. The filter F(co) can be implemented in the enhancer either in the room simulator 592 or the master 30 equalizer 5 SO, or in both.
Another filter may be employed to compensate for ambient noise. Ambient room noise compensation is obtained by boosting specific spectral bands of the music over the corresponding bands of ambient room noise. Such boosting improves the signal-to-noise ratio, and hence the clarity, of the music without
IS

WO 03/104924 FCT/U5Q3/J77S8
resorting to turning up the overall volume. This noise reduction technique will
penorm well when the noise spectrum is essentially unchanging. As with the
filter for acoustics, the microphone 302 (sec FIG. 3) maybe employed to obtain a
measure of the ambient noise within the listening room. The transduction from
5 sound to electricity is described by a microphone transform function, M(co).
Therefore, tlie transform describing the trans formation from the original sound
spectrum to the spectrum of the signal transduced by the microphone is given by
U(m) ■ T(co) = M(Qi) • R(CD) * S(ci) • C(u) ■ l( The sound heard by the listener is most accurately monitored by placing
10 the microphone 302 near the location of the listener. The spectrum of the filter to compensate for ambient noise will typically have the same general shape as the ambient noise spectrum. Such filter can also be implemented in the enhancer either in the room simulator 592 or the master equalizer 5SQ, or in both.
Further enhancement may be obtained by compensating for the
15 environment in which the music was recorded or a simulated recording environment (which may actually differ from the environment in which the music was recorded). The client is given a choice of multiple recording environments. According to ihe preferred embodiment, the following six simulated recording environments may be selected by a client: studio (A, B)> hall (A, E), and stadium.
20 For instance, in a studio environment there will be an enhancement of early reflections. Or. in a simulated hall environment there v.'ill be short reverb times, while a simulated stadium will have considerably longer reverb times. In a sense, the user becomes a "producer" in that the user simulates how the music was recorded.' Alternatively, the application of simulated recording environments may
25 be based solely on the actual environment in which the music was recorded, rather than the user's preference. In this case, the system would correct for unwanted artifacts from the recording, and downloaded or streamed files may include a tag, .such as the ID3 tag of MP3 files, which will identify the appropriate recording room acoustics.
30 The output of the room simulator 592 is connected to the karaoke element
:593. The karaoke element 593 has inputs from the ro:>m simulators from both
stereo channels. These left and right channel signals are compared, and musical ;
components, such as voices, that have equal energy rn both channels may be
19 '

WO 03/104914 PCT/US03/n?SS
removed to provide a karaoke effect. This is preferably clone in a similar manner in the 3D enhancer 595, discussed below, except that the karaoke element 593 docs not re-introduce the original stereo signals.
The output c: fha karaoke element 595 is connected to the wide element 5 594. The wide dement 594 compares left and right channels and then performs arithmetic and delay functions io the two channels in order to change the ■perceived distance between (hem. This effect changes the perceived stereo-separation spread o: the music. Whereas other attempts to produce an enhanced wideness result in £ loss of the low-frequency portion of the signal, the wide
ID element 594 can produce this separation while leaving the low-frequency
components substantially unaltered. Processing of this effect is integrated into
■ standard PL-2 processing, a positioning algorithm distributed by Dolby
Corporation of San Francisco, California. Specifically, the karaoke element 593,
the wide element 594, and the 3D enhancer 595 (discussed below), winch each
15 require interaction between the left and right channel, accomplish PL-2 decoding with the combined use of both channels.
The output of the wide element 594 is connected to the 3D enhancer 595. The 3D enhancer 595 removes "equal energy" (common-mode) signal content from the stereo signal, (usually solo vocals and instruments) delays it, then
20 re-mixes it with the raw signal using a combination of frequency and time-domain functions. This provides a "widened11 sound stage to the listener without delocalizing the equal-energy material.
The output of the 3D enhancer 595 is then connected to the leveling . amplifier 596. In turn, the leveling arfiplifier 596 is connected to the AI level
25 control 597. The AX level control 597 circuit functions to lower the audio level dining peak events and then return it after a peak event has passed. _To keep ■sound from distorting during the listening process or while recording it, a human engineer would always drop the volume, by moving the volume control down of ::he offending instrument or vocal. By essentially simulating a human engineer,
30 'he AI level control 597 rapidly moves the audio level down by analyzing the digital stream for distortion and signal overload; to identify peak events. It then returns the volume towards the initial volume seninsr after the peak event has occurred, wirhout the need for an "always-or." audio compressor circuit, which undesirably leads to a loss of dynamic edge and fiat sound.
20

WO 03/10492-1 PO7US03/17788
The output oi the AI level control 597 is connected to the master expander
593. which is used to selectively increase the dynamic range of the mastered
stsrso signal. Output from the master expander 59S is connected to an amplifier
599.
5 The master sxpander 59S controls the foil output volume level of the
system. It allocs the listener to set the volume level as high as he or she likes without having to ^-vorry about overdriving the speaker driver circuitry or the speakers, This feature is accomplished by a process that detects a speaker-overdriving peek sound level by monitoring for distorted samples. According to
10 the preferred embodiment, a fuzzy logic tally of The amount of clipping is used xo
determine the degree to which ths volume level should be reduced. Alternatively,
the process may lock ahead at ths music stream inc predict the arrival of a
.speaker-overdriving peak sound level. If such a level is reached or predicted to be
reached, the master gain level is automatically tumsd down using a non-linear
15 attenuation-versus-time curve which simulates the atienuation-versus-tirns that a live person would use.
The master expander 598 is the final stage of enhancement processing and provides the enhanced signal to channel output 504, which, in turn, connects to the speaker driver circuitry. The speaker driver circuitry converts the processor's
20 ' enhanced digital representation of the signal into a hardware analog signal, and provides the necessary amplification and connectivity to the speaker.
The sound-level decompression described herein provides a widening of the dynamic range of the music to help correct for compressions of the audio signal that have occurred at any time from the recording of the original audio
25 source onwards. Typically, the recording and mixing of music includes sound-level compression of many of the tracks so as to fake advantage of the limited dynamic range of the recording medium, Also, some form of compression may be applied post-recording, to reduce the bandwidth for Internet broadcast purposes. This latter type of compression may be substantially removed by the reception
30 codec, but may have been insufficiently corrected for, or otherwise be in need of farther expansion to improve the ':lfveness." or other subjective qualities, of the music. A processing feature using dynamics with different time constants and expansion factors for each emphasis band is preferably employed.
21

WO 03/104924 PC'T/L'S03/1"7S8
The various processing dements shown in FIG. 5 may be controlled by a master control program that can bypass any of the processes, and can specify the .parameters of each process. The "skin" is the interface which allows the client to control parameters and presets, i.e... the "skin" is the visual and inter active part at 5 the enhancement pre gram displayed on the listener's PC screen. Fader controls are available for the listener to specif}' each parameter ifl the system, and "radio buttons'' (i.e. on/off switches) are available to select groups of preset parameters. The enhancement parameters may be adjusted separately, or various presets may be chosen.
10 The system may include a '"'"bigness" control that simultaneously controls
the.parameters of the individual band processors. For low values of the irbigness"': parameter, less dynamic processing occurs, and the sound-level dynamic range is equal to that of the music as recorded: For high values of the lvbjgncss" parameter, each band's processing dynamics are increased relative to the sound-
15 level dynamic range of the recorded music.
Preset parameter groups are of two types; listener denned and built-in. Listeners can select presets from their own previously labeled groups, or can select from a menu of built-in presets. Built-in presets are designed based on considerations of bandwidth, code type, listeners' speakers, and music type. Once
20 a listener selects a built-in preset, the listener may than adjust any individual parameter or group of parameters to customize the built-in preset. That adjusted group of parameters can then be labeled and saved as a new preset. For example, if a built-in preset is selected, then the listener may subsequently select a group of room-compensation parameters that may be-applied to the selected built-in preset.
25 FIG. 6 is a block diagram illustrating a ?TJ> enhancer in accordance with a
preferred embodiment. As with other elements* this element has a left input 602 and a right input 604 as well as a left output 650 and a right output 652. One mixer 640 is associated with left output 650, while another mixer 642 is associated with right output 652. The signal associated with left input 602 is passed through
30 £. low pass filter 606 and a high pass filter 608. Similarly, the signal associated with right input 604 is passed through a low pass filter 610 and a high pass filter 612. The outputs of the low pass filters 606 and 610 are respectively passed through amplifier 622 and amplifier 62S. the outputs of which are respectively directed onto mixer 6^0 and mixer 6^2, Similarly, the outputs of the high pass
22

WO 03/10492-1 PC'T.'l'SOj.'l 77S8
filters 60S and 612 are respectively passed through amplifier 624 and amplifier 626, the outputs of which are respectively directed onto mixer 640 and mixer 642. The outputs of the high pass filters 608 and 612 are also surged together at adder 652 and then directed toward amplifier 634. The output of amplifier 634 is passed 5 onto mixer 640 as well as onto time delay element 636, the output of which is further directed to mixer 642.
The 3D enhancer element is suitably configured to provide ?. widened souudstage to the lis'.tncr. The 3D enhance: clement, which is similar to the spatial enhancer element described below in connection with FIG. 11, removes
10 "equal energy" (common-mode) signal content from the stereo signal (usually solo vocals and instruments), delays it. then re-mixes it with the raw signal using a combination of frequency and time-domain functions. This provides a "widened" sound stage to the listener without dslocalizingihe equal-energy material.
FIG. 7 is a block diagram illustrating a wide element, in accordance with a
15 preferred embodiment. As with other elements, this element has a left input 702 and a right input 704 as well as a left output 750 and a right output 752. One mixer 740 is associated with left output 750, while another mixer 742 is associated with right output 752. The signal associated with loft input 702 is passed through a high pass filter 706 and a low pass filter 70S. Similarly, the signal associated
20 with right input 704 is passed through a high pass filter 710 and a low pass filter 712. The outputs of the low pass filters 70S and 712 arc respectively directed onto mixer 740 and mixer 742. Similarly, the outputs of the high pass filters 706 and 710 are respectively passed through time delay elements 724 and 726, the outputs of v/hich are respectively directed onto mixer 740 and mixer 742. Preferably, the
25 time delay provided by time delay element 724 is greater than the time delay provided by time delay element 726. For example, the time delay associated with slemeat 724 may be 0.05-2.0 milliseconds while the time delay associated with element 726 maybe 0.5-30 milliseconds.
The wide element is preferably configured to produce a desired time.
30 -iUfferential between the left and right channel high frequency information, as processed by the respective high pass filter:; 706/710. The respective time delay elements 724/726 can be adjusted to provide the desired differential time delay. In practical embodiments, the differential time delay is between 5 and 22 milliseconds, and preferably about 20 milliseconds, which falls within the Haa?
23

li
WO 03/10492-1 PCT/USU3/17?8S
effect (or precedence effect) range. In operation, one of the time delay elements can be set to a fixed delay value while the other time delay element is varied to achieve the desired Kaas effect.
FIG. 8 is a block diagram illustrating an alternative embodiment or the 5 enhancement processor according to the disclosed method/system. The system depicted in FIG. 8 includes many of ihs same elements depicted in FIG. d and also operates in the same manner as described above. It should be noted, however, that FIG. S includes the following additional elements: a bass dynamics processor ?02; time delay elements 905, 91S and 919; a DFM wall simulator 909; an offset
10 device 907; a wave generator 915; a gain window threshold processor 917 and a voice V detection circuit 91S. Also depicted in FIG. S are a speaker 92: (with an accompanying amplifier 920) and a microphone 922. The bass dynamics orocesso: 902 comprises a special filter combined with & variable time delay circuit and compressor and expands: blocks to enhance a dynamic bass sound,
15 The wall simulator 909 performs the same functions as described above with respect to the previous figures. In embodiments deployed on XSG-compatible processors (PCs and derivative devices), *hc wave generator 915 is used to prevent Intel FPU 'denormal*' operation during periods of silence. The offset device 907 is used to allow communications between the Al dynamics compressor 901 and
20 "he AI dynamics decompressor 913. Il should also be noted that the AI fader and distortion detection device 916 can be used to monitor the listening environment 923 and provide feedback so that an appropriate gain level can be applied to the output signal. This can be performed through the use of a Fietcher-Munson look-up table.
25 FIGS. 9-16 illustrate various aspects of another preferred embodiment of
the invention that can be implemented at a client-side processing component such as a personal computer or other device capable of processing digital audio files for playback to a user,
FIG. 9 is a block diagram illustrating signal processing functions
30 associated with thent-side enhancement of limited1, bandwidth music, in accordance with a preferred embodiment. In a practical embodiment, the architecture 900 depicted in FIG. 9 can be realised in hardware, software, .firmware, or any combination thereof. While only one channel of processing is shown in FIG. 9, it should be appreciated that multiple processing channels may
24

WO 03/104924 PCT/T503/177SS
bs so employed. For example, although a single channel, mono channels, c-r stereo channels are described, herein, multiples of ;hese described-channels may be employed to provide additional functionality and sound processing., as needed. Further: within a channel, although a specific number of pathways may be 5 described herein, ii is to be understood that fewer or more such pathways may be employed within the spirit of this invention.
Further, the below-described decoding and enhancement processes are preferably software routines running on a processor, and therefore references to signal paths refer to common programming techniques of passing data from one
10 routine to another. Thus, consistent with the preferred embodiment, a signal path or pathway is not intended to refer to a physical connection; however, distinct connections may be used in some practical embodiments.
The enhancement process starts with the audio signals outputted from the reception codec. Initially, the signal is direc:ed through a channel inpuc 902 to a
15 compressor 904. The compressor 904 is preferably a standard audio limits:, i.e., a processing function that keeps the louder sections of the sound from 'overwhelming the downstream processing due to lack of dynamic range. In response to the sound levels, the compressor 904 makes gam changes which may have a coloring effect on the sound, such as "pumping'' and 'clipping.'1 Changes
20 in gain, which occur as the result of limiting or decompression, are often noticeable by Hie listener, and this is referred to as "pumping." "Clipping7' occurs when the signal exceeds the maximum possible value available in a system.
The output of die compressor: 904- splits the signal into a plurality of . discrete pathways or bands, at least one of which corresponds to a full bandwidth
25 signal. In the preferred embodiment, the output of the compressor 904 is directed to four streams. They arc referred-to as the full bandwidth pathway 906, the bass pathway 90S. the midrangs pathway 910, and the treble pathway 912. Each pathway is preferably processed independently. The full bandwidth pathway 906 is for the fall-band width sound to reach an output rrrixer 913. In contrast with the
30 processing of the various filtered bands discussed b;low\ the nil! bandwidth pathway 9C6 is preferably not sound-level decompressed. The bass, midrange, and treble pathways 90S/910/912 preferably niter the signal into non-overlapping frequency bands.
25

WO 03/jOJy24 PCT/US03/1T7SS
It should be appreciated that mors or fewer pathways way be employed. For example, there may bs an additional pathway for a sub-woofer band and the. mid-frequency band may be divided into two separate mid-frequency bands. When the number of frequency bands used in an alternate embodiment is very
5 high, the filtering may be provided by an ARBI filler. For example, the compressor 904 may be an ARBI filter having three hundred stereo channels for dynamic, parametric filtering,
Prior to processing, the respective inputs of the full bandwidth, bass, midrange, and treble pathways 906/908/910/912 are, amplified by respective
10 variable gain amplifiers 9l4a-d, In a practical embodiment, each of the variable gain amplifiers employed by the processing architecture 900 has an adjustable gain between -30 d3 and +25 d3, with an adjustment resolution of 0.1 dB. In operation, a number of settings and/or adjustable features of the processing architecture, including the adjustable gain settings of the amplifiers 914. may be
15 determined according to the requirements of other processing functions described herein which are performed in connection with the operation of the present invention. After processing, the respective outputs of the full bandwidth, bass, midrange, and treble pathways 906/90S/910/912 3re amplified by variable gain amplifiers 9l6a-d and then combined at the mixer 913.
20 Each frequency band formed by the niters is processed independently by
the various processing elements shown in FIG-. 9 and described in more detail below, A specific, programmable, sound-level expansion or compression for each frequency band is carried out by dynamic processing elements included in each of the bass, midrange, and treble pathways 903/910/912. Such processing elements
25 preferably comprise various filters together with an expander and/or compressor. For example, the bass pathway 908 preferably includes at least a low pass filter 913 and a compressor 920. The midrange pathway 910 preferably includes at least %. bandpass pass filter 922 and a compressor 924. The treble pathway 912 preferably includes at least a high pass filter 926 and a compressor 92S. In the
30 example embodiment, the full bandwidth pathway 906 includes a compressor 930 and need not utilize any filtering elements, It should be appreciated that the processing elements used in each pathway can vary depending on the number znc type of bands associated with the pathway as well as other design choices.
26

WO 03/10492-1 PCI7US0J/17-SS
As mentioned above, the processed signal corresponding to each band pathway serves as a respective input to the .mixer 913. The mixer 913 provides i signal balance among the four pathways, and directs the mixed signal 932 to 2. number of selectable (i.e., capable of being bypassed) 01 optional processing 5 elements. FIG. 9 depicts u preferred ordering of these processing elements. Alternate embodiments of the invention, however, may utilize a different ordering of such processing elements and/or employ additional or alternative processing elements.
In the exsmpie embodiment, the mixed signal 932 serves as an input to a
10 vocal enhancer element 934, which is suitably configured to enhance voices and
solo instruments in the time domain without additional frequency domain coloring
or overtone unbalancing with- relation to the fundamental frequencies of the solo
instruments or voca! materials in the stereo waveform. Ons example vocal
enhancer element is described in more detail below in connection with FIG. 10.
15 The output of the vocal enhancer element 934 is then (optionally) passed through
highly equalised isscnani filters to reinforce the subwoofer and bass frequencies.
Such filters preferably comprise a high-shelf filter 936, a low pass filter 938, and a
high pass filter 940. The high-shelf filter 936 emphasizes the range of frequencies
above a given "crossover" frequency. The "steepness" of the crossover is
20 adjustable by varying the "Q" or quality factor of the filter.
The filtered output signal may be directed to a spatial enhancer element 942, which is configured to provide a widened soundstage to the listener. The spatial enhancer element 942 removes "equal energy" (common-mode) signal content from the stereo signal (usually solo vocals and instruments), delays it, then 25 re-mixes it with the raw signal using a combination cf frequency and time-domain functions. This provides a "widened" sound stage to the listener-without delocalizing the equal-energy material. -
One example spatial enhancer element is described in more detail below in connection with FIG. 11. In the example embodiment, the output of the spatial 30 enhancer element 942 senses as an input to a walls simulator element 944. The walls simulator element 944 preferably uses diffuse-field matrix (DFM) techniques to produce time delays simulating the reflections from, an actual stage. Simulation of such B sound-reflecting environment can add a liveliness, or reverb quality to the music, without introducing unwanted resonant peaks. One example.*
27

walls simulator element is described in more detail below in connection with FIG. 12.
Conventional DFM techniques use number theory algorithms for non-haimonic, non-resonant wave reflection. For- example, the quadratic residues described in Section o.S and the primitive roots described in Section 13.9 of Number Theory in Science and Communication, by MR. Schroeder, Springer-Verlag, Berlin. Second Edition (1936), can be applied in this context. Those conventional techniques only, however, provide for long-time reflections thai would simulate the "reverb" of a room. A primitive root calculation, which improves upon the methods taught by Schroeder by applying a "diffuse field matrix" ("DFM") technique so as to provide for early reflections of the sound, i.e.. reflections within 5 to 30 milliseconds of the direct .sound, is preferably employed.
The walls simulator element 944 can also help to break-up, re-shape, or remove the unwanted effects of strong periodic processing artifacts or troublesome periodic features. The DFM techniques used m the stage simulator do not use regeneration, i.e., feedback from the output to the input of this processing element. Control parameters of this processing stage include the size and distance from the wall.
In the example embodiment, the output of the walls simulator element 944 is directed to a room simulator element 946. One example room simulator element is described in more detail below in connection with FIG. 13. The room simulator element 946 uses DFM techniques to produce time delays and resonances that are similar to natural room acoustics. The DFM techniques are similar to those used in the walls simulator element 944, but use regeneration. The room simulator element 946 can add reverb and decay, or can add DFM without reverb, to enhance dry musical material, and further obscure subtle codec-induced distortions. Other parameters of this processing stage include room size, room aspect ratios, and the wet/dry mix (where "dry" refers to a lack of effects and "wet" refers to the use of effects). Mother use of die room simulator element 946 is to compensate for poor room acoustics in the listener's listening environment, The same DFM techniques used for adding natural room or stage acoustics to a dry signal, as described above, can also be used to de-emphasize » resonances or filtering in the listener's room, and to provide for a reduction in the n room's perceived ambient noise level.
2S

WO 03/J04924 PCT/US03/J7TSS
Various filters may be applied based on the qualities of the client site or listening room, which may be measured and compensated for by the room simulator element 946. One filter may compensate- for the acoustics of the listening room, which is bassd on a transform function, l^(co), having a number of 5 resonances, If much of the room has sofl surfaces, such ss carpet, drapes, or cushioned furniture, then it is likely that the room transform R(co) will fall off at high frequencies. However, if the listening room has many hard surfaces, then it is likely that the high frequency end of the room transform R(co) will not fall off to such a degree.
10'. Further enhancement may be obtained by compensating for the
environment in which the music was recorded, or a simulated recording environment (which may actually differ from the environment in which the music was recorded). The client is given a choice of multiple recording environments. According to the preferred embodiment, the following ten simulated recording
15 environments may be selected by a client; audio studio, jazz session., nightclub.
game space, bass jam, theater, rock concert, sonic wide, symphony; or cathedral.
For instance, in a srudio environment there will be an enhancement of early
\ reflections (DFM). Or, in a simulated hall environment there will be short reverb
times, while a simulated stadium will have considerably longer reverb times. In a
20 sense, the user becomes a "producer" in that the user simulates how the music was
recorded. Alternatively, the application of simulated recording environments may
be based solely on the actual environment in which the music was recorded, rather
than the user's preference. In this case, the system would correct for unwanted
. artifacts from the recording, and downloaded or streamed files may include a tag,
25 such as the ID3 tag of MP3 files, which will identify the appropriate recording room acoustics.
The output of the room simulator element 946 is connected to a subsonic enhancer element 948, which is suitably configured to provide low-bass reinforcement of the signal. One example subsonic enhancer element is described
30 in more detail below in connection with FIG. 14.
The output of the subsonic enhancer element S4S is connected to a look-ahead automatic gain control (AGC) element 950. The look-ahead AGC element 950 is suitably configured to provide control of the output dynamic range of the entire process. The "look-ahead" terminology refers to the delay of the signal,
' 29

,. u VJ/JUHWJ VC7/VSQ3H 71SS
vvjuch gives the control amplifier enough time to change gain smoothly, without introducing transients, or "pumping" in the output. This feature operates to lower the audio level during peak events and then return it after a peak event has pissed. To keep sound from distorting during the listening process or while recording it, i 5 human engineer would always drop the volume, by moving the volume control down of the offending instrument or vocal. By essentially simulating a human engineer, the look-ahead AGC element 950 rapidly moves the audio level dowr. by analyzing the digital stream for distortion and signal overloads to identify peak events. It then returns the volume towards the initial vohrme setting after the peak
10 event has occurred, without the need for an ualways-on" audio compressor circuit, which undesirably lead? to & loss of dynamic edge and .Oat sound.
One example look-ahead AGC element is described in more detail below -..n connection with FIG. IS. Notably, the look-ahead AGC clement 950 may include one or more delay elements (not shown) that compensate for different
15 lime delays that the various processing elements may generate, or which may have been produced during recording or processing at the s&rver side. Typically, the -time delay for proper alignment will be on the order of microseconds.
In this example embodiment, the look-ahead AGC element 950 is the final stage of enhancement processing and provides the enhanced signal to a channel
20 output 952, which, in mm, connects to the speaker driver circuitry. The speaker driver circuitry converts the processor's enhanced digital representation of the signal into a hardware analog signal, and provides the necessary amplification and connectivity to the speaker.
The preferred ordering of the individual processing components (between
25 the mixer 913 and the channel output 952) is shov/n in FIG. 9. Practical embodiments, however, may employ a different ordering of such components as necessary to suit the needs of the particular application or to meet'the demands of the particular listener. Furthermore, additional and/or alternative processing elements may be utilised in alternate embodiments of the invention.
30 The sound-level decompression described herein provides a widening of
the dynamic range of the music to help correct tor compressions of the audio signal that have occurred at any time from the recording of the original audio source onwards. Typically, the recording and mining cf music includes sound-level compression of many of the tracks so as to take advantage of the limited
30

WO 03/104924 PCT/DS03/177SS
affecting the primary fundamental^frequencies that give voices depth and fullness. In operation, a number of settings and/or adjustable fei-circs of the vocal enhancer element 1000 may be determined according to the requirements of other processing functions described herein which are performed in connection with the 5 operation of the present invention.
The vocal enhancer element 1000 is a stereo processing component - it receives a Jen Input signal 1002 'dud a right input signal 1004. and produces a corresponding left output signal 1006 and a corresponding right output signal ] 00S. The left channel input signal 1002 is routed i.o an absolute value generator
10 '.010, which generates an output signal 1012 that represents the absolute value of ths left input signal 1002, The right channel input signal 1004 is routed to an absolute value generator 1014, which generates an output signal 1016 that represents the absolute value of the right input signal 1004. In other words, the left and right channel input signals are full-wave rectified. A comparator 1018
15 receives the two output signals 1012/1016 and produce;, a difference signal 1020 ±at represents the output signal 1012 subtracted from the output signal 1016. The voltage of the difference signal 1020 is proportional to the differences between ths left and right inputs.
The derived difference voltage is then filtered -.o remove fast transients,
20 becoming a control voltage. The output of the comparator 101S is connected to ons end of a variable resistance 1022. The second ead of the variable resistance 1022 is connected to (or corresponds to) a node 1024. The first end of another variable resistance 1026 is also connected to node 1024. The second end of the variable resistance 1026 is connected to the first end of a variable capacitance
25 102S, and the second end of the variable capacitance 1Q2S is connected to a reference voltage, e.g., ground. The variable resistance 1022, the variable resistance 1026, and the variable capacitance 1028 can be independently'adjusted to provide a suitable level and cross over frequency. These variable components from an adjustable low pass filter arrangement that conditions the difference
30 signal 1020 into a suitable control signal 1029 preswr. at node 1024.
The left input signal 1002 also serves as ar, input to a first voltage controlled amplifier 1030, and the right input signal 1004 also serves as an input TO a second voltage conrroiled amplifier 1032. The differential nature of the voltage controlled amplifiers equalizes the signal amplitude of the lef; and right
32

WO 03/104924 PCX/US03/1 "X$
channel audio levels over time. The control signal 1029 adjusts the gain of the two voltage controlled amplifiers 3 030/1032 - the output signal 1054 of the voltage controlled amplifier 1050 represents an. amplified version of the left input signal 1002 and the output signal 1036 of the volhige controlled amplifier 1032 5 represents an amplified version of the right input signal 1004, These two output signals 1034/1036 are fed into a summer 1032. which produces a summed output signal 1040. The summer 1038 effectively removes any opposite-phase material, and creates a synthesized "vocal" or "center" channel. This takes advantage of the fact that most voc2l tracks are mixed with equal energy into the left and right
10 channels when originally recorded. The summed output signal 1040 serves as an input to an adjustable gain amplifier ,1042, to provide a suitable signal level. The output of amplifier 1042 is then processed by a band pass filter arrangement 1044 ':o produce a filtered signal 1046. The band pass filter arrangement 1044 removes bass and treble content outside of the desired vocal range.
15 The left input signal 1002 also serves as an input to a summer 104S, and
the right input signal 1004 also serves as an input to a summer 1050. The summer 1048 generates the sum of the left input signal 1002 and the filtered signal 1046: this sum represents the left output signal 1006. The summer 1050 generates the sum of the right input signal 1004 and the filtered signal 1046: this sum represents
20 the right output signal 100S. These summers 104S/1050 mix the vocal output with the original left and right channel signals, thus emphasizing the vocal content of the source material.
The spatial enhancer element creates a complex sound field enhancement by stripping common-mix material from the stereo signal, then mixing the result
25 back into the left channel directly, and xhe right channel with an appropriate delay. Bass content is removed from the original signals before processing, then re¬applied in the "final" left and right channel mixers, thus preventing low frequency bass energy from compromising the effectiveness of the "stripper" circuit. FIG. 11 is a schematic representation of an example spatial enhancer element 1100
30 suitable for use with the architecture depicted in FIG. 9. hi operation, a number of settings and/or adjustable features of the spatial enhancer element 1100 may be determined according to the requirements of other processing functions described herein which are performed in connection with the operation of the present invention.
33

WO 03/104924 PCT/USU3/177SS
The spatial enhancer element 1100 is a stereo processing component - it
receives a left input signal 1102 and a right input signal 1104. and produces a
corresponding left output signal 1106 and a corresponding light output signal
110S. One mixer II10 is associated with the left ouiput signal 1106, while
: another mixer U12 is associated with the right output signal 1108.
The left input signal 1102 is passed through a low pass filter 1114 and i high pass filter 1116. In the example embodiment, the low pass filter 1,114 is realized as a second order filter having an adjustable cutoff frequency that is typically set at approximately 300 Hz. This filter is utilized to isolate the low
10 frequency content such that it does not imbalance the spatial enhancer element 1100 or generate undesirable artifacts. In ibe example embodiment, the high pass filter 1116 is realized as a second order filter having an adjustable cutoff frequency that is typically set at approximately 300 Hz. Similarly, the right input signal is passed through a low pass filter 1US and a high pass filter 1120. In the
15 . preferred embodiment; the characteristics of the low pass filter 1113 match the characteristics of the low pass filler 1114. and the characteristics of the high pass filter 1120 match the characteristics of the high pass Slier 1116.
The outputs of the low pass filters 1114 and 11 IS are respectively passed through a variable gain amplifier 1122 and a variable gain amplifier 1124, the
20 outputs of which are respectively directed into mixer 1110 aud mixer 1112. Similarly, the outputs of the high pass filters 1116 and 1120 are respectively passed through a variable gain amplifier 1126 and a variable gain amplifier 1128, the outputs of which are respectively directed into mixer 1110 and mixer 1112. In a practical embodiment, each of the variable gain amplifiers employed by the
25 spatial enhancer elsment 1100 has an adjustable gain between -30 dB and +25 dB: with an adjustment resolution of 0.1 dB. The outputs of the high pass filters! 116 . and 1120 are also used as inputs to a subtracter 1130. The ouiput of the subtractor 1130 represents the output of the high pass filter 1116 minus the Output of the high pass filter 1120. This operation effectively phase-cancels any material
30 common to both channels. This creates the "stripped"' signal The output of the subtractor 1130 is then directed toward a variable gain amplifier 1132. The output of the variable gain amplifier 1132 serves as an additional input to mixer 1110, as well 2S an input to a time delay element 1134,
34

WO 03/104924 PCT/L'St)3/l7?S8
The time delay element 1134 is configured to introduce a delay of between
0.05 ms to 30 ms (e.g.. 1 io 1440 samples at a sampling frequency of 4S kHz). In
. operation, the specific amount of delay may be determined according to the
requirements of other processing functions described herein which are performed
5 i:.i connection with the operation of die present invention. The time delay
■ simulates a spatial fu^ciion related to the distance between the listener's ears. In
practical implementations, the time delay should DOT exceed approximately 2.2
ms. In one preferred embodiment, the time delay is about 1.1 ms. The output of
the time delay element 1134 serves as an additional input to ths mixer 1112.
10 The mixer 1110 mncrions as a summer to combine its input signals. In
practice, ths mixing results in a more complex sound field and spatial
displacement having a wider stereo image. Thus, the spatial enhancer element
1100 emphasizes discrete left and right channel content and remixes that content
with the original signal content. The mixer 1112 functions in a similar manner.
15 " The output of the mixer 1110 serves as an input to a variable gain amplifier 1136,
■.he output of which represents the left channel output signal 1106. The output of
the mixer 1112 serves as an input to a variable gain amplifier 113S? the output of
which represents the right channel output signal 11 OS. The left and right output
signals 1106/1108 can be routed to additional processing elements utilized in the
20 architecture, such as the walls effect element 944 (see FIG. 9).
The Wall Effect element is used to add artificial early reflections to the
signal, simulating the effect of nearby reflective surfaces close to the performance
source. No regeneration is used with this element. In the example embodiment,
the signal path maybe summarized as follows:
25 • Predetermined "lap' points are created in a circular delay line, by
calculating the distribution of primitive roots across a reflective
surface.
• The signal is low-pass filtered to approximate the frequency
response of the desired reflective surface.
30 • The filtered signal is applied to the circular delay line.
• The delayed signal is "tapped'1 at ths predetermined tap points
down ibe delay line. The tapped values are summed in decreasing
35

WO 03/104924 PCT/LS03/J77SX
amplitude, approximating the effect of a:.r losses over distance
points along the reflective surface.
• The synthesized reflective "wet" signal is mixed in ratio with the
original ''-dry' signal to provide the block output.
5 FIG. 12 is a schematic representation of an example Wall Effect element
1210 suitable for use with the architecture depicted in FIG- 9. The Wall Effect element 1210 uses diffuse-field matrix (DFM) techniques to produce time delays simulating the reflections from an actual stage. Simulation of such a sound-reflecting environment can add a liveliness, or can arid diffuse field matrix type 10 energy without reverb to add a "live" quality to the music., without introducing unwanted resonant peaks.
Conventional DFM techniques use number theory algorithms for non-harmonic, non-resonant wave reflection. Fox example, fljs quadratic residues described in section 15.S and the primitive roots described in Section 13.9 of 15 ' Number Theory in Science and Communication, by M.R. Schioeder. Springer-Verlag, Berlin, 1986, 2nd Edition can be applied m this context. Those conventional techniques only, however, provide for long-time reflections that would simulate the "reverb' of a room. A primitive root calculation, which improves upon the methods taught by Schroeder by applying a diffuse field matrix 20 DFM technique so as to provide for early reflections of the sound, i.e., reflections within 5 to 30 milliseconds of the direct sound, is preferably employed.
The Wall Bffect element 1210 can also help to break-up, re-shape, or remove the unwanted effects of strong periodic processing artifacts or troublesome periodic features. The DFM techniques used in the stage simulator 25 do not use regeneration, i.e., feedback from the -output to the input of this processing element. Control parameters of this processing stage include the size and distance from the wall.
Referring to FIG. 12.. an implementation of Wall Effect element 1210 will now be described. It is to be understood that while wall effect processing for a 30 single channel is illustrated in FIG. 12, for a stereo effect, two such channels may be used.
The channel input follows two paths: a direct path 1212 to an input of wet/cry nailer 1214, and a niter, delay and summing path 1216. the output of
36

WO 03/304924 PCT/US03/177S8
which is applied ro another input of wet/dry mixer 121.4. The output of Wall
Effect clement 1210 can be adjusted to provide different ratios or proportions of
information from the direct path 1212 and the processed path 1216,. as indicated
by arrow 121S.
5 Along path 1216, each incoming sample if applied to a low pass filter
1220. Then the filtered sample is applied to a circular delay line 1222. As can be seen from FIG. 12. r,-multiplier taps may be employed at different points in the delay line 1222, to form the sum:
10 where the number of taps equals x+\, D(n) represents the delayed sample nt and S(iJ represents the coefficient to be applied to the product. The value ofx will be governed by the amount of available, processing power in a practical implementation. Thus, the sum of £)*S is formed for HI positions of multiplier ■caps. As a part of the operation, the position indexes for the multiplier taps are
15 [Shifted to the right, and. should the position index run past the end of the delay line, the position indexes are wrapped around to the beginning of delay line 1222. The output of this summing operation is the sum "/'which is applied to one of the inputs to wet/dry mixer 1214.
In the example of the Wall Effect element 1210 provided in FIG. 12., the
20 total length of circular delay line 1222 may be 90 msec at a sample rate of Fs - 4S kH2, and there may be six (x^S) multiplier taps. Also, the longest reflection {W) may be less than or equal to 30 msec at a sample rate of Fs = 48 kHz. The length of the' W axis influences the "size" of the wall effect. Also, the "mix" of the wall effect is a function of the wet/dry ratio set (symbolically) by arrow 121S.
25 It is to be understood that as implemented in FIG. 12, Wall Effects.dement
1210 is not a finite impulse response filter (FIR) since a complete convolution is
not performed.
The output of the Wall Effec; element 1210 may be directed to the room
effects element 1310.
30 FIG. 13 is a schematic representation of an example Room Effect element
suitable for use with the architecture depicted in FIG. 9. Referring to FIG. 13, an implementation of Room Effect element 1310 will new be described. While one section of a room effect element implementation is shown in FIG- 13. it is to be
37

WO 03/104924 PCT,T$03/177SS
understood that two or mors such sections may be used for a stereo or multichannel embodiment.
The room effects element 1310 uses DFM techniques io produce time delays and resonances that are similar to natural room acoustics. The DFM 5 techniques arc similar to those used in the Wall "Efecxi element 1210, hut use regeneration. The room effects element 1310 can add reverb and decay to enhance dry musical material, and further obscure subtle codec-induced distortions. Other parameters of this processing stage include room size, room aspect ratios, and the wet/dry mix. The room effects clement 1310 is used to add
10 artificial ivlate,: reflections to the signal, simulating the ambient reflectivity of a ical room environment. The example embodiment uses a combination of eight hand-tuned comb filters in parallel, feeding four all-pass filters in series. The synthesized reflective "wet" signal is mixed in ratio with the original ;;dry" signal to provide the output.
15 Further enhancement may be obtained by compensating for the
environment in which the music was recorded, or a simulated recording environment (which may actually differ from the envirotunent in which the music was recorded). The client is given a choice of multiple recording environments. According to the preferred embodiment, the following ten simulated recording
20 environments may be selected by a client: audio studio, jazz session, nightclub, game space, bass jam, theater, rock concert; sonic wide, symphony, cathedral. For instance, in a studio environment there will be an enhancement of early reflections. Or, in the "night club' environment there will be short reverb times, while a 'cathedral'* will have considerably longer reverb times. In a sense, the
'25 user becomes a "producer" in that the user simulates how the music was recorded. Alternatively, the application of simulated recording environments may.be based solely on the actual environment in which the music was recorded, rather than the user's preference. In this case, the system would correct for unwanted artifacts from the recording, and downloaded or streamed files may include a tag, such as
30 the IDS tag of MPS files, which will identify the appropriate recording room acoustics.
The implementation of Room Effect element 1310 illustrated in FIG. 13 employs s multiplicity of parallel paths (eight (8) such paths I312a-h in tins example) each being processed by a comb filter 13i4a-h. respectively. The
3$

WO 03/10-1924 , PCT/US03/177>
outputs of each of these comb niters 1314 are then summed in summer 13)6: and
then applied to severs! all-pass filter blocks 131S, 1320, 1322, and 1324. Each of
. the comb filters 1314 is parameterized individually to provide 2 different amount
of reverb enhancement to reduce the amount of "metallic" or "tinny" artifacts that
5 are typically produced by conventional processing techniques. The parameters of
the all-pass filter blocks 131S, 1320, 1322, and 1324 are adjusted such that thsir
phase characteristics also contribute to the reduction of such 'metallic" or "tinny"
artifacts. In practical embodiments, the comb filters and all-pass filters iv.ay be
hand-timed by an experienced sound engineer to provide the desired output signal
10 , characteristics.
Following the processing of the sound signals in room effect element
■ 1310, the signals proceed to the subsonic enhancer element.
In the example embodiment, the subsonic effect element uses a
■ combination of an adjustable-Q low-pass filter and a compressor to provide low-
15 bass reinforcement of the signal. The subsonic effect element may have the
following features and/or characteristics:
• The low-pass filter edge frequency and "Q" are both adjustable to
provide either a smooth or "humped" response in the frequency
domain.
20 ' • The compressor raises the average energy of the bass signal by
tracking the amplitude over time. High energy material is limited,
and low energy material is amplified, raising the average energy.
*. The filtered "wet" signal is gain-controlled, then summed with the
original "dry" signal to provide variable control of the block
25 ; output,
FIG. 14 illustrates a functional block level implementation of -subsonic 1 effect element 94S of FIG. 9. In FIG. 14, although a single channel is illustrated. it is to be understood that two such sections may be usee, for a stereo presentation. In the preferred embodiment of the invention, the subsonic effect function 1410 is 30 implemented by combining versions of the channel input signal which have propagated down two paths: (1) a path 1412 with no filtering or compression so that the original channel input sound is preserved. r.nd (2) a path 1414 over which the sound is filtered and compressed, preferably with a low pass filter 1416 and a
39

WO 03/10.1924 PCT/US03/177SS
compressor 141S, respectively. These two signals are preferably summed, as depicted by summing slemcnt 1420. to provide the channel output for the subsonic effect elemen', 1410. It is to be noted that in the summing element 1420. the arrowhead 1422 indicates that the clement may be operated to provided a selectable ratio of the filtered/compressed signal to the uiifiltered/uncompresssd signal, to enhance or reduce the amount of lower frequency components of the channel input signal.
Preferably, the filter characteristics of low pass filter 1416 and of compressor 141S are determined according to the processing requirements of other processing funcioiis described herein which are performed in connection with the operation of ths present invention.
As described in connection with FIG. 9 above, the Look-Ahead AGC element 950 provides a look-ahead automatic gain control function. This feature operates to lower the audio level during peak events and then return it after a peak even! has passed. To keep sound from distorting during the listening process or while recording it, a human engineer would always drop the volume, by moving the volume control down of the offending instrument or vocal. By essentially simulating a human engineer, the Look-Ahead AGC dement 950 rapidly moves the audio level down by analyzing the digital stream ibr distortion and signal overloads to identify peak events. It then returns the volume towards the initial volume setting after ihs peak event has occurred, without the need for an "always-on" audio compressor circuit, which undesirably leads to a loss of dynamic edge and flat sound. In the example embodiment, the signal path may be summarized as follows:
• The signal is applied to a circular delay line.
• The signal is full-wave rectified, and the resultant value 'is measured against the "target" amplitude (the target amplitude represents the maximum signal value for the desired dynamic
range).
• If the rectified signal exceeds the target value, the gain of the
co-;-ol amplifier is decreased by a predetermined "negative ramp'"'
value.
40

WO 03/104924 PCT/US03/177S5
* If the rectified signal is less than the target value, the *ain of the
control amplifier is increased by a predetermined "positive ramp"
value.
• The output signal sample is taken from an earlier position in the
5 delay line and applied to the control amplifier. The amplified
signal becomes the output of the block. FIG, 15 provides a functional block level implementation of the Loolc-Aheacl AGC element 950, While the Look-Ahead AGC element is described at a functional block levsi. one skilled in the art will recognise in light of the detailed
10 description provided herein thai these functions may readily be- implemented in software, hardware, firmware, or any combination therec-f, within the spirit of the invention. Further, although a single channel is presented in FIG. 15, two such sections may be used for a stereo presentation.
In the Look-Ahead AGC implementation 1510 illustrated in FIG. 15, the
15 channel input signal is received at the input of a delay line 153.2. Preferably, delay line 1512 is a digital delay line, and may accommodate one thousand (1000) samples of riie channel input at a sampling frequency o: about 4S kHz, The output of the delay line 1512 is applied to an input of a voltage controlled amplifier 1514. The operation of the voltage controlled amplifier is controlled by a signal level
20 obtained by applying a filtering junction 1516 to the sample from delay line 1512, preferably the sample residing in input element 15IS. Preferably, as the filtered sample level increases, the gain of the voltage controlled amplifier 1534 is decreased, and vice versa, as depicted by the minus (-) sign which labels the control input of voltage controlled amplifier 1514.
25 Preferably, the filtering function 1516 provides a low pass function, and is
represented in HG. 15 by a variable capacitance 1520 in series with a variable resistance 1522 and.which is connected between the output of the first block of delay line 1512 and a reference voltage, such as ground. Thus, frequencies below the cut-off frequency of the low pass function 1516 will have the greatest impact
30 on the gain adjustment of voltage controlled amplifier 1514, while frequencies above the cut-off frequency will have a proportionally reduced effect. As will be understood by these skilled in the art, the settings of the variable capacitance and the variable resistance of filtering function 1516 will affect the frequency
41

WO 03/104924 PCT/US03/177S8
characteristics of the filtering fraction. In operation these settings may be determined according to the processing requiressnts of other processing functions described herein which are performed in connection with die operation of the " present invention.
It is also to be Doted that Look-Ahead AGC element 1510 provide? an inherent time delay it the output end of the signal processing flow.- It has been found for the present invention that implementing a time delay function at this point in the process flow is preferred over the use of time delays in each of the banded channels at the front end of the signal flow. Among the advantages of such a configuration is a buffering feature that allows modification of the waveform before it reaches the listener.
FIG. 16a provides an illustrative example of one implementation of ths Adaptive Dynamics type processing block (labeled core process) in FIG. 9. FIG. 16b is an illustration of the time response characteristics of the sound paths of FIG. 16a.
The input signal is received at the input 1602 to the AI (artificial intelligence) dynamics pre-coiirprcssor. The signal is distributed equally to a fill "ange buffer amp 1612, low pass buffer amp 1611, band pass buffer amp I6l0and a high pass buffer amp 1609.
The full range stream is routed to the full range stream compressor 1601. modified in the time domain with respect to ratio, envelope attack and envelope release and a maximum target level is set. The signal is then routed to a buffer amp 1613 and then to a summing amp 1617.
The low pass range stream is routed to the buffer amp 1611, through the lov/pass filter 1605, to the Jowpass stream compressor 1632, modified in the time domain with respect to ratio, envelope attack and envelope release and a maximum target level is set. The signal is then routed to a buffer amp T614 and then to a summing amp 1617.
The mid or band pass stream is routed to the buffer amp 1610, through ths band pass filter 1606, modified in the time domain with respsct to ratio, envelope attack and envelope release and a maximum target level is set. The signal is then routed io a buffer amp 1615 and then to a summing amp 1617.
The high pass stream is routed to 1-he buffer amp 1609, through the high pass filter 1607, modified in the time domain with respeet to ratio, envelope attack
42

WO 03/104924 PCT/US03/1778S
and envelope release arid a maximum target level is set. The signs! is then routed to 2 buffer amp 1616 and then to a summing amp 16:7.
The addition of the Full, Low, Mid, and High streams simulates live direct
sound impinging on the ear of a live concert listener combined with the low
. 5 frequency dynamics of the room environment (pressure acoustics) combined the
mid range sounds (wave 4- pressure acoustics) and combined with high, frequency
sound (wave acoustics). The sum of these waves creates a combination waveform
in the time domain that can be normalized in the frequency domain to remove
undue frequency non-linearities if desired.
10 The output 1631 of surmnirig amplifier 1617 is routed to the Voice
Enhancer block 934 of FIG. 9.
Included in FIG. 16a are actual parameters for one implementation of the disclosed embodiment. As can be seen from these values, there is a distinct difference in attack, release, gain ratio, and target level used for the compressor 15 blocks in each of the streams. As described above, this difference in parametric settings for the compressor; filter, and gain blocks in each of these streams, is meant to create a time beat or unpredictable character in the processed sound signal.
The attack parameter for the compressor blocks determines how quicklv 20 the path responds to changes in increases in the sound levels. The larger the setting for the attack, the quicker the response. The release parameter controls how much the output of the compressor will lag the fall of a sound signal applied to the input of the compressor. The larger the magnitude of the release setting, the greater the lag. The gain ra'.io is. a dynamic ratio of the envelope of the signal of 25 input versus output up to the target level for the compressor block. It is to be noted that the target level is not used as a threshold, but rather as a maximum number of bits (in the digital signal processing sense) allowed for that compressor output.
The settings for the unnltered, full range stream path [1612-*1601-*1613] 30 are intended to provide a full bandwidth, high SPL simulation which provides a sound that would be expected, from a stage setting without any surrounding environment.
The settings for the low stream path [1611-»1632-*16I4], which handles low frequency sounds, are intended to provide a simulation of sound
43




Referring new to Fie. 16b. the left hand set cf graphs illustrate for each of
5 the different sound paths or streams, the relationship berween the attack, release, target level, and gain ratio. Also, the tirne relationship of the response characteristics as between streams can be seen. Finally, the graph on the right hand side of the sheet illustrates the combined response characteristics of the process. Therefore, from these curves it can be seen that environment dynamics
10 are provided by each of )ow stream, mid stream anc high stream sound paths, and that direct sound dynamics are provided by the fall range stream path.
In this embodiment, the full range stream path provides direct sound reinforcement, the low range stream path provides pressure acoustics reinforcement, the mid range stream path provides both wave and pressure
15 reinforcement, and the high range stream path provides wave reinforcement.
It is to be noted that the graphs for each of these streams illustrates the differences in attack, release, gain ratio and target level between ihe streams as a .function of time. Thus, the envelope for the full range stream has the largest ■aiergy level relative to the indicated base line, find sharper rise and fall times than
20 "he other streams. It is also to bs noted that, relative 10 the pomts in time of U and ".2 for each of the corves, the high stream path concentrates most of its energy in 'ihe middle porlicn of the time period between 11 trA t2. On the other hand, the
45

WO 03/104«J24 PCT/US03/1778S
energy distribution for the low range stream occupies much of the period between. U ar.o 12. and even extends to points before tl and beyond -2.
With continued reference to FIG, 16a, the preferred embodiment include; a "proximity control" feature that allows ths listener to adjust the ratio of the 5 direct sound stage versus the reflected (or otherwise simulated) sound stage. The proximity control feature can be implemented in the example embodiment by providing adjustable access to ths gain ratio element c; the .full range s::ea:ri compressor 1601. As this gain ratio is increased, the output signal received by (he listener will be more direct in nature, with less reflective content. Conversely, &5
10 u.is gain ratio is decreased, the output signal received by ths listener will be Isss cireci in nature, with more reflective content. In practical embodiments, this gain ratio will have a range of O.S to 5.0, with a nominal range of 1.2 to 2.5.
Although preferred embodiments are illustrated in ths accompanying drawings and described in the foregoing detailed description, it will be understood
15 that the inventions arc not limited to the embodiments disclosed, but are capable of numerous rearrangements, modifications and substitutions without departing from the spirit of the inventions as set form and defined by the claims and equivalents thereof.



CLAIMS
1. A method for enhan-ing transmitted i".idio data, comprising:
coding audio data into a digitally formatted signal;
enhancing the digitally formatted signal by pre-emphasizing frequencies
find dynamics expected \o be losi or distorted, resulting in an enhanced audio
.signal;
transmitting the enhanced audio signal to a client site;
decoding data contained in tat enhanced audio signal after transmission to die client site, resulting in a decoded audio signal; and
processing the decoded audio signal to recover frequencies and dynamics •preserved by pre-eraphasis of the frequencies and dynamics expected to be lost or 15 distorted,
2. The method or" claim 1, wherein the frequencies and dynamics
expected to be lost or distorted are attributable at least in part to compression of
the audio signal
20
3. The method of claim i. wherein the frequencies and dynamics
expscted to be lost or distorted are attributable at least in part to transmission of
the audio signal.
25 4. The method of claim 1, further comprising compressing the
enhanced audio signal prior to its transmission, resulting in a compressed enhanced audio signal.
5, The n:eihod of claim 4. further comprising decompressing the
30 . compressed enhanced audio signal subsequent to its transmission.
47

WO03I104W "'. PCT/US03/1T788
6. A method for enhancing audio signals, comprising:
receiving an audio signal;
separating :hs audio signal into component signals corresponding 10
discrete bands;
5 processing one or more o: the component signals with distinct processing
pathways, resulting in processed component signals;
aggregating '.he processed component signals to recreate a standard signal in one or more channels; and
performing additional post-processing en the standard signal to maslc 10 artifacts and response anomalies introduced by a codec and equipment used, resulting in an enhanced audio signal.
7. A method according to claim 6, wherein the audio signal is a
compressed audio signal.
15
8. A method according to claim 6. wherein the separating step
separates the audio signal into at least one full bandwidth component signal and at
least one limited bandwidth component signal.
20 9. A method according to claim 8, wherein the at least one limited
bandwidth component signal comprises at least one of; a bass component signal, a midrange component signal, and a treble component signal.
10. A method according to claim 6, wherein the post-processing 25 comprises at least one of:
3D/live enhancement for adding life and stereo perspective to the sound field of the enhanced audio signal;
recording environment simulation for adding diffusion, reverb, depth,
regeneration, and room decay to the enhanced audio signal;
-0 voice elimination for reducing vocals in the enhanced audio signal;
wide stcrso enhancement for adding wider stereo perspective to the sound field of the enhanced audio signal;
parametric equalisation for providing b.oad spectrum shaping of the enhanced audio signal;
4S

PCT/US03/J778S WO 03/104924
.fiHerins the enhanced audio signal to reinforce sub-woofer, and bass frequencies;
wail simulation for producing time delays that simulate reflections ironi a
stage;
5 room simulation for producing time delays that simulate natural room
acoustics;
karaoke enhancement for removing equal energy components from left
and right signal channels;
vocal enhancement for clarifying vocal features;
10 subsonic enhancement for low-bass reinforcement of Hie enhanced audio
signal; and
look-ahead automatic gain control for controlling output dynamic range.
11. A method according to claim 6, wherein the post-processing
15 includes room simulation for compensating for poor room acoustics in a listening
environment for the enhanced audio signal.
12. A method for compensating for audio equipment operated in a poor
acoustic environment, comprising;
20 obtaining a measured impulse response of a listening environment in
which aiidio equipment is present;
deriving a compensatory process using the ."Measured impulse response; and
compensating for flaws in the listening environment and audio equipment 25 during audio playback by employing the compensatory process,
13. A method according to claim 12. wherein the obtaining step
measures ths impulse response with a microphone located within ths listening
environment.
30
14. A method according to claim 12. wherein the obtaining step
comprises;
producing sound, using the audio equipment, having a known frequency spectrum;
49

PCT/US03; 1*788 WO03/10J92-*
transducing a ie=t signal generated in response to the sound, the test signal
indicating E.coust;cs of the listening envu"or.meru: avid
calcularb j a room transform function based upon the spectrum of the lest
signal and the known frequency spectrum of the sound,
5
15. A system for enhancing audio signal comprising:
a full bandwidth pathway for processing a iUli bandwidth component of an
audio signal the- fall bandwidth pathway producing a processed full bandwidth
signal;
10 at least on± limited bandwidth pathway for processing a limited bandwidth
component of the audio signal, the limited bandwidth pathway producing a
processed limited bandwidth signal;
a mixer configured to combine the processed ful". bandwidth signal and the
processed limited bandwidth signal to create a mixed audio signal; and
15 one or mere post-processing elements for further enhancement of the
mixed audio signal.
16. A system according to claim 15, wherein the at least one limited
bandwidth pathway comprises at least one of;
20 a bass pathway for processing a bass component of the audio signal;
a midrange pathway for processing a midrange component of the audio signal; and
a treble pathway for processing a treble component of the audio signal,
25 17. A system according to claim 15, wherein the one or more post-
processing elements comprises at least one of:
a 3D/live enhancement element configured to add life and stereo perspective to the sound field of the mixed audio signal;
a recording environment simulator configured to add diffusion, reverb, 30 depth, regeneration, and room decay to the mixed audio signal;
a voice elimination element configured to reduce vocals in the mixed audio signal;
a wide stereo enhancement element configured to add wider stereo perspective to the sc-uid Sold of ihe mixed audio sisrtial:

WO05/HU924 PCT/US0J/1T7SJ
a parametric equalizer configured to provide broad spectrum shaping of the mixtd audio signal;
a*, least one filter configured to reinforce subwoofsr and bass frequencies
in the mixed audio signal;
5 a wall simulator configured to produce lime delays that simulate
reflections from a stage;
a room simulator configured to produce time delays that simulate natural room acoustics;
a .karaoke enhancement element configured to remove equal energy 10 ' components from leu and right signal channels;
a vocal enhancement element configured to clarify vocal features;
a subsonic enhancement element configured to reinforce low-bass components of the enhanced audio signal; and
a look-ahead automatic gain control element configured to control output 15 dynamic range.
13. An apparatus for playback of digital audio files, said apparatus comprising:
a digital audio signal source;
20 at leas', one processor coupled (o the digital audio signal source, said at
least one processor being configured to carry out a method comprising:
receiving an audio signal from the digital audio signal source;
separating the audio signal into component signals corresponding
to discrete bands;
25 processing one or more of the component signals with distinct
processing pathways, resulting in processed component signals;
aggregating me processed component signals to recreate a standard signal in one cr more channels; and
performing additional post-processing on the standard signal to 30 mask artifacts and response anomalies introduced by a codec and equipment used, resulting in an enhanced audio signal; and
one or more speaker drivers coupled to the processor, the one or more speaker drivers being configured to drive one or more speakers for playback of the enhanced audio signal.
51

WO03/104M4 .. PCT/US03/I7788
19. A method for enhancing delivery of audio signals, the method
comprising:
modifying an audio signal by creating a thus beat among sound streams in 5 the audio signal, resulting in a modified audio signal; and
inserting reflection components into the modified audio signal.
20. A method according to claim 19,. wherein the modifying stsp
includes the steps of:
10 forming at least a first sound stream and a second sound stream for the
audio signal; and
altering the characteristics of the audio signal m the second sound stream
while maintaining an alignment in time between signals in the first sound stream
and the second sound stream. 15
21. A method according to claim 20, wherein:
the forming step includes the s:sp of compressing the audio signal in the
first sound stream in accordance with a first set of parameters; and
the altering step includes the steps of:
20 filtering the audio signal in the second sound stream to obtain a
filtered audio signal; and
compressing the filtered audio signal according to a second set of parameters different from the first set of parameters.
25 22. A method according to claim 21, further including the step of
selecting the second set of parameters to provide an altered audio signal from the second sound stream which has the characteristics of s-ounds emanating" from a selected environment.
30 23. A method according to claim 19, wherein the inserting step
includes inserting a; least one of early, mid, and late reflection comporunts.
24. A method according to claim 19, wherein the modifying step includes ths seeps of:
52

WO 03/10492-1 PCT/T5Q3/177S8
forming a full ringc sound stream, a Low range sound stream, a mid range sound stream; and a high range sound stream;
low-pass filtering the audio signal in the low range sound stream to obtain
alow-pass filtered audio signal;
5 compressing the low-pass filtered audio signal according to a "dead**
environment set of parameters;
band-pass filtering the audio signal in the mid range sound stream to obtain a band-pass filtered audio signal;
compressing the band-pass filtered audio signal according to a "scoring" 10 stage environment set of parameters;
high-pass filtering the audio signal in the high range sound stream to obtain a high-pass filtered audio signal;
compressing the high-pass filtered audio signal according to a "plaster wail" environment set of parameters. 15
25. A method of enhancing delivery of audio signals to a listener, the
method comprising:
modifying an audio signal by creating a difference in the dynamics, of
sound streams in the audio signal, resulting in a modified audio signal;
20 adding predictable environmental characteristics to the modified audio
signal to form an enhanced audio signal; and
delivering the enhanced audio signal to the Iistensr. .- „.,
26. '"'A method according to claim 25, further including the step of *-".."
25 inserting sound field enhancing features into the enhanced audio signal to provide
alternative listening sound field controllability lo che listener.
27. A system for enhancing audio signals, comprising:
a full bandwidth pathway for processing £ full bandwidth component of an 30 audio signal \he full bandwidth pathway producing a processed full bandwidth signal, ths full bandwidth pathway comprising:
a first input amplifier having an input for the audio signal, a first ouiput amplifier having an output for the processed full bandwidth signal, and e
53

WO03/10JK4 PCT/US0J/J77W
first compressor connected between the fust input amplifier and the first output amplifier;
at least one limited bandwidth pathway for processing a limited bandwidth component of the audio signal, the limited bandwidth pathway producing a 5 processed limited bandwidth signal, the at least one limited bandwidth pathway comprising:
n second input amplifier having an input for the audio signal, a
second output amplifier having an ouiput for the processed limited bandwidth
signal, a second compressor connected between the second input amplifier and the
1.0 second output amplifier, and a filter connected between '!he second input amplifier
and the second output amplifier; and
a mixer corc'.gured to combine the processed full bandwidth signal and the processed limited bandwidth signal to create 2 mixed audio signal.
15 28. A system according to claim 27, further comprising one or more
post-processing elements for further enhancement of the mixed audio signal.
29. A system according to claim 27, wherein at least one of the first
input amplifier, the first output amplifier, the second input amplifier, and the
20 second ouiput amplifier is a variable gain amplifier.
30. A system according lo claim 27, wherein the at least one limited
bandwidth pathway comprises at least one of;
a bass pathway for processing a bass component of die audio signal;
25 a inidrange pathway for processing a midrange component of the audio
signal; and
a treble pathway for processing a treble component of the audio signal.
31. A system according to claim 30. wherein;
30 for the bass pathway, the filter is a law-pass niter:
for the inidrange pathway, the filter is a band-pass filter: and for the trebls pathway, the filter is a high-pass filler.
54

I
WO 03/10492-. "'. PCT/US03/1778S
32. A system according to claim 27, farther comprising a pre-comprsssor configured to receive an input audio signal and to generate ths audio signal as a compressed representation of the input audio signal.
5 33. An apparatus fen playback of digital audio files, said appavans
comprising;
a digital audio signal source;
at least or.; processor.coupled to ihe digit-:! audio signal source, said at
least one processor being configured to cany out a method comprising:
10 modifying an audio signal by creating s. time beat among sound
streams in the audio signal, resulting in a modified audio signal; and
inserting reflection components into the modified audio signal; and one or mors speaker drivers coupled to the processor, the one or more speaker drivers being configured to drive one or mors speakers for playback of the ] 5 enhanced audio signal.
34. An apparatus for playback of digital audio files, said apparatus comprising:
a digital audio signal source;
20 at least one processor coupled to the digital audio signal source, said at
least one processor being configured to carry out a method comprising:
modifying an audio signal by creating a difference in the dynamics of sound streams in hhc audio signal, resulting in a modified audio signal;
adding predictable environmental characteristics to the modified 25 audio signal to form an enhanced audio signal; and
inserting sound field enhancing features into the enhanced audio signal to provide alternative listening sound field controllability to the listener, and
one or more speaker drivers coupled to the processor, the one or more 30 speaker drivers be:'ng configured to drive one or more speakers for delivery of the enhanced audio signs! to the listener.
55

o. A method tor enhancing delivery substantially as herein described with reference to the accompanying drawings.


Documents:

2752-CHENP-2004 AMENDED PAGES OF SPECIFICATION 05-04-2011.pdf

2752-CHENP-2004 AMENDED CLAIMS 05-04-2011.pdf

2752-chenp-2004 assignment 04-07-2011.pdf

2752-CHENP-2004 CORRESPONDENCE OTHERS 04-07-2011.pdf

2752-chenp-2004 form-1 04-07-2011.pdf

2752-chenp-2004 form-3 04-07-2011.pdf

2752-CHENP-2004 OTHER PATENT DOCUMENT 1 04-07-2011.pdf

2752-chenp-2004 other patent document 2 04-07-2011.pdf

2752-chenp-2004 amended pages of specification 04-07-2011.pdf

2752-chenp-2004 amended claims 04-07-2011.pdf

2752-CHENP-2004 CORRESPONDENCE OTHERS 12-08-2010.pdf

2752-CHENP-2004 EXAMINATION REPORT REPLY RECEIVED 05-04-2011.pdf

2752-chenp-2004 form-1 05-04-2011.pdf

2752-CHENP-2004 POWER OF ATTORNEY 05-04-2011.pdf

2752-CHENP-2004 FORM-3.pdf

2752-chenp-2004 assignment.pdf

2752-chenp-2004 claims.pdf

2752-chenp-2004 correspondence others.pdf

2752-chenp-2004 correspondence po.pdf

2752-chenp-2004 description(complete).pdf

2752-chenp-2004 form-1.pdf

2752-chenp-2004 form-26.pdf

2752-chenp-2004 form-3.pdf

2752-chenp-2004 form-5.pdf

2752-chenp-2004 form-6.pdf

2752-chenp-2004 pct.pdf


Patent Number 248369
Indian Patent Application Number 2752/CHENP/2004
PG Journal Number 28/2011
Publication Date 15-Jul-2011
Grant Date 07-Jul-2011
Date of Filing 06-Dec-2004
Name of Patentee ARC INTERNATIONAL PLC
Applicant Address VERULAM POINT, STATION WAY, SAINT ALBANS, UNITED KINGDOM AL1 5HE
Inventors:
# Inventor's Name Inventor's Address
1 PADDOCK, THOMAS 12153 NUTHATCH COURT, TRUCKEE, CA 96161
2 BARBER, JAMES 681 WENAS VIEW DRIVE, SCLAH, WA 98942,
PCT International Classification Number G10K15/00
PCT International Application Number PCT/US03/17788
PCT International Filing date 2003-06-05
PCT Conventions:
# PCT Application Number Date of Convention Priority Country
1 60/386,541 2002-06-05 U.S.A.
2 60/472,180 2003-05-20 U.S.A.