Audio Coding

La tonalité est un attribut perceptif lié aux composantes tonales (émergences spectrales) présentes dans de nombreux types de bruit (avions, systèmes de climatisation, voiture électrique, ...). Plusieurs facteurs influencent la tonalité perçue (fréquence du ton, niveau d'émergence du ton, etc.). La largeur spectrale du bruit de fond est l'un de ces facteurs. Son impact sur la perception de la tonalité n'est pas nettement caractérisé dans la littérature. S'appuyant sur la théorie du masquage, certains indicateurs de tonalité ne prennent en compte qu'une largeur de bande critique de bruit autour du ton émergent. D'un point de vue plus pratique, il existe différentes méthodes de test d'écoute pour étudier l'impact d'un facteur sur la tonalité perçue (évaluation directe, méthode adaptative, …). La durée d'un test ainsi que la précision des résultats perceptifs dépendent en partie de la méthode choisie. Dans le cadre de cette étude, 18 sons ayant des largeurs spectrales de bruit de fond différentes ont été évalués par plusieurs sujets en utilisant trois méthodes de tests d'écoute : une méthode d'évaluation directe, une méthode adaptative à choix forcé (2AFC 1down/1up) et une méthode d'égalisation à iso-tonalité La méthode d'égalisation à iso-tonalité apporte le meilleur compromis entre la précision des résultats et la durée du test d'écoute. Les résultats perceptifs suggèrent que le bruit contenu en dehors de la bande critique autour du ton devrait être considéré lors de l'évaluation de la tonalité.

In this paper we investigate a special-purpose application of MPEG-1 layer II audio streaming. First, we discuss how two or more already coded MPEG-audio bitstreams can be manipulated and mixed within the coded subband domain by using an appropriate algorithm for combining and recalculating the bit allocation. Next, we will show that very significant speedups can be obtained compared with mixing in the temporal domain. We will also illustrate that, at the same time, the system can act as a bit rate scaling device, using one of the bit allocation-combination algorithms which we implemented. The results obtained with these different algorithms are briefly discussed and compared. Finally, we report on how the developed software was integrated into an internet audio streaming system, which is now able to allow simple, yet efficient, real-time mixing of several MPEG-audio coded signals.

A multichannel extension to the RVQGAN neural coding method is proposed, and realized for data-driven compression of third-order Ambisonics audio. The input-and output layers of the generator and discriminator models are modified to accept multichannel input. We also propose a loss function which accounts for spatial perception in immersive reproduction, and transfer learning from single-channel models. Listening test results with 7.1.4 immersive playback show that the proposed extension is suitable for coding ambient scene-based, third-order (16-channel) Ambisonics content with good quality at 16 kbps when trained and tested on the EigenScape database. The model presented is the first neural codec dedicated to immersive audio to the authors' knowledge and has potential applications for learning other types of content and multichannel formats.

Recently lifting-based integer transforms have received much attention, especially in the area of lossless audio and image coding. The usual approach is to apply the lifting scheme to each Givens rotation. Especially in the case of long transform sizes in audio coding applications, this leads to a considerable approximation error in the frequency domain. This paper presents a multidimensional lifting approach for reducing this approximation error. In this approach, large parts of the transform are calculated without rounding operations, only the output is rounded and added. The new approach is applied and evaluated for both the Integer Modified Discrete Cosine Transform (IntMDCT) and the Integer Fast Fourier Transform (IntFFT).

Proper coding and transmission of medical and physiological data is a crucial issue for the effective deployment and performance of telemedicine services. This chapter presents a platform for performing proper medical content adaptation based on context awareness. Sensors are used in order to determine the status of a patient being monitored through a medical network. Additional contextual information regarding the patient's environment (e.g., location, data transmission device and underlying network conditions, etc.) is represented through an ontological knowledge base model. Rule-based evaluation determines proper content (i.e., biosignals, medical video and audio) coding and transmission of medical data, in order to optimize the telemedicine process. The paper discusses the design of the ontological model and provides an initial assessment.

Substantial progress has been made recently in finding acoustic features that describe perceptually relevant aspects of sound. This paper presents a general framework for synthesizing audio manifesting arbitrary sets of quantifiable acoustic features of the sort used in music information retrieval and other sound analysis applications. The methods described have broad applications to the synthesis of novel musical timbres, processing by analysisresynthesis, efficient audio coding, generation of psychoacoustic stimuli, and music information retrieval research.

We present a general framework for synthesizing audio manifesting arbitrary sets of perceptually motivated, quantifiable acoustic features. Much work has been done recently on finding acoustic features that describe perceptually relevant aspects of sound. The ability to synthesize sounds defined by arbitrary feature values would allow perception researchers to more directly generate stimuli "to order," as well as providing an opportunity to directly test the perceptual relevance and characteristics of such features. The methods we describe also provide a straightforward way of approaching the problem of mapping from data to synthesis control parameters for sonification.

Codage audio perceptuel a bas d?ebit par D?ecomposition Modale Empirique (EMD) Kais KHALDI1,3, Abdel-Ouahab BOUDRAA2, Monia TURKI1, Thierry CHONAVEL3 1Unit?e Signaux et Syst emes, ENIT BP 37, Le Belved ere 1002 Tunis, Tunisie 2IRENav, Ecole Navale, ...

This paper describes the use of sorted sinusoidal parameters to produce a fixed rate, scalable, wideband audio coder. The sorting technique relies on the perceptual significance of the sinusoidal parameters. Sinusoidal coding permits the representation of a given signal through the summation of sinusoids. The parameters of the sinusoids (the amplitudes, phases and frequencies) are transmitted to allow signal reconstruction. The sinusoidal parameters are sorted according to energy content and perceptual significance. The most significant parameters are transmitted first, allowing the use of only a small set of the parameters for signal reconstruction. The proposed scheme incurs a low delay and uses a 20 ms frame length. The results presented show the advantages gained for scalable audio coding by sorting the parameters.

La problématique abordée dans cet article concerne le développement et l'optimisation d'une métrique de qualité pour les images compressées. Nous utilisons une méthode d'optimisation basée sur l'utilisation conjointe d'une échelle perceptive des différences et d'un algorithme génétique pour trouver les paramètres de la fonction de qualité développée. Les résultats obtenus montrent que les valeurs optimisées des paramètres induisent une augmentation significative de la corrélation avec les valeurs objectives obtenues par les observateurs humains.

Bark-scale warped linear prediction [WLP] is a very potential core for a monophonic perceptual audio codec [2]. In the current paper the WLP scheme is extended for processing complex valued signals (CWLP). Three different methods of converting a stereo signal to one complex valued signal are introduced. The philosophy behind the coding scheme is to integrate some aspects of modern wideband audio coding (e.g. perceptuality and stereo signal processing) into one computational element in order to find a more holistic and economic way of processing.

An inherent property of many DSP algorithms is that they tend to exhibit uniform frequency resolution from zero to Nyquist frequency. This is a direct consequence of using unit delays as building blocks; a frequency independent delay implies uniform frequency resolution. In audio applications, however, this is often an undesirable feature since the response properties are typically specified and measured on a logarithmic scale, following the behavior of the human auditory system. In this paper we present an overview of warped filters and DSP techniques which can be designed to better match the audio and auditory criteria. Audio applications, including modeling of auditory and musical phenomena, equalization techniques, aurahation, and audio coding, will be presented.

Frequency-warped signal processing techniques are attractive to many wideband speech and audio applications since they have a clear connection to the frequency resolution of human hearing. A warped version of linear predictive coding (LPC) is studied in this paper. The performance of conventional and warped LPC algorithms are compared in a simulated coding system using listening tests and conventional technical measures. The results indicate that the use of warped techniques is beneficial especially in wideband coding and may result in savings of one bit per sample compared to the conventional algorithm while retaining the same subjective quality.

MPEG-1 Layer 3 (MP3) is one of the most popular compression formats used for sound and especially for music. However, during the coding process, the MP3 algorithm negatively affects the spectral and dynamic characteristics of the audio file being compressed. The aim of this study is to evaluate the effect the MP3 coding format for different music genres and different bitrates via listening tests in which the original uncompressed files and the compressed files are compared. For this purpose, five different music genres were selected (rock, jazz, electronic, classical and solo instrument), and the files were compressed in three different bitrates (96 kbps, 160 kbps and 320 kbps). The semantic differential method was used, and ten bipolar scales were selected for the listening tests (e.g., better-worse, more distortion-less distortion, etc.). The following are the most important findings of this study: classical music was negatively affected the most among the genres due to the MP3 compression (lowest ratings in 8 out of 10 bipolar scales), the solo instrument was least affected among the genres (highest rating in 7 out of 10 bipolar scales), and for higher bit rates, the differences in ratings were small for all music genres. The findings of this study could be used to optimize and adapt the standard, depending on the music genre and the musical piece that needs to be encoded.

The papers at this Convention have been selected on the basis of a submitted abstract and extended precis that have been peer reviewed by at least two qualified anonymous reviewers. This convention paper has been reproduced from the author's advance manuscript, without editing, corrections, or consideration by the Review Board. The AES takes no responsibility for the contents.

La mise en service et l'arrêt de certains circuits audionumériques, tels que ceux intégrés dans les téléphones portables multimédia, entraîne la génération de bruits impulsionnels (« pop ») qui peuvent s'avérer gênants pour l'utilisateur, même à faible niveau. Ce travail est consacré à l'étude d'un indicateur objectif pour prédire l'audibilité de ce type de bruit en fonction de ses caractéristiques électriques et du casque d'écoute utilisé. Pour cela, un test psychoacoustique est réalisé sur un panel de 30 sujets, portant sur 10 signaux impulsionnels différents, diffusés aux sujets au moyen de 4 modèles de casque d'écoute. En appliquant une procédure adaptative du type « 1up -2 down », ce test a permis de déterminer le gain à appliquer pour que ces signaux soient justes audibles (probabilité de détection de 71%). Différents indicateurs objectifs (L A,eq,Tms,max , N 30ms , STL max , LTL max , LMIS, PEAQ) sont ensuite calculés à partir des signaux électriques et des fonctions de transfert électroacoustiques mesurées sur les casques d'écoute, pour déterminer celui qui est le mieux corrélé avec les résultats du test.

This paper deals with pre-echo reduction in low bit-rate audio compression. [1] proposed an attack restoration method based on the correction of the temporal envelop of the decoded signal. A small set of coefficients were then transmitted through a limited bit-rate auxiliary channel. However, the transmission of the transient position computed on the original audio signal was required. In this paper, we deployed a new method of attack localization based on differential algebraic, which guaranties a successful detection on the decoded audio signal. The algebraic method has also a reduced complexity compared to the index stationary detector used in [1]. The new proposed approach is evaluated for single audio coding-decoding, using objective perceptual measures. The experimental results for MP3 coding exhibits an efficient restoration of the attacks and a significant improvement of the audio quality.

This paper studies the quality of multimedia content at very low bitrates. We carried out subjective experiments for assessing audiovisual, audio-only, and video-only quality. We selected content and encoding parameters that are typical of mobile applications. Our focus were the MPEG-4 AVC (a.k.a. H.264) and AAC coding standards. Based on these data, we first analyze the influence of video and audio coding parameters on quality. We investigate the optimal trade-off between bits allocated to audio and to video under global bitrate constraints. Finally, we explore models for the interactions between audio and video in terms of perceived audiovisual quality.

We carried out a number of subjective experiments for audiovisual, audio-only, and video-only quality assessment. We selected content and encoding parameters at very low bitrates that are typical of mobile applications. Using these data, we explore the influence of video codecs and frame rate as well as audio channels and sampling rate on quality. Finally, the optimal trade-off between bits allocated to audio and video inside a bitstream is investigated.

New approaches to hybrid in band on channel (HIBOC) FM systems for digital audio broadcasting based on multistream transmission methodology and multidescriptive audio coding techniques are introduced in this paper. These ideas combined with a lower per sideband audio coding rate and more powerful channel codes result in robust transmission and graceful degradation in variable interference channels. By also using orthogonal frequency division multiplexing techniques with a nonuniform power profile combined with unequal error protection and sideband time diversity, we arrive at new HIBOC FM schemes with extended coverage and better peak audio quality than previously proposed. The paper provides approximate performance analysis for potential systems including audio coding quality.

Emerging digital audio applications for broadcast radio and multimedia systems are presenting new challenges such as the need to code mixed audio content, error robustness, higher audio bandwidth and quality at low bit rates; demanding a paradigm shift in the existing speech coding techniques. This paper describes continuation of our research in low bit rate audio/speech coding and in the recently introduced Audio Bandwidth Extension Toolkit (ABET). Several new modes of operation have been introduced in the codec, in particular making innovative use of perceptual coding tools. Additionally, a new mode in ABET is added to improve the efficiency of the temporal shaping tool, Multi Band Temporal Amplitude Coding (MBTAC), by exploiting time and frequency correlation. The structure of the codec and its performance in these modes of operation are detailed. Further information is available at

In this paper we describe the components of a novel audio coding algorithm capable of delivering high-fidelity CDlike stereo audio at the bit rates of 40-48 kbps and natural sounding FM grade mono at the bit rates of 18-22 kbps. Bandwidth Extension has emerged as an important tool for the satisfactory performance of low bit rate audio codecs. Recently we proposed two new bandwidth extension algorithms, Fractal Self-Similarity Model (FSSM) and Accurate Spectral Replacement (ASR), which belong to a new class of Bandwidth Extension techniques which are applied directly to the high resolution frequency representation of the signal (e.g., MDCT or ODFT). The proposed coding scheme uses FSSM and ASR in an adaptive and complementary framework. Another important component of the proposed codec is a wideband psychoacoustic model that makes an explicit use of the Comodulation Release of Masking (CMR) phenomenon. It also includes a novel parametric stereo coding technique. The proposed audio coding scheme is geared towards broadcast applications where codec latency and encoder complexity is generally not an overriding concern. In this paper we present algorithmic details of the new codec, audio demonstrations, and, comparison to other audio coding schemes. Further information and audio demonstrations are available at .

In the application of conventional audio compression algorithms to low bit rate audio coding one is faced with the unsatisfactory tradeoff between coarser quantization and audio bandwidth reduction. Frequency Extension has therefore emerged as an important tool for the satisfactory performance of low bit rate audio codecs. In this paper we describe one of a newer class of Frequency Extension techniques which are applied directly to the high frequency resolution representation of the signal (e.g., MDCT). This particular technique is based on a Fractal Self-Similarity Model (FSSM) for the short-term frequency representation of the signal. The FSSM model, which may include multiple dilation and translation terms, has been found to be effective for a wide variety of speech and music signals and provides a compact description for long term correlation that may exist in frequency domain. The high frequency resolution of MDCT aids in accurate parameter estimation for the model, which in turn has shown promise as a Frequency Extension tool that offers a detailed and natural sounding quality at low bit rates. Structure of the FSSM model, issues related to parameter estimation, and its application to audio coding for bit rates of 8-48 kbps is discussed. Audio demos are available at . 1 Support for this work was provided by Sirius Satellite Radio (see Acknowledgements for details).

R sum L'algorithme OLA Overlapp and Add est classiquement utilis pour le ltrage adaptatif dans le domaine des fr quences. Cette m thode introduit cependant un retard, qui peut nuire son utilisation dans certaines applications temps r el. Dans cet article, nous proposons de r organiser le calcul de la Transformation de Fourier Rapide permettant de r duire ce retard. Nous indiquons aussi une solution pour parall liser la succession de deux op rations Transformation de Fourier Rapide et Transformation inverse. L' valuation du gain fourni par ces modi cations nouvelles est aussi donn e.

This paper outlines an adaptive wavelet-based perceptual audio coding scheme attending to various entropy-type criteria. Its performance using some different wavelet families and various filter lengths and decomposition depths has also been investigated. An optimal choice of these parameters is accomplished in order to evaluate both quality and bit rate of compressed signals for four different entropy-type criteria and four representative samples of audio material. The proposed coding scheme performs a periodized wavelet packet transform for each audio frame leading to a decomposition tree which is adapted to the characteristics of the audio frame attending to some entropy criterion. After time-frequency mapping, a thresholding to zero step is carried out to take advantage of entropy coding methods. Next, an uniform quantifier controlled by a psychoacoustic model taking advantage of the masking effect in human hearing is used. Finally, statistical redundancies of audio signals are r...

In a previous work we presented a low delay audio coder [1]. This coder responded to a Subband-ADPCM hybrid structure and required 2'5 bits/sample. In this paper we present a new low delay audio coder that demands only 2 bit/sample and responds to the same structure Subband-ADPCM but incorporating adaptive vector quantization [2]. In such sense, each component of the vector (prediction error) is weighted according to its variance estimation. The total coding delay (< 2 ms) is due only to the perfect reconstruction filter bank [3]. Our objective is to find an audio coding scheme that allows to achieve nearly-transparent quality with low delay. In applications where the delay is a critical parameter, i.e. when is important to use a feedback channel, there is not a standard well suited for this case. The profusely used standards ISO-MPEG for audio handle perceptual models that require high frequency precision and introduces high delay [4].

This paper deals with the application of adaptive signal models for parametric audio coding. The matching pursuit algorithm is used for extracting sinusoidal components and transients in audio signals. The resulting residue is perceptually modelled as a noise like signal. First of all, an accurate detection of transients in audio signals is required. When a transient is detected, psychoacoustic-adapted matching pursuits are accomplished using a wavelet-based dictionary followed of an harmonic one. Otherwise, matching pursuit is applied only to the harmonic dictionary. This multi-part model (Sines + Transients + Noise) is successfully applied for audio coding purposes, assuring high perceptual audio quality at low bit rates (lower than 24 kbps for most of the CD-quality one channel audio signals considered for testing).

Nowadays, it does not exist an audio coding standard for getting nearly-transparent quality with low delay. The standard ISO-MPEG is profusely used in audio for getting high quality [1]. It uses a perceptual model that requires high frequency precision and introduces high delay. However, in applications where the delay is a critical parameter, e.g. when is important to use a feedback channel, the ISO-MPEG is not adequate and there is not a standard for this case. This paper describes the different steps followed for the obtainment of an audio coder with nearly-transparent quality and low delay. The complexity was another design criterion because the final objective is the real time implementation of the proposed coder on a commercial DSP. A comparative analysis of low delay codification schemes is done (ADPCM, Subband, and Subband-ADPCM coding). For each technique we modify its different parameters (e.g. numbers of bits of the quantizer, number of bands, bit allocations criterion, aliasing filters, etc) in order to get the best configuration. The proposed coder uses a Subband-ADPCM hybrid structure with dynamic bit allocation. In such sense, we propose a very low complexity bit allocation strategy with similar results to the classic strategy [2][3] and no delay. The coder output is scalable from high quality at higher bit-rate to lower quality at lower bit rates, supporting a wide range of service and resource utilization. Nearly-transparent quality is accomplished with 2.5 bits/sample. The less than 2 ms algorithm delay is due only to the perfect reconstruction filter bank [4].

ITU-R BS.1387 states a method for objective assessment of perceived audio quality. This Recommendation, known also as PEAQ (Perceptual Evaluation of Audio Quality) is based on a psychoacoustic model of the human ear and was standardized by the International Telecommunications Union as an alternative to subjective tests, which are expensive and timeconsuming processes. PEAQ combines various physiological and psycho-acoustical properties of the human ear to give a measure of the quality difference between a reference audio and a test audio. The reference audio signal could be considered as a distortion-free source, whereas the test signal is a distorted version of the reference, which may have audible artifacts because of compression. The algorithm computes the Model Output Variables (MOVs) which are mapped to a single quality measure, Objective Difference Grade (ODG), using a three-layer perceptron artificial neural network. The ODG estimates the perceived distortion between both audio signals. In this paper we propose a new metric of low computational complexity called FQI (Fuzzy Quality Index) which is based on Fuzzy Logic reasoning and has been incorporated into the existing PEAQ model to improve its overall performance. Results show that the modified version slightly outperforms PEAQ.

The very great volume of information and data in a digital image can cause practical problems. Transmitting an image from one computer to another and/or archiving are very expensive due to the abundance of data representing the image in the form of bits. We present in this article a compression method, which takes in account the coding hybrid of huffman and shanon Fano, which has been applied to imagery and data. We've determined the method limit by considering different forms of images histogram. The results showed that the method is efficient when the number of bit by pixel is of the same order of magnitude as entropy.

Computational methods for the perceptual evaluation of lossy audio processing systems are required, if subjective listening tests are either too expensive or not applicable (e.g. for real-time monitoring). One approach to assess the transmission quality of such systems adequately to human perception is to estimate the perceived audio quality of the systems output signal relative to the input signal. According to [1], the basic audio quality of an audio object is a single, global attribute to judge any and all differences between the reference and the object. In this sense, the more similar the input (= reference) and output signal are perceived, the better the audio quality of the output signal and therefore the transmission quality of the system is. According to this understanding of audio quality, it is not an absolute quality measure, but a quality difference measure. A common concept to predict the perceived audio quality (difference) by computational means is to compare simulat...

This paper illustrates the suitability of ADSP 21160 for implementation of high performance DSP applications requiring 32 bit floating point precision. We present a real-time implementation of a multi-channel MPEG-2 AAC-LC encoder. The performance results of this implementation are also discussed in this paper.

Audio Lossless Coding (ALS) is a new addition to the suite of MPEG-4 audio coding standards. The ALS codec is based on forward-adaptive linear prediction, which offers remarkable compression even with low predictor orders. Nevertheless, performance can be significantly improved by using higher predictor orders, more efficient quantization and encoding of the predictor coefficients, and adaptive block length switching. The paper describes the basic elements of the ALS codec with a focus on these recent improvements. It also presents the latest developments in the standardization process and describes several important applications of this new lossless audio format in practice.

The spectrum interpolation synthesis model has recently been applied in the high quality synthesis of harmonic musical sounds. In this work we investigate the performance of the model in the compression of music signals. Efficient methods for the automatic analysis, parameter extraction and synthesis of musical signals are presented. The system is tested on several examples of segments from wind and bowed string instruments. It is found that typically a perceived quality matching the original is obtained even when large portions of the waveform are generated by interpolation, implying that a high degree of compression is possible. Further, there is a graceful degradation in quality as the extent of interpolation is increased which makes the model well suited for use in a scalable audio coding framework.

This paper illustrates the suitability of ADSP 21160 for implementation of high performance DSP applications requiring 32 bit floating point precision. We present a real-time implementation of a multi-channel MPEG-2 AAC-LC encoder. The performance results of this implementation are also discussed in this paper.

This paper illustrates the suitability of ADSP 21160 for implementation of high performance DSP applications requiring 32 bit floating point precision. We present a real-time implementation of a multi-channel MPEG-2 AAC-LC encoder. The performance results of this implementation are also discussed in this paper.

This paper proposes a modification of ADPCM algorithm to be suitable for hardware implementation of portable audio devices. This is accomplished by a simplification of predictor coefficient algorithm which normally requires a high complexity correlation algorithm, while the sound quality measured as SNR remains relatively the same to those of the original ADPCM. The less complicated calculation results in a reduction of the chip area and power consumption, thus it is suitable for pocket solid-state speech and audio recorder. Simulations on MATLAB verify the performances and experimental results on XILINX FPGA synthesis confirming the functionalities and significant reduction of device count by 63% of the implementation by the original algorithm.

This paper describes a novel technique for audio coding, a lossy compression algorithm, that considers perceptual and rate-distortion criteria. It is based on matched finite impulse response (FIR) wavelet-packet-like filter banks, the filter coefficients being produced adaptively according to the input signal. This technique achieves perceptually transparent coding of high-quality audio, signals sampled at 44.1 KHz-16 bits PCM, at bit rates of about 54-64 Kbps. The matched filter-bank makes a time-frequencyshape analysis, reducing the number of sub-bands requiring quantization. The decoder that implements this algorithm works effectively in real time. This reassures the efficacy of our technique.

With the proliferation of broadband access and continuous decline of storage prize per gigabyte, there has been an increasing demand of audio solution that provides high sampling rate and high resolution. Lossless audio is undoubtedly the ultimate solution. In response to this demand, MPEG issued a call for proposal soliciting technology contributions that provides a state-of-art solution. At the technology end, lossless compression requires the usage of integer transform. The Integer Modified Discrete Cosine Transform (IntMDCT) has been adopted in MPEG-4 Scalable to Lossless (SLS) coding to enable this efficient lossless operation. Because of rounding operations, rounding errors introduced by IntMDCT exist during the whole coding process. With the SLS having capability of using operations that spreads over the bitrate spectrum which ranges from lossy to lossless, it is of interest to study the effect of rounding errors in IntMDCT for operation of SLS in lossy mode. This paper analyzes the contributions of noise due to these errors. It is found that the noise introduced by rounding operations of IntMDCT does not affect the perceptual quality of the coded audio under any circumstances. As such, it concludes that the MDCT and IntMDCT filterbanks are interchangeable at lossy bitrate. With the fact that SLS uses both MDCT and IntMDCT, the finding in this paper suggests the possibility of using only IntMDCT filterbank.

Wide bitrate range scalability is now the latest trend in audio coding. A lot of efforts has been devoted to the development of algorithms for more efficient scalable audio coder that scales from very low bitrate. Scalable audio coding technique such as MPEG-4 Scalable Lossless coding (SLS) offers a unified solution for high-compression perceptual audio and high-quality lossless audio. SLS provides a fine-grain scalable extension of the well-known MPEG-4 Advanced Audio Coding (AAC) perceptual audio coder up to fully lossless reconstruction. Recently, the combination of SLS and AAC coder is renamed as "High Definition Advanced Audio Coding" (HD-AAC). It is observed that HD-AAC can be further improved at intermediate enhancement bitrate when the core bitrate is low. In this paper, a Perceptually Prioritized Bit-Plane Coding (PPBPC) is proposed. With this novel coding scheme, the bit-plane coding is performed with priorities according to the perceptual information of the signal to be coded. By using this low-complexity structure with trivial extra side information, the bit-plane coding for scalable audio can be implemented in a perceptually more efficient manner and the quality of the audio under aforementioned scenario is greatly improved.

A simple and flexible bit-plane coding method is developed for scalable audio coding. It is different from the traditional bit-plane coding in that the optimal bit-plane scanning order is adapted to the scale between the energy of the residual signal and the audio mask. Both the perceptual quality and the lossless compression ratio at common lossy bitrates are greatly improved compared with the performance of state-of-the-art scalable audio.

This paper presents a double improvements to reduce the speech coding complexity of the pitch prediction in a Code-Excited Linear Prediction (CELP) coder. First, the pitch analysis structure is modified. A new fast Pitch Modelling by linear-Filtering (P MF) procedure will determine the adaptive and stochastic codebook contributions of the excitation signal. Afterwards, an e cient implementation of the P MF CELP coding is proposed by using Number Theoretic Transforms which can significantly reduce the algorithm computation complexity.

We propose an approach to digital audio effects using recombinant spatialization for signal processing. This technique, which we call Spatio-Operational Spectral Synthesis (SOS), relies on recent theories of auditory perception. The perceptual spatial phenomenon of objecthood is explored as an expressive musical tool.

Speech coding is a very mature research area and many coding schemes are available that provide speech qualities ranging from highly intelligible synthetic speech at about 2 kbit/s, till wideband natural speech at about 16 kbit/s. However, emerging application scenarios such as information services on broadcast radio are eliciting additional concurrent challenges not easily addressed by current speech coding technology, namely the need to code mixed audio material, the need to permit flexible bitrate coding configurations, the need to scale effectively in quality in the range 2-8 kbit/s, and the need to offer pleasant natural sound. In this paper we present a new very low rate speech/audio coding technology addressing those concurrent challenges thanks to the use of innovative approaches regarding accurate reconstruction of harmonic complexes, optimal coding of the excitation, efficient side information coding, and suitable combination of new bandwidth extension techniques. The structure of the speech/audio coder is detailed and its performance in the range 2.4-12 kbit/s is illustrated and compared to that of reference coders.

In this paper, we propose an improved sinusoidal audio modeling method for perceptual matching pursuits driven by a perceptual distortion measure. A linear model derived from the effective signal processing in the ear is used for computing the perceptual distortion measure. This distortion measure allows us to select the most perceptually meaningful sinusoid at each iteration of the pursuit. Furthermore, the distortion measure can be used to define a psychoacoustic stopping criterion for the matching pursuit algorithm. The proposed sinusoidal modeling method is designed to be used in sinusoidal audio coding. Our method provides significant advantages regarding previous works because it achieves a better separation between tones and noise, as can be seen in results.

A new audio transform coding technique is proposed that reduces the bitrate requirements of the Perceptual Transform Audio Coders, by utilizing the stationarity characteristics of the audio signals. The method detects the frames which have significant audible content and codes them in a way similar to conventional Perceptual Transform Coders. However, when successive data frames are found to be similar to those sections, then their audible differences are only coded. An error analysis for the proposed method is presented and results from tests on different types of audio material are listed, indicating that an average of 30% in compression gain (over the conventional Perceptual Audio Coders bitrate) can be achieved, with a small deterioration in the audio quality of the coded signal. The proposed method has the advantage of easy adaptation within the Perceptual Transform Coders architecture and add only small computational overhead to these systems. n recent years the introduction of digital audio as a method for storing, processing and transmitting high-fidelity acoustic signals has helped in the evolution of numerous applications in the field of consumer electronics and professional audio. It is also envisaged that, in the near future, significant new techniques will become commercially available and novel applications will emerge based on the manipulation of audio data within multimedia or audiovisual technologies. However, the feasibility of such future applications, as well as some current ones, greatly depends on the use of data compression techniques which reduce the data transmission rate and memory storage requirements. Given the existence of such techniques, terrestrial or satellite transmission channels can be economically employed for single or multi-channel audio data transmission, and also data storage media can be efficiently utilized for storing lengthy segments of acoustic signals. The storage and transmission of such high-quality audio data (here it will be considered as reference the Compact Disc format, based on a 44.1 kHz sampling rate and 16-bit resolution) results in the relatively high bit-rate of 706 kBits/s, per data channel. This data rate can be technically or economically prohibitive for many applications, and this necessitates the introduction of data compression, preferably by using low-complexity methods (so that real-time implementations are not impeded), and without the insertion of perceptually detectable distortions. Applications which have emerged or are expected to appear with strong dependence on such signal compression technology , are in the area of high-fidelity audio for radio broadcasting (especially for the Digital Audio Broadcasting -DAB format [2]), in the area of multichannel audio for HDTV, in storing and processing of audio signals for domestic (e.g. multimedia or home studio) and professional applications (multichannel music recording), in transmitting audio data through computer or communication networks (e.g. ISDN), etc. Coding and data compression methods for acoustic signals have been known for at least 4 decades, but until recently they were mainly concerned with speech signals [3], [4]. More recently, Transform Manuscript