The perception of one’s own voice influences the acceptance of hearing devices, such as headphones, headsets or hearing aids. When these devices fully or partially occlude the ear canal, the wearer’s own voice sounds boomy or like talking in a barrel. This is calledocclusion effect. Occluding the ear canal results in an amplification of body-conducted sounds, mainly at low frequencies, and an attenuation of air-conducted sounds, predominantly at high frequencies, compared to the open ear. This contribution provides a comprehensive treatment of Occlusion Effect Cancellation (OEC) and its relation to Active Noise Cancellation (ANC) using digital signal processing. A novel effective filter structure is presented which offers some degree of adaptability and adjustability. Furthermore,digitally openingandclosingthe ear is evaluated by listening tests and objective measurements.

The intelligibility of speech from a telephone or a public address system is often affected by acoustical background noise in the near-end listening environment. Speech intelligibility and listening effort can be improved by adaptive pre-processing of the loudspeaker signal. This is called Near-End Listening Enhancement (NELE). The speech spectrum is dynamically modified, taking the acoustical background noise at the near-end into account. In this paper, two opposite NELE strategies with either Noise-Masking-Proportional Shaping or Noise-Masking-Inverse Shaping are proposed which are appropriate for different noise characteristics. Both strategies are formulated in closed form in the frequency domain. They do not require to optimize an intelligibility measure but use explicitly the masking threshold. Motivated by the frequency domain approach, a simpler time domain solution is derived which is based on linear prediction techniques and does not need the masking calculations. The proposed NELE solutions outperform state-of-the-art in terms of computational complexity, memory requirement, continuous processor load, and latency.

Vector quantization is an integral part of modern speech and audio codecs. This study proposes the logarithmic pyramid vector quantizer (LPVQ), which is a gain-shape vector quantizer specifically designed for the quantization of Laplace distributed memoryless sources. The objective of the study is a theoretical analysis of the LPVQ: We determine the distortion of the shape quantizer with respect to rate and vector dimension for the quantization of an i.i.d. Laplace source. Furthermore, we derive formulas for the quantization signal-to-noise ratio (SNR) of the shape quantizer and the quantization SNR of the LPVQ, giving reliable results for an effective bit rate per vector coordinate of 2bits and higher. We study the SNR-behavior of the LPVQ for infinite dimensions and compare it with the maximal achievable SNR according to the rate-distortion bound. After having proposed a strategy for the allocation of the bit rate for the gain quantization and the quantization of the shape vector, respectively, we verify the derived formulas by simulations.

Occluding the ear canal alters the perception of the own voice. This is called the occlusion effect. Low frequency components of the own voice are amplified and high frequency components are attenuated. In this contribution, we extend our previously published Active Occlusion Cancellation (AOC) system for headphones and hearing aids. It uses the principle of Active Noise Cancellation (ANC) to enhance the perception of the own voice for an occluded ear canal. We combine robust feedback controller design with an adaptive mechanism to attenuate the low frequency components of the own voice. Furthermore, we include an equalizing hear-through filter to enhance high frequency components. The AOC system is evaluated in a hearing test with an in-ear headphone connected to a real-time processing system. Objective and subjective metrics show a clear improvement of the perception of the own voice. The occlusion effect is significantly reduced by the system.

Algorithms for Near-End Listening Enhancement (NELE) improve the intelligibility of speech from the far-end, played back in a near-end noisy environment, by adaptively filtering the speech signal and taking into account the near-end background noise characteristics. In contrast to previous works, this study considers that the speech from the far-end is also disturbed by additive noise. A noise reduction (NR) subsystem is concatenated to NELE and a joint control of NR and NELE is proposed. Listening tests confirm that the joint control leads to a significantly higher intelligibility and quality.

In studies on artificial bandwidth extension (ABE), there is a lack of international coordination in subjective tests between multiple methods and languages. Here we present the design of absolute category rating listening tests evaluating 12 ABE variants of six approaches in multiple languages, namely in American English, Chinese, German, and Korean. Since the number of ABE variants caused a higher-than-recommended length of the listening test, ABE variants were distributed into two separate listening tests per language. The paper focuses on the listening test design, which aimed at merging the subjective scores of both tests and thus allows for a joint analysis of all ABE variants under test at once. A language-dependent analysis, evaluating ABE variants in the context of the underlying coded narrowband speech condition showed statistical significant improvement in English, German, and Korean for some ABE solutions.

In announcement and communication systems, clean speech is often played back by loudspeakers in the presence of local background noise which decreases intelligibility. Speech reinforcement is a technique that enhances the intelligibility by adaptively filtering the speech, usually based on measured noise characteristics. In this contribution, solutions known from literature are applied to the case of public address systems where the microphone for noise measurement and the loudspeaker are close to each other and are far away from the listener on the ground. After establishing an acoustical model, the main problems are identified to be strong echoes between loudspeaker and microphone which make conventional noise estimation impossible and lead to an unstable system. A new approach including echo path estimation and echo-aware noise estimation is proposed and evaluated by means of simulations.

A drawback of digital transmission of analog signals is the unavoidable quantization error which leads to a limited quality even for good channel conditions. This saturation can be avoided by using analog transmission systems with discrete-time and quasi-continuous-amplitude encoding and decoding, e.g., Analog Modulo Block codes (AMB codes). The AMB code vectors are produced by multiplying a real-valued information vector with a real-valued generator matrix using a modulo arithmetic. Here, algorithms for improving the decoding performance are presented. The Lattice Maximum Likelihood (LML) decoder, a variant of the Discrete Maximum Likelihood (DML) decoder, is derived and analyzed. It refines the Zero Forcing (ZF) result if necessary, thus achieving near-ML signal quality with a reduced decoding complexity. A reduced complexity is essential for decoding high-dimensional code words. Additionally, pre- and post-processing methods are presented and analyzed, which increase the signal-to-distortion ratio (SDR) of the received symbols.

A novel noise-robust soft Voice Activity Detector (VAD) operating in the short-time Fourier domain is presented. A speech energy gain is obtained by frame-wise processing of a noisy speech signal with a speech codebook algorithm. This gain can be used for robust voice detection. A speaker-independent speech codebook, consisting of spectral envelopes, is created in the training process. While applying the algorithm, the codebook is adapted in every frame to the current speaker by combining the harmonic pitch structure of the actual noisy speech frame with the codebook entries. Soft VAD values ranging from zero to one are calculated by post-processing of the speech gain which is obtained using gain shape vector quantization. A binary VAD is carried out by applying a threshold. The proposed method does not rely on noise a-priori knowledge and is robust w.r.t. highly non-stationary noise and adverse SNR conditions. In addition, it is possible to compromise between the detection-rate and the false-alarm-rate by varying a threshold without increasing the total number of mis-detections. Compared to state-of-the-art VAD systems, the proposed method is characterized by better detection-rates at significant lower false-alarm-rates.

A novel noise power spectral density (PSD) estimator for disturbed speech signals which operates in the short-time Fourier domain is presented. A noise PSD estimate is provided by constrained tracing with time of the noisy observation separately for each frequency bin. The constraint is a limitation of the logarithmic magnitude change between successive time frames. Since speech onset is assumed as sudden rises in the noisy observation, a fixed and adaptive tracing parameter β has been derived to track the contained noise while preventing speech leakage to the noise PSD estimate. The experimental evaluation and comparison with state-of-the-art algorithms, SPP and Minimum Statistics, confirms a lower logarithmic noise estimation error and superior speech enhancement rated in a standard noise reduction system. The proposed concept has extremely low computational complexity and memory usage. Thus, it is well suited for applications where processing power and memory is limited.

In this paper, a novel noise-shaping method for Multi-Channel Linear Prediction (MCLP) is presented. Without special consideration, the quantization noise of the prediction error poses a serious problem in multi-channel prediction as each noise component distorts the reconstruction of every channel at the decoder.

This paper presents a method to enhance a speech signal disturbed by wind noise. The wind noise is generated by turbulences in an air stream close to the microphone which picks up the desired speech signal. As the majority of speech enhancement algorithms works in the frequency domain, the short term power spectrum (STPS) of the unwanted noise must be estimated to reduce the wind noise. Conventional algorithms for background noise estimation fail in the case of wind noise due to its non-stationary characteristics. Hence, it is necessary to use special methods for the estimation and reduction of wind noise. The proposed system exploits the spectral characteristics of speech and noise to estimate the wind noise STPS. The spectral power distribution of wind noise and the pitch frequency of speech are used to generate a binary mask for the noise STPS estimation. This method is dependent on a precise pitch estimation. To reduce estimation errors a robust pitch estimation method using knowledge from prior estimates is presented. An evaluation and comparison with other wind noise reduction techniques shows improved speech enhancement of the proposed method.

A new noise production and propagation model for open- and closed loop linear predictive coding (LPC) is proposed in this paper. The model allows to accurately predict the overall SNR even at lower bit rates where the conventional high rate theory fails. Moreover, a source of LPC encoder instabilities is pointed out which is due to the interaction between the quantizer and the (filtered) feedback of the quantization error. The new model is verified by measurements.

This contribution presents an efficient technique for the enhancement of speech signals disturbed by wind noise. In almost all noise reduction systems an estimate of the current noise power spectral density (PSD) is required. As common methods for background noise estimation fail due to the non-stationary characteristics of wind noise signals, special algorithms are required. The proposed estimation technique consists of three steps: a feature extraction followed by a wind noise detection and the calculation of the current wind noise PSD. For all steps we exploit the different spectral energy distributions of speech and wind noise. In this context, the so-called signal centroids are introduced. Investigations with measured audio data show that our method can cope with the non-stationary characteristics and enables a sufficient reduction of wind noise. In contrast to other wind noise reduction schemes the proposed algorithm has low complexity and low memory consumption.

In most communication systems, being able to adapt the error protection strength of the physical layer is essential to ensure the functionality of the system in potentially strongly varying conditions. Traditional forward error correction code design focuses on the optimization of codes with a fixed code rate for a worst-case channel condition. So-called rate-compatible codes, on the other hand, are tailored to support multiple code rates. In this paper we present an optimization strategy for a rate-compatible system based on shortening and puncturing of Low-Density Parity-Check (LDPC) codes. While the technique itself is well known, we propose a novel joint optimization of the resulting multi-rate LDPC code ensemble at all rates. This joint optimization achieves close to capacity decoding thresholds over an arbitrarily wide range of target code rates using only a single encoder and decoder implementation.

\gls{HARQ} schemes are employed in mobile communication systems to reduce the residual Frame Error Rate (FER) after channel decoding. If decoding fails, the receiver requests the transmission of additional redundancy about these frames to improve the decoding success. Signaling is carried out over a feedback link from the receiver back to the transmitter. However, errors on the feedback link severely influence the behavior of the HARQ system and may degrade the overall system performance. In this contribution, the impact of HARQ signaling errors on the FER, the average number of HARQ transmissions per frame, and the system throughput is investigated. We derive analytical expressions for all performance measures based on FER simulations of the HARQ system with reliable feedback. The theoretical results are verified by the example of the LTE physical layer, however, they are also applicable to other communication systems. Moreover, target requirements for the maximum tolerable feedback error probabilities are derived which are used to develop design guidelines for HARQ feedback transceivers.

Efficient digital transmission of analog speech, audio or video requires source coding which introduces unavoidable quantization errors. The bit stream produced by the source encoder needs to be protected against transmission errors by channel coding. The split of a given gross bit rate between source and channel coding is a compromise taking into account the design target of the worst case channel. Thus, even in clear channel conditions the quality of the decoded source signal is limited because of the quantization errors. Hybrid digital-analog (HDA) codes address this limitation by additionally transmitting the quantization error with quasi-analog methods (discrete-time, quasi-continuous-amplitude) with neither increasing the transmit power, nor the occupied frequency bandwidth on the radio channel. However, the decoding complexity of existing HDA codes is infeasible for the required block lengths. In this paper, the decoding complexity problem is solved by a new design approach. These HDA codes benefit from well-known digital codes. Furthermore, theoretical bounds are derived and explicit guidelines for the design of superior HDA systems are given. By experimental verification it is shown that the HDA concept may outperform conventional purely digital transmission systems at all channel qualities while additionally eliminating the quality saturation effect.

Rate-compatible channel codes are commonly used in mobile communication systems to adapt the protection against transmission errors to the varying channel conditions. In the LTE radio system, e.g., the code rate adaptation is carried out based on a fixed-rate mother Turbo code by means of bit puncturing and bit repetition. However, lower code rates can alternatively be provided by \gls{IC} Turbo codes. In RCIC Turbo codes, known bits (dummy bits) are inserted into the information bit sequence during Turbo encoding, providing perfect \apri knowledge to the mother decoder. In this contribution, a detailed analysis of RCIC Turbo codes is provided. It will be shown that these codes outperform rate-compatible Turbo codes using bit repetition in terms of frame error rate and convergence speed. Furthermore, their EXIT charts will be derived from the EXIT chart of the mother Turbo code. This EXIT chart analysis enables a novel semi-analytical optimization tool for RCIC Turbo codes and the efficient EXIT chart comparison of competing codes, avoiding complex simulations. Finally, a HARQ scheme for LTE, based on RCIC Turbo codes, is presented which achieves a considerably higher system throughput compared to the standard LTE solution.

A new method for wideband speech transmission is proposed which is fully backwards compatible with narrowband telephone systems. For this purpose, a pitch-scaled version of the higher speech frequencies (4 - 6.4 kHz) is inserted into the previously “unused” 3.4 - 4 kHz frequency range of standard telephone speech. This operation is reverted at the decoder side. A consistently good wideband speech quality can be achieved, even after transmission over common codecs and codec tandems. The quality impact on the narrowband part of the signal is insignificant.

A novel near field filter-and-sum beamformer using non uniform frequency sub-bands is presented. The concept is based on numerical optimization of the reception characteristic of the microphone array. In order to improve the reception characteristic over frequency and space, a non uniform filterbank is utilized to subdivide the frequency range. Individual optimization processes for each sub-band result in a clearly improved reception characteristic. The new system is able to closely approximate a target (independently of the frequency) which can be defined according to the application.

This contribution addresses the enhancement of noisy speech signals picked up by a dual microphone mobile phone in hands-free position. A novel technique to estimate the noise power spectral density is presented which combines two methods: a single microphone algorithm based on the speech presence probability and a dual microphone technique exploiting the coherence properties of the target signal and the background noise. Due to the novel approach, the weakness of both methods can be overcome. Since the proposed method requires knowledge of the current coherence properties, a technique is presented which estimates the coherence of the speech and noise signals, which is usually not known in practice.

One problem of conventional digital transmission systems is that the transmission fidelity always saturates at the design channel quality (worst-case channel condition); this saturation is due to the irreversible quantization error introduced by the source coder. Hybrid digital analog (HDA) codes address this problem by additionally transmitting the inherent quantization error by pseudo-analog methods. However, the design and the decoding of these HDA codes is complex or even impossible for long block lengths (look-up table decoding). In this study we propose an HDA transmission system for sources with correlation supporting long block lengths. This is achieved by combining transform coding with HDA transmission using well-known digital channel codes. For the evaluation Monte Carlo simulations are used with a normalized discrete cosine transform (DCTN) and Turbo codes as the digital part and LMMSE estimation in the pseudo-analog part of the HDA system. The proposed HDA system with transform coding outperforms the purely digital transmission system at all channel qualities and exploits the expected gain due to source correlation.

Instrumental evaluation of the perceived audio signal quality is an important tool for the development of audio signal enhancement and transmission systems. There are various single channel measures which can be used for different application scenarios. Binaural signals have not received much focus so far and no sophisticated model of spatial perception is utilized in the available measures. In this contribution, an extension to Perceptual Evaluation of Audio Quality (PEAQ) is presented which makes use of a recently proposed binaural hearing model. It is shown that the inclusion of spatial information into the instrumental quality measurement leads to a strongly increased correlation between the instrumental measure and a listening test.

Fast and robust acoustic system identification is still a research topic of interest, because of the typically time-variant nature of acoustic systems and the natural performance limitation of electroacoustic measurement equipment. In this paper, we propose NLMS-type adaptive identification with perfect-sweep excitation. The perfect-sweep is derived from the more general class of perfect sequences and, thus, it inherits periodicity and especially the desired decorrelation property known from perfect sequences. Moreover, the perfect-sweep shows the desirable characteristics of swept sine signals regarding the immunity against non-linear loudspeaker distortions. On this basis, we first demonstrate the fast tracking ability of the perfect-sweep NLMS algorithm via computer generated simulation of a time-variant acoustic system. Then, the robustness of the perfect-sweep NLMS algorithm against non-linear characteristics of real measurements in a time-invariant case is presented. By finally addressing the measurement of quasi-continuous head-related impulse responses, we face the combined challenge of time-variant and possibly non-linear distorted acoustic system identification in a real application scenario and we can demonstrate the superiority of the perfect-sweep NLMS algorithm.

Artificial Bandwidth Extension (ABWE) exploits spectral dependencies of speech signals and aims at recovering missing high frequency components if only the narrowband speech signal is available. This contribution provides an information theoretic view on ABWE when used in noisy conditions. Based on the results of [1], a performance bound of ABWE is formulated if the narrowband signal is disturbed by additive noise. The performance bound is evaluated using real entropy measurements and the influence of noise suppression prior to ABWE is investigated.

This paper discusses the application of noise reduction algorithms for dual-microphone mobile phones. An analysis of the acoustical environment based on recordings with a dual-microphone mock-up phone mounted on a dummy head is given. Motivated by the recordings, a novel dual-channel noise reduction algorithm is proposed. The key components are a noise PSD estimator and an improved spectral weighting rule which both explicitly exploit the Power Level Differences (PLD) of the desired speech signal between the microphones. Experiments with recorded data show that this low complexity system has a good performance and is beneficial for an integration into future mobile communication devices.

A novel numerical design approach for allpass transformed DFT filter-banks with subsampling is proposed. Such frequency warped filter-banks possess a non-uniform time-frequency resolution which is of interest, e.g., for speech and audio processing. The coefficients of the FIR synthesis filters are determined by a constrained least-squares error (CLSE) design. Thereby, the stopband energy of the synthesis filters is minimized with the constraint for perfect reconstruction (PR). This design by quadratic programming accounts also for polyphase network (PPN) filter-banks where the prototype filter degree exceeds the number of subbands to facilitate an enhanced frequency selectivity. In contrast to known designs for allpass transformed analysis-synthesis filter-banks (AS FBs), perfect signal reconstruction can be achieved effectively with synthesis filters featuring a pronounced bandpass characteristic.

Mobile communication systems as, e.g, UMTS LTE perform rate matching and Hybrid Automatic Repeat reQuest (HARQ) to provide adequate decoding performance over a wide range of channel conditions. In UMTS LTE this is realized by means of a fixed rate convolutional Turbo code and a rate matching scheme based on bit puncturing and bit repetition. A promising alternative to convolutional Turbo codes with simple bit repetition are Rate-Compatible Insertion Convolutional (RCIC) Turbo codes which have proven considerable performance gains in systems without HARQ. Since modern systems conventionally employ HARQ, a novel HARQ scheme for UMTS LTE based on RCIC Turbo codes is presented in this paper. Extensive simulations have demonstrated that the conventional UMTS LTE HARQ scheme is outperformed by our novel HARQ scheme in terms of system throughput.

Low-density random linear fountain (LDRLF) codes are a type of LT (Luby transform) codes with optimum erasure correction under maximum likelihood (ML) decoding given a certain density or average check node degree. The upper bound on the residual symbol erasure rate is very tight and can be used for the design of LDRLF codes instead of performing time-consuming simulations. Using LDRLF codes for unequal error protection (UEP), the excellent erasure correction performance is maintained as well as the tightness of the upper bounds for each importance class. Since the UEP upper bounds may be complex to compute, we provide an extremely good approximation thereof which is well suited to design UEP LDRLF codes. Furthermore, we provide a heuristic criterion that has to be fulfilled in order to yield good approximations.

In this paper, the impact of HARQ signaling errors caused by unreliable feedback on the throughput of communication systems like UMTS LTE is investigated. We derive analytical expressions for the calculation of the average number of transmissions per data frame as well as the system throughput based on a general feedback channel model. In combination with the residual frame error rates (FERs) after channel decoding, measured for a system with error-free feedback, a semi-analytical evaluation of the influence of unreliable feedback information on the system throughput can be carried out for any feedback channel condition. Consequently, time-consuming simulations of the respective system with unreliable feedback can be avoided. The theoretical results are verified by extensive simulations of the UMTS LTE physical layer. They are also applicable to other communication systems.

Robust speech enhancement relies on the estimation of stationary as well as non-stationary background noise. This contribution presents a novel approach for estimating the short-term power spectral densities (ST-PSDs) of rapidly time-varying harmonic noise as produced, e.g., by cars or motorcycles. The well-known Minimum Statistics algorithm is modified by frequency warping controlled by the fundamental frequency of the harmonic noise which is assumed to be known a priori. The resulting noise estimates are used for the enhancement of the noisy signal. A detailed description of the algorithm is given and it is shown by a thorough analysis that the new solution considerably outperforms two conventional noise ST-PSD estimation techniques.

Insertion convolutional codes and the underlying principles attract more and more attention, e.g., in iterative decoding, joint channel coding and cryptography or even in channel estimation. Known bits (dummy bits) are inserted into the information bit sequence before convolutional (turbo) encoding. These bits support the decoding of the information bits resulting in an improved decoding quality. Although this concept is widely employed, there does not exist any theoretical evaluation of these codes. We provide a basis for such an evaluation by means of their EXIT charts. We will analytically derive their EXIT charts as a function of the EXIT charts associated with the underlying convolutional code for a transmission over a BEC as well as a binary input/continuous output AWGN channel. These results can be adopted to all applications where perfect information is exploited by a SISO decoder and provide an excellent basis for the analytical prediction of the expected gains.

This contribution presents a modified Kalman filter approach for single channel speech enhancement which is operating in the frequency domain. In the first step, temporal correlation of successive frames is exploited yielding estimates of the current speech and noise DFT coefficients. This first prediction is updated in the second step applying an SNR dependent MMSE estimator which is adapted to the (measured) statistics of the speech prediction error signal. Objective measurements show consistent improvements compared to estimators which do not take into account the temporal correlation or the influence of the input SNR on the statistics of the prediction error signal.

This paper introduces a novel and highly efficient realization of a spherical vector quantizer (SVQ), the "Gosset Low Complexity Vector Quantizer" (GLCVQ). The GLCVQ codebook is composed of vectors that are located on spherical shells of the Gosset lattice E8. A high encoding efficiency is achieved by representing the spherical vector codebook as aggregated permutation codes. Compared to previous algorithms, the computational complexity and memory consumption is further reduced by exploiting the properties of so called classleader root vectors and by a novel approach for the codevector-to-index-mapping. The GLCVQ concept can be generalized to vector dimensions that are multiples of eight. In particular, GLCVQ for 16-dimensional vectors is used in Amd. 6 to ITU-T Rec. G.729.1.

An improved approach for the estimation of the frequency dependent reverberation time (RT) by means of allpass transformed filter-banks is presented. It is shown that by means of these warped filter-banks, a much more accurate RT estimation at lower frequencies can be obtained than by octave filter-banks, which are commonly used for the estimation of the frequency dependent RT. Furthermore, allpass transformed filter-banks can achieve a much better approximation of the non-uniform frequency resolution of the human auditory system than octave filter-banks. A uniform or non-uniform (auditory) frequency resolution can thereby be simply adjusted by a single allpass coefficient. The RT estimation can be done with an allpass transformed DFT or DCT filter-bank. The warped DCT filter-bank is of special interest as it provides real-valued subband signals. This facilitates the use of a maximum-likelihood (ML) estimator for either a non-blind estimation of the frequency dependent RT from a room impulse response or a blind estimation from a reverberant speech signal.

In this paper a novel approach for broadband feedback active noise control (ANC) is presented which is based on the combination of classical non-adaptive feedback and adaptive feedback ANC techniques. The non-adaptive part is suitable to attenuate low frequency ambient noise whereas the adaptive part attenuates periodic components of the ambient noise. The proposed technique yields a higher overall noise attenuation performance compared to a purely classical non-adaptive feedback or purely adaptive feedback ANC system. In addition to that, the combination of both techniques is also beneficial for practical realizations since the adaptive feedback ANC stabilizes the overall system. With regard to low cost headset devices, the impact of practical hard ware constraints such as low-cost analog-to-digital and digital-to analog converters (ADC, DAC) is discussed. As a conclusion, a mixed analog-digital realization of the new approach is proposed.

Low-density parity-check (LDPC) codes have proved to be very powerful channel coding schemes with a broad range of applications. However, as maximum-likelihood decoding is utterly complex, suboptimal decoders have to be employed. One of the most popular decoding algorithms of LDPC codes is belief propagation (BP) decoding. In this paper, we present a novel scheduling for belief propagation decoding of LDPC codes. The new approach combines probabilistic scheduling with the known shuffled and check-shuffled serial scheduling algorithms. The resulting probabilistic shuffled and probabilistic check-shuffled decoders show a superior performance in terms of residual bit error rate. The drawback is that the convergence speed is slightly decreased. However, the convergence is still faster than for the standard and probabilistic flooding algorithms. Furthermore, we have adapted the probabilistic flooding schedule and the proposed probabilistic shuffled schedule to the non-binary case. We show that the aforementioned effects on the binary decoder can similarly be observed when applying the different schedules to the decoding of LDPC codes over higher order Galois fields.

In this paper, two particular instances of LT codes with short message blocklength k and maximum likelihood (ML) decoding are investigated, i.e., random linear fountain (RLF) codes and (nearly) check-concentrated LT codes. Both show an almost equally good performance. The focus of this paper will be on RLF codes, a type of LT codes whose generator matrices are constructed from independent Bernoulli trials and have a binomial check node degree distribution. A new simple expression for an upper bound on the bit erasure probability under ML decoding is derived for RLF codes with density Δ = 0.5, i.e., with check node degree distribution Ω(x) = 2-k(1+x)k. It is shown that RLF codes with a minimum density far less than 0.5 are equally well suited to achieve a certain bit erasure probability for a given reception overhead. Furthermore, a characteristic term from a general upper bound on the bit erasure probability under ML decoding is identified that can be used to optimise check node degree distributions. Its implications on the performance of LT codes are qualitatively analysed.

Beside the well known iterative Belief Propagation algorithm several alternative decoding schemes for Low-Density Parity-Check (LDPC) codes providing better performance in terms of residual error rate, convergence speed or computational complexity have been developed in the last years. Recently, Informed Dynamic Scheduling has been proposed in [1] providing different decoding strategies that dynamically decide which messages are passed throughout the decoding process. It was shown that the overall convergence can be sped up considerably and also more errors can be corrected compared to other (non-dynamic) decoding strategies. However, these improvements are somehow overshadowed by a significant amount of additional computational complexity that is needed for the selection of the messages to be updated in each decoding step. We propose two novel dynamic decoding strategies that allow for a flexible adaptation of the decoder's dynamics and reduce the additional complexity remarkably while maintaining, and in some cases even exceeding, the convergence speed and error rate performance of currently known dynamic schedules.

Channel sounding, channel modeling and their parameter extraction in particular considering MIMO configurations, plays an important role for todays realistic wireless system design and evaluation. The analysis of the gathered channel sounding data has to match the specifics in the measurement system configuration. Within this contribution directional antenna arrays are used at both sides of the measured link. Furthermore, the measurement system applies automatic gain control between subsequent MIMO snapshots and additionally within the MIMO matrix. Both features lead to improved measurement accuracy. A new extension of existing power and delay domain analysis procedures considering the automatic gain control within the MIMO matrix is introduced and evaluated. The results show that delay spread and K-factor are more affected, while shadow fading and cross polarization ratios are almost not.

Extensive analysis of channel measurement data for the investigation of new channel models plays a fundamental role in designing and evaluating todays wireless communication systems. Besides the knowledge of the parameters and their distribution functions, the cross correlation between them is of great interest and can be exploited for more sophisticated channel simulations. In this contribution the cross correlation characteristics of the large scale parameters (LSP) in the power and delay domain are investigated for different parameter estimation procedures (taking the spatial filtering due to directional antennas into account or neglecting it), for various distance ranges between base station (BS) and mobile station (MS) as well as for different antenna heights.

In modern wireless communication systems different services with different bit rates and individual error protection requirements are supported by the same network. Rate matching by repetition is often employed after convolutional turbo encoding in order to achieve flexibly the required decrease of the code rate. This ensures near error-free transmission down to a specific target channel quality. We propose the optimization of the rate matching using EXIT charts. The rate for random repetition required for error-free decoding is determined semi-analytically avoiding extensive simulations. This includes the analytical optimization of the random repetition scheme and the derivation of the overall EXIT chart taking into account the convolutional turbo encoding and rate matching.

A novel semi-analytical signal processing model for the binaural coherence of homogeneous isotropic noise fields is presented in this contribution. This is derived from a simplified geometrical model of the human head, where the shadowing between the left and right ear is modeled by two nonreflecting circular plates. Based on Kirchhoff's diffraction theory, it is shown how the corresponding coherence is calculated. This model can be used as part of various binaural signal processing algorithms, such as speech enhancement for digital hearing aids or binaural speech transmission systems. In experiments using an artificial head in a highly reverberant environment, it is confirmed that the proposed theoretical model shows a good match with the coherence obtained from measurements.

Audio-visual source encoders for digital wireless communications extract parameter sets on a frame-by-frame basis. Due to delay and complexity constraints these parameters exhibit some residual redundancy which manifests itself in non-uniform parameter distributions and intra- as well as inter-frame correlation. This residual redundancy can be exploited by iterative source-channel decoding (ISCD) to improve the robustness against impairments from the channel. In the design process of ISCD systems the well known EXIT charts play a key role. However, in case of inter-frame parameter correlation, the classic EXIT charts do not provide reliable bounds for predicting the convergence behavior of ISCD. We explain the reasons for the so-called overshooting effect and propose a novel extension to the EXIT chart computation which provide significantly better bounds for the decoding trajectories. Four advanced ISCD system configurations are proposed and investigated using the benefits of the improved EXIT chart based system design. These configurations include regular and irregular redundant index assignments. In addition, we incorporate unequal error protection in the optimization of irregular index assignments. We show how to realize a versatile multi-mode ISCD scheme which operates close to the theoretical limit.

This contribution presents a wideband (50 Hz-7 kHz) speech enhancement system that is operating in the frequency domain. As a novel feature, techniques known from artificial bandwidth extension (BWE) are used to improve the spectral estimation process by exploiting the statistical dependencies between the low band (50Hz - 4kHz) and the high band (4-7kHz). Conventional noise suppression is used in the low band, while a novel approach is applied to the high band. Features from the processed (enhanced) low band signal are extracted and used to estimate subband energies of the high band. The weighting gains determined from these energy estimates are adaptively combined with conventional gains obtained in addition for the high band. The performance of the proposed method is shown to be consistently better than the conventional approach, especially at low input SNR values.

In this paper a novel speech enhancement algorithm for binaural dereverberation is proposed. It is based on a multichannel Wiener filter approach, which is optimized for the application to digital hearing aids and binaural telephony headsets. This is mainly done by two different modifications. First, an optimized model for the binaural coherence which takes the shadowing effects of the head into account. Second, a binaural input-output structure which does not affect the most important binaural cues, i.e., interaural time difference (ITD) and interaural level difference (ILD), and hence, keeps the localization ability. Evaluations with measured binaural room impulse responses (BRIR) show that this approach is capable of reducing reverberation especially in highly reverberant environments.

Artificial bandwidth extension (ABWE) of speech signals aims to estimate wideband speech (50 Hz – 7 kHz) from narrowband signals (300 Hz – 3.4 kHz). Applying the source-filter model of speech, many existing algorithms estimate vocal tract filter parameters independently of the source signal. However, many current methods for extending the narrowband voice source signal are limited to straightforward signal processing techniques which are only effective for high-band estimation. This paper presents a method for ABWE that employs novel data-driven modelling and an existing spectral mirroring technique to estimate the wideband source signal in both the high and low extension bands. A state-of-the-art Hidden Markov Model-based estimator evaluates the temporal and spectral envelopes in the missing frequency bands, with which the ABWE speech signal is synthesized. Informal listening tests comparing two existing source estimation techniques and two permutations of the proposed approach show an improvement in the perceived bandwidth of speech signals, in particular towards low frequencies. Subjective tests on the same data show a preference for the proposed techniques over the existing methods under test.

Source codec parameters determined by modern source encoders are usually very sensitive to noise on the transmission link. Natural residual redundancy of these parameters can be utilized by Soft Decision Source Decoding (SDSD)to increase the error resistance. Serially concatenating Forward Error Correction (FEC) and SDSD leads to a Iterative Source-Channel Decoding (ISCD) scheme which offers further improvements in robustness. ISCD exploits natural residual source redundancy by SDSD and artificial channel coding redundancy due to FEC in an iterative Turbo-like process. The convergence behavior of ISCD schemes can be analyzed by EXtrinsic Information Transfer (EXIT) charts. However, if an advanced ISCD system design is applied, offering incremental quality improvements for many iterations, the EXIT curves for SDSD do not specify a tight bound for the decoding trajectory anymore. In the initial iterations, the decoding trajectory "overshoots" the EXIT curves of SDSD, e.g.. In this paper, we give reasons for this overshooting effect. In addition, we propose a novel solution for determining the EXIT curve of SDSD which allows a more precise convergence analysis.

Iterative source-channel decoding (ISCD) exploits the residual redundancy of source codec parameters by using the Turbo principle. In this paper we extend the excellent capabilities of ISCD to a hybrid automatic repeat request (HARQ) scheme by novel incremental redundant index assignments. The incremental redundancy for HARQ is supplied by the source encoder and not the channel encoder. Simulation results show an excellent performance over a wide range of channel conditions, with inherently adapted bandwidth and complexity.

In some applications the transmission of discrete-time but continuous-amplitude (or multilevel) source symbols is required which might be more bandwidth efficient than conventional digital transmission. An appropriate method is to apply a source channel mapping (SCM) of M source symbols to N channel symbols. A geometrical approach for SCM has been introduced by Shannon and Kotel'nikov (Shannon-Kotel'nikov mappings). These systems are used to map M continuous-amplitude and discrete-time source symbols to N continuous-amplitude and discrete-time channel symbols without the intermediate step of a binary representation. These schemes are usually decoded using a maximum likelihood (ML) decoder which leads to optimum results in the mean square error sense for very good channels, but is suboptimal for noisy channels. In this paper the performance of an improved decoder, the minimum mean square error (MMSE) decoder is assessed. As a special case of a 1:2 expansion case (rate 1/2) Shannon-Kotel'nikov mapping, the Archimedes spiral is considered. The properties of the ML and MMSE decoder are examined and a graphical interpretation of the superior performance of the MMSE decoder is given. Furthermore, the robustness of the MMSE decoder w.r.t. an inaccurate estimation of the channel quality is determined. The concepts of the MMSE decoder which lead to a superior performance to the ML decoder can be generalized and applied to all Shannon-Kotel'nikov mappings.

Wireless communication of multimedia signals is often based on the transmission of source codec parameters which are protected against channel noise by channel coding. Commonly, a cyclic redundancy check (CRC) is introduced in order to check for erroneous frames after error correction. If the CRC fails, a bad frame indication (BFI) flag is set which controls a codec-specific frame error concealment. This results in the substitution of complete multimedia frames by their concealed versions even if only a few parameters may be corrupted. The proposed system is based on iterative source-channel decoding. A low complexity bad parameter indication (BPI) scheme is introduced which identifies faulty decoded parameters and enables more effective parameter-individual concealment strategies. This concept may be applied to any multimedia codec (speech, audio, video, images).

In this paper, we present a near-lossless compression scheme for scalar-quantized source codec parameters based on iterative source-channel decoding (ISCD). The scheme is comparable to a Turbo source encoder and can inherently incorporate protection against transmission errors. In order to realize a close-to-capacity transmission/compression scheme, we employ irregular redundant bit mappings and irregular inner channel codes. The irregular inner code is composed of (pseudo) randomly punctured convolutional codes of rate > 1, i.e., the compression is performed by strong puncturing of the inner code. The optimization goal is a minimum number of transmitted bits given certain source and channel conditions. We show how the inner and outer component can be jointly optimized resulting in a constrained nonlinear programming problem. An additional successive approximation based on linear programming is also presented. Simulation examples show the advantage over systems employing only a single irregular component.

The ability of the human auditory system for sound localization mainly depends on the binaural cues, especially interaural time and level differences (ITD and ILD). In the context of digital hearing aids and binaural audio transmission systems, these cues can be severely degraded by independent bilateral signal processing such as dereverberation or noise reduction. This contribution presents a novel two-stage binaural dereverberation algorithm which explicitly preserves the binaural cues. The first stage is based on a statistical model of the room impulse responses (RIR) and comprises a spectral subtraction rule which reduces late reverberation only. It includes a smoothing process of the spectral gains to reduce musical tones. In a second stage, the residual reverberation is attenuated by a dual-channel Wiener filter. This is derived from a coherence model of the reverberant sound field taking into account shadowing effects of the head. The overall binaural-input binaural-output structure efficiently reduces both early and late reverberation. In experiments as well as informal listening tests using measured binaural room impulse responses, the proposed algorithm significantly improves speech quality according to objective and subjective measures.

Noise reduction techniques that are relying on spectral weighting rules often generate annoying musical noise artifacts in the processed signal. In this paper, we present a postfilter (PF) for the spectral weighting gains that is capable of reducing musical noise in a simple but efficient way. It includes a robust detector for speech pauses and low SNR conditions and adaptively smoothes the weighting gains over frequency based on soft-decisions. Objective and subjective measurements show consistent improvements if the postfilter is applied to conventional noise reduction techniques.

This paper describes the speech and audio codec that has been submitted to ITU-T by Huawei and ETRI as a candidate for the upcoming super-wideband and stereo extensions of Rec. G.729.1 and G.718. The core codec in the current implementation is G.729.1 and the encoded frequency range is increased from 7 kHz to 14 kHz. Therefore, the maximum bit rate is raised from 32 kbit/s to 64 kbit/s by adding five bitstream layers. A comprehensive overview of the codec is presented with a focus on the mono coding components. The results of the listening tests that have been conducted during the ITU-T qualification phase are summarized. The proposed codec passes all quality requirements for mono input signals.

In this paper we investigate the application of adaptive postfiltering for the enhancement of reverberant speech. The considered method is commonly used in code excited linear prediction (CELP) speech coding to lower the impact of quantization noise in the excitation signal and the spectral envelope. We show that the underlying additive noise model is accurate enough to enhance speech which is recorded in an enclosed space where the resulting early reflections are usually modeled as a convolutive distortion. By means of adaptive filtering, the amplitudes of the unwanted peaks in the excitation signal are attenuated and the signal components at the harmonic peaks are emphasized. Both, single- and multi-channel dereverberation algorithms are proposed having a moderate computational complexity. Experiments have shown that this approach is capable of reducing early reverberation and attenuate the ldquodistance-effectrdquo arising from room reflections.

A new design for a recursive DFT analysis-synthesis filter-bank (AS FB) with critical subsampling and near-perfect reconstruction is proposed. The analysis filter-bank consists of allpass polyphase filters to achieve a high frequency selectivity with a low algorithmic complexity and low signal delay. The condition for perfect reconstruction (PR) leads to either unstable or anti-causal synthesis filters. This problem is solved by stable allpass polyphase synthesis filters which are designed by analytical closed-form expressions. The first design can achieve arbitrarily small aliasing, amplitude and phase distortions in dependence of the tolerable signal delay and algorithmic complexity. The second design avoids aliasing and amplitude distortions and minimizes the phase distortions at the expense of an increased system delay. The proposed IIR filter-bank possesses also a significantly lower algorithmic complexity than comparable FIR filter-banks.

This paper proposes a new speech enhancement algorithm for the suppression of background noise and late reverberation without a priori knowledge. A generalized spectral weighting rule based on estimations for the spectral variances of the late reverberant speech and background noise is used for speech enhancement. This allows to suppress speech distortions due to late room reflections without knowledge about the complete room impulse response. In contrast to existing methods, all needed quantities are estimated entire blindly from the reverberant and noisy speech signal. The new algorithm has also a low signal delay which is important, e.g., for speech enhancement in mobile communication devices or hearing aids.

We propose the combination of orthogonal frequency division multiplexing (OFDM) and turbo decodulation (TDeC) - a multiple turbo process consisting of iterative demodulation and iterative source-channel decoding - for transmission of correlated source codec parameters over wireless frequency-selective broadband fading channels. As OFDM systems split up frequency-selective fading channels into orthogonal flat-fading channels, individual modulation signal constellations sets with different bit mapping rules can be assigned to subcarriers taking into account the iterative reception and decoding process. The superior performance of the proposed OFDM-TDeC system is demonstrated by analyzing the achieved residual bit error rate (BER) and parameter signal-to-noise ratio (SNR).

In this paper, we present a novel near-lossless compression scheme for scalar-quantized source codec parameters. The scheme is comparable to a Turbo source coding approach and can inherently incorporate protection against transmission errors. We show that using the concept of EXIT charts and irregular codes, a linear programming optimization problem can be formulated and the solution of this problem leads to an irregular index assignment offering the best possible compression given the considered system model and a fixed channel quality. The performance of the compression scheme is demonstrated by a simulation example.

In this contribution a new system for the rapid development of real-time prototypes for digital audio signal processing algorithms on Windows PCs and a Digital Signal Processor (DSP) platform is presented. The goal of this system is to enable even unexperienced developers to transform the conceptual idea of a new algorithm into a standalone real-time demonstrator as quickly and conveniently as possible. In order to achieve this goal, a software architecture is defined where hardware and algorithm related programming issues are separated to allow the algorithm developer to completely focus on the implementation of the algorithm only. Compared to the earlier version [7], the new system supports all real-time prototype development phases from first Matlab simulations to the final highly efficient implementationin fixed point arithmetic and covers all relevant aspects such as e.g. control interface generation, function verification, complexity measuring and real-time data tracking. Development of prototypes based on general purpose PCs(RTProcPC) and the ADSP-21369 EZKIT [1] embedded DSP target (RTProcDSP) are currently supported (Fig. 1). The system has been successfully used to implement various real-time prototypes such as noise reduction, acoustic echo compensation, and digital hearing aid simulation.

This paper presents a modified Kalman filter operating in the frequency domain for single channel speech enhancement. The proposed scheme uses a two step approach. In the first step, information from previous, enhanced speech DFT coefficients is exploited to perform an estimation of the current speech coefficients. Investigations show that the highest prediction gain is achieved by modeling the temporal trajectory of the speech DFT coefficients as a complex autoregressive (AR) process. In the second step, the first prediction is updated using three alternative spectral estimators, including the conventional Kalman filter gain. Instrumental measurements show the improvement of the proposed scheme compared to purely statistical weighting rules.

A new method for hiding digital data in the bitstream of an ACELP speech codec is proposed in this paper. The key element of our method is an alternative search strategy for the ACELP codebook which allows for joint data hiding and speech coding. The concept has been examplarily applied to the AMR speech codec (12.2 kbit/s mode) and it is shown that steganographic data can be reliably transmitted at a rate of up to 2 kbit/s both with a negligible effect on the subjective quality of the coded speech and with reasonable computational complexity. Apart from data hiding, it is further pointed out that our method can also be exploited to reduce the codec bit rate.

A new design approach for an allpass transformed analysis-synthesis filter-bank (AS FB) with subsampling is proposed, which can be used for adaptive subband processing with non-uniform time-frequency resolution. Signal reconstruction is performed by an FIR synthesis filter-bank whose coefficients are determined by two different linear least-squares (LS) error designs. The presented unconstrained LS optimization leads to perfect reconstruction (PR), but an insufficient bandpass characteristic for the synthesis subband filters. This problem is solved by a second, equality constrained least-squares (CLS) error design, which achieves near-perfect reconstruction (NPR). In the process, the linear distortions are minimized with the constraint for a linear, time-invariant (LTI) overall transfer function. The aliasing-free reconstruction error is significantly lower than for known designs of allpass transformed DFT filter-banks with NPR. Moreover, the new filter-bank design allows for a low and adjustable signal delay.

A novel design for a two-channel IIR quadrature-mirror filter (QMF) bank with near-perfect reconstruction (NPR) is presented. The analysis filter-bank is given by an efficient polyphase network (PPN) implementation based on allpass filters. The arising phase distortions are almost compensated by stable allpass filters, designed via analytical closed-form expressions. In a first design, the remaining aliasing, amplitude and phase distortions become arbitrarily small in dependence of the tolerable system delay and algorithmic complexity, respectively. In a second design, aliasing and amplitude distortions are completely canceled and phase distortions are minimized at the expense of an additional signal delay. The proposed QMF banks have a lower algorithmic complexity than comparable designs.

We address the optimization of joint source-channel coding schemes for iterative source-channel decoding of first- order Markov sources. Compared to the traditional design, we propose two novelties: (1) source encoders, providing code words with a minimum Hamming distance dminges2, realized by linear block codes, and (2) irregular channel encoders which are optimized for both the source characteristics and the conditions on the channel. Inner code rates RC> 1 may be chosen in order to compensate for the additional source redundancy if required. Design examples for the AWGN channel and an overall code rate R=0.66 show that the proposed system is able to establish reliable communication within 0.3 dB of the capacity limit for an interleaver length of approximatively 200000 bits.

In this paper, we investigate turbo-coded transmission over a temporally correlated flat Rayleigh fading channel. Conventionally, channel estimation is performed prior to decoding and is solely based on periodically inserted pilot symbols. We here consider the case that the pilot spacing is not small enough to satisfy the Nyquist criterion for the bandlimited channel process. It is demonstrated that iterative code-aided channel estimation can then significantly improve the bit error rate (BER). Furthermore, we show that the convergence speed of the iterative channel estimation can be significantly accelerated without additional computational power by just carefully choosing the mapping strategy of systematic bits to data symbols.

Turbo DeCodulation is the combination of iterative demodulation and iterative source-channel decoding in a multiple Turbo process. The receiver structures of bit-interleaved coded modulation with iterative decoding (BICM-ID) and iterative source-channel decoding (ISCD) are merged to one joint Turbo system, which we further enhance in this paper by using a low- density parity check (LDPC) code for channel coding, resulting in a third iterative loop. We propose to use a special LDPC code structure with short sub-codes, which can be implemented very effectively in parallel. The quadripartite Tanner graph of the Turbo DeCodulation is presented, showing that the processing of all receiver nodes could be parallelized. Simulation results including an EXIT chart analysis demonstrate the excellent capabilities of Turbo DeCodulation with its performance gain exceeding the combined gain of BICM-ID and ISCD.

We consider the problem of transmitting a wideband speech signal with a cut-off frequency of fc= 7 kHz over a standardized narrowband (fc= 3.4 kHz) communication link in a backwards compatible manner. In a previous contribution we have shown that backwards compatibility can be achieved by using digital watermarking: we embedded compact side information about the missing high frequency band (3.4 - 7 kHz) into the narrowband speech signal. Here, we present a related system which is especially tailored to state-of-the-art narrowband speech communication networks such as GSM or UMTS. Therefore, we propose an improved low bit rate bandwidth extension algorithm along with a robust watermarking scheme for CELP-type speech codecs. The practical relevance of our system is shown by speech quality evaluations and by link-level simulations for the "enhanced full rate traffic channel" (TCH/EFS) of the GSM cellular communication system.

Current speech, audio, and video coding and transmission systems are either analogue or digital, with a strong shift from analogue systems to digital systems during the last decades. We have combined both digital and analogue schemes for the benefit of saving transmission bandwidth, complexity, and of improving the achievable quality at any given signal-to-noise ratio on the channel. The combination is achieved by transmitting pseudo analogue samples of the unquantized residual signal of a linear predictive digital filter which is called mixed pseudo analogue-digital (MAD) transmission. In this paper a new modulation scheme based on QPSK for digital information and an Archimedes spiral for the time discrete, pseudo analogue residual signal is introduced and evaluated.

This paper discusses a potential extension of the ITU-T G.729.1 speech and audio codec. The G.729.1 coder is hierarchically organized, i.e., the obtained quality increases with the amount of bits that is received for each frame. In particular, the bit rates of 8 and 12 kbit/s offer narrowband (50 Hz – 4 kHz) speech transmission. With a received bit rate of at least 14 kbit/s, the output bandwidth is increased to the wideband frequency range (50 Hz – 7 kHz). Here, we investigate efficient methods to provide the full wideband frequency bandwidth already for the lower bit rates of 8 and 12 kbit/s while maintaining interoperability with the standard implementation of G.729.1. These techniques are not necessarily limited to G.729.1 and thus may serve in other applications as well.

In our contribution we analyze the exit characteristics of feed forward convolutional codes employed in a parallel concatenated turbo scheme or as inner component in a serially concatenated scheme. It turns out that the exit characteristics reveal some general limitations. First we give an illustrative reason for the limited capabilities of feed forward and non-terminated recursive convolutional codes and secondly, we provide novel analytical means to determine the maximum mutual information available at the extrinsic output of feed forward convolutional decoders. This upper bound for the mutual information is only dependent on the channel quality and on the Hamming weight of the impulse response of the convolutional code. The paper concludes by determining the maximum mutual information for a wide range of channel conditions and code parameters, showing for which codes and channel conditions a maximum mutual information of ap1 bit can be reached.

Iterative source-channel decoding (ISCD) aims at the exploitation of the time-variant residual redundancy of the source samples, e.g., source codec parameters, for error concealment and quality improvements. In most previous publications the receiver had perfect knowledge of the amount of residual redundancy. This assumption would require a reliable, i.e. highly redundant, transmission of side information. In contrast, in this paper we present a relatively simple scheme, yet efficient and robust, by which the residual redundancy at the receiver can be estimated accurately without any side information, and then can be exploited adaptively. We present the achievable performance gains in an ISCD system including the estimation of the residual source redundancy at the receiver for various scenarios. Several methods of different performance and computational complexity are proposed, with some of them even outperforming a system with perfect side information. The latter, not quite intuitive fact, is explained in the paper.

Recommendation G.729.1 is a new ITU-T standard which was approved in May 2006. This recommendation describes a hierarchical speech and audio coding algorithm built on top of a narrowband core codec. One challenge in the codec design is the generation of a wideband signal with a very limited additional bit rate (less than 2 kb/s). In this paper, we describe the respective codec layer, which extends the transmitted acoustic bandwidth from the narrowband frequency range (50 Hz-4 kHz) to the wideband frequency range (50 Hz-7 kHz). The underlying algorithm uses a fairly coarse parametric description of the temporal and spectral energy envelopes of the high frequency band (4-7 kHz). This parameter set is quantized with a bit rate of 1.65 kb/s. At the decoder side, the high-frequency components are regenerated by appropriately shaping a synthetically generated ldquoexcitation signal.rdquo Apart from the algorithmic description and a discussion, we state a complexity evaluation as well as some listening test results.

Iterative source-channel decoding (ISCD) exploits the residual redundancy of the source, e.g., codec parameters, for quality improvements. In contrast to the well-known convolutional coded ISCD systems, we propose in this paper an ISCD scheme which features a superior performance but is based solely on two short block codes. The block codes serve as highly redundant index assignment and rate-1 inner channel code respectively. The improved capabilities are confirmed by EXIT charts. Finally, the effects of imperfect knowledge of the receiver on the residual redundancy are analyzed. With the proposed flexible index assignment no feedback from the receiver to the transmitter is required and competitive results are obtained even for slightly erroneously estimated residual redundancy

We present a technique to extend narrowband (NB) speech communication systems, using e.g. the GSM enhanced full rate (EFR) codec, with wideband (WB, 50-7000 Hz) capability. The limited acoustic bandwidth of narrowband speech coding is extended using a fairly coarse description of the missing high frequency band (3.4-7 kHz) in terms of temporal and spectral envelopes. The high-band parameters are quantized, transmitted and then used at the receiver side to regenerate the high frequency components. The parameter encoding is done by applying split vector quantization in a transformed domain. This quantization scheme can be scaled to match any given target bit rate. Several example configurations have been implemented and tested in MUSHRA-style listening tests

In contrast to common noise reduction systems, this contribution presents a digital signal processing algorithm to improve intelligibility of clean far end speech for the near end listener who is located in an environment with background noise. Since the noise reaches the ears of the near end listener directly and therefore can hardly be influenced, a sensible option is to manipulate the far end speech. The proposed algorithm raises the average speech spectrum over the average noise spectrum and takes precautions to prevent hearing damage. Informal listening tests and the speech intelligibility index indicate an improved speech intelligibility

In contrast to most of the clinical examinations, the sonotubometric assessment of the Eustachian tube function allows the registration of the tube opening at the perforated, as well as at the intact eardrum in physiological conditions. In this paper, a new procedure is introduced which performs a measurement of the transmission path nose/ear. With digital signal processing algorithms the generation of a new feature relevant for medical diagnostics is made accessible. The method is based on techniques known from speech processing: the acoustic tube model, the normalized least-mean-square (NLMS) algorithm with so-called perfect sequence (PSEQ) (Luke, 1992 and Luke and Schotten, 1995) excitation and the Levinson-Durbin algorithm. Within the framework of an interdisciplinary cooperation we developed a new real-time acoustic measurement prototype which opens up entirely new possibilities in otological diagnostics.

We present an embedded and hierarchical 8-32 kbit/s speech and audio coding algorithm that has been successfully submitted to the ITU-T as a candidate for ITU-T Rec. G.729.1 (ex G.729EV). The submitting consortium consisted of Siemens AG, Matsushita Electric Industrial Co. Ltd., and Mindspeed Technologies Inc. This contribution gives a comprehensive overview of the proposed codec, describes the implemented algorithms, and states a detailed characterization as well as results of the official G.729EV qualification tests

Today's speech coding and transmission systems are either analogue or digital, with a strong shift from analogue systems to digital systems during the last decades. In this paper, both digital and analogue schemes are combined for the benefit of saving transmission bandwidth, complexity, and of improving the achievable speech quality at any given signal-to-noise ratio (SNR) on the channel. The combination is achieved by transmitting pseudo analogue samples of the unquantized residual signal of a linear predictive digital filter. The new system, Mixed Pseudo Analogue-Digital (MAD) transmission, is applied to narrowband speech as well as to wideband speech. MAD transmission over a channel modeled by additive white Gaussian noise (AWGN) is compared to the GSM Adaptive Multi-Rate speech codec mode 12.2 kbit/s (Enhanced Full-Rate Codec), which uses a comparable transmission bandwidth if channel coding is included.

The SCELP (spherical code excited linear prediction) audio codec has recently been proposed as a new candidate for low delay audio coding based on linear prediction (LP). The new codec applies closed-loop vector quantization employing a spherical code in a gain shape manner. The spherical code is based on the apple-peeling code construction rule and in general does not require a codebook table for the encoding and decoding process. In this contribution, however, we propose to employ auxiliary information gathered in advance to reduce the computational encoding and decoding complexity at runtime significantly. This auxiliary information can be considered as the SCELP codebook. Due to the consideration of the characteristics of the apple-peeling code construction principle, this codebook can be stored very efficiently in read-only-memory. With the proposed principle, low computational as well as low memory complexity can be achieved simultaneously in the SCELP codec

A cost function for the optimization of mappings for bit-interleaved coded modulation systems with iterative decoding (BICM-ID) for a Rayleigh channel with applied IQ interleaving is derived. For this channel model the optimum mapping depends on the channel quality and on the rotation of the signal constellation set. Simulation results reveal that mappings optimized with the new cost function yield a better performance than previously known mappings. Furthermore, a second, easy procedure to obtain the optimum rotation angle for a signal constellation set in conjunction with IQ interleaving is described

In this contribution, we present a novel soft demodulation (SDM) method for higher order digital modulation schemes based on softbit source decoding (SBSD). The soft demodulator makes a soft decision on signal point level using a mean square (MS) estimator and can also exploit a priori knowledge about the source. Furthermore, we present a novel concept for unequal error protection (UEP) by higher order modulation with unequal power allocation (MUPA). Usually, the bits representing the quantized source codec parameters are protected by channel coding with UEP according to the individual bit error sensitivities. For communication systems where channel coding is not applicable, we propose the MUPA method with periodically time-varying power allocation. The average transmitted energy per bit remains unaffected. The significant SNR gains of the SDM combined with MUPA in comparison to systems with hard decision and a fixed signal constellation are demonstrated by simulation.

In recent years research in the area of artificial bandwidth extension (BWE) of speech signals has made significant progress. The intention of BWE is to produce wideband speech with a cut-off frequency of for instance 7 kHz from a narrowband version (e.g., with telephone bandwidth, i.e., 300 Hz − 3.4 kHz). The respective algorithms are based on the estimation of parameters of a source model for speech production given the knowledge of thenarrowband signal. Atheoretical performance bound on this estimation has been formulated in [1]. In order to overcome this boundary we propose the transmission of (compact) side information which can support the parameter estimation. Since a common additional channel would conflict with the requirement of backwards compatibility in narrowband communication systems, the side information is embedded as digital watermark into the narrowband speech.

In this contribution a modified scheme for linear prediction analysis is presented which controls the degree of decorrelation by a parameter α. In consideration of this parameter the signal to noise ratio of a linear predictive coding scheme is investigated and finally maximized for open loop quantization. Also, the new parameter can be used to control the spectral shape of the quantization error in the decoder output. As result the signal to noise ratio of the reconstructed signal can be increased for open loop quantization compared to conventional prediction analysis, and at the same time the perceptual encoding quality benefits from a moderate spectral shaping of the quantization noise in the decoded signal.

An efficient realization of a low delay filter-bank, termed as generalized filter-bank equalizer (FBE), will be proposed for noise reduction with low signal delay. The FBE is equivalent to a time-domain filter with coefficients adapted in the frequencydomain. This filter-bank structure ensures perfect signal reconstruction for a variety of spectral transforms with less restrictions than for an analysis-synthesis filter-bank (AS FB). A non-uniform frequency resolution can be achieved by an allpass transformation. In this case, the FBE has not only a lower signal delay than the AS FB, but also a lower algorithmic complexity for most parameter configurations. Another advantage of the FBE is the lower number of required delay elements (memory) compared to the AS FB. The noise reduction achieved by means of the AS FB and the FBE is approximately equal.

This contribution presents a stereo input-output beamformer for dual channel noise reduction. The computationally very efficient beamformer adapts superdirective filter design techniques to binaural input signals to optimally enhance signals from a given spatial direction. The beamformer outputs a stereo enhanced signal and thus preserves the spatial impression. Experiments in a real environment using a dummy head and various speech sources indicate that the proposed algorithm is capable of improving the speech intelligibility significantly.

In UMTS speech transmissions, the Adaptive Multi-Rate (AMR) speech codec is employed to allow for a dynamic assignment of data rates to individual users. The control of AMR modes is based on quality measurements of the transmission channel and aims at the maximization of speech quality by selecting the mode which is best suited to the current interference situation. In contrast to GSM, a reduction of AMR modes also leads to an increased cell capacity in UMTS. In this paper, a decentralized method for AMR mode switching is presented which is based on individual softbit measurements at the Viterbi channel decoder and optimizes the speech quality for a wide range of channel conditions. This method was developed by optimizing the AMR mode switching thresholds with respect to PESQ speech quality scores. The second part of the paper describes a new instrumental non-intrusive method for speech quality measurement by the evaluation of UMTS transmission parameters for application to UMTS-AMR speech transmissions. High correlations with the reference PESQ scores were observed.

A well known technique to protect bits with different sensitivities against channel errors is unequal error protection (UEP) by selective channel coding. However, some transmission systems such as Bluetooth and DECT include only weak or no channel coding. For those systems modulation with unequal power allocation (MUPA) is an appropriate technique to realize UEP without channel coding, MUPA allocates different transmission power to the modulation symbols according to the individual bit error sensitivities. In this paper we apply the MUPA concept to systems with softbit source decoding and soft demodulation, and enhance the MUPA approach by taking a 1storder Markov model in terms of 1storder a priori knowledge (AK1) into account (MUPA-AK1)

Turbo decodulation is the combination of iterative demodulation and iterative source-channel decoding to a single multiple turbo process. The receiver structures of bit-interleaved coded modulation with iterative decoding (BICM-ID), iterative source-channel decoding (ISCD), and iterative source coded modulation (ISCM) are merged to one novel turbo system, in which in two iterative loops reliability information is exchanged between the three single components, demodulator, channel decoder, and (softbit) source decoder. Simulations show quality improvements compared to the different previously known systems, which use iterative processing only for two components of the receiver. We propose the enhancement to turbo decodulation by an additional link for a direct supply of reliabilities from the softbit source decoder to the demodulator. We will demonstrate that using three-dimensional EXIT charts a convergence analysis of turbo decodulation similar to single turbo processes is possible.

We propose the application of iterative demodulation at a DVB-S2 receiver. The demodulation is improved by feedback of extrinsic information from the decoder of the LDPC channel code. The presented simulation results show an enhancement of about 0.3 dB, which is notable taking the gap to the Shannon limit into account. The additionally required computational demand is negligible and the receiver only needs to be slightly modified. With the unmodified transmitter the proposed scheme is completely standard compliant.

This paper examines the application of UDP-lite for unequal error detection in packet-switched speech transmission via Internet protocols (voice-over-IP) over UMTS radio channels. Traditionally, UDP is used as transport layer protocol, which contains a checksum that covers the complete packet. Thus, any packet with residual bit errors is discarded. Speech codecs like AMR, however, can tolerate bit errors in less sensitive parts of the bitstream. A more recent development, UDP-lite, provides unequal error detection with a partial checksum that covers only the sensitive parts of a packet. Thus, only packets with errors in important bits are discarded. We compare the use of UDP-lite for UMTS channels with convolutional and channels with turbo coding. The results show that the achievable quality improvement by applying UDP-lite depends on the residual bit error distribution of the chosen UMTS channel coding method. While we determined a quality improvement for channels with convolutional coding, we did not get an improvement for turbo coding. Furthermore, when combined with header compression, the convolutional coder with use of UDP-lite can reach the performance of the turbo coder with use of UDP

We present a new class of nonlinear block codes called source-optimized channel codes (SOCCs), which are particularly designed for parametric source encoding of speech, audio, and video. In contrast to conventional channel codes, the new codes are not optimized for minimizing residual bit-error rate, but maximizing the signal-to-noise ratio of transmitted source codec parameters. The decoding of SOCCs is not based on bit-error correction, but on parameter estimation. We compare SOCCs with other approaches to joint source/channel coding such as channel-optimized vector quantization, channel-constrained vector quantization, unequal error protection, and source-controlled channel decoding. In terms of performance, SOCCs show better robustness if under channel mismatch conditions. For real-world applications, SOCCs are attractive, since the separation of source and channel codec is preserved.

Bit-interleaved coded modulation with iterative decoding (BICM-ID) is a transmission scheme with excellent performance on fading channels. The mapping of the bit patterns to the elements of the signal constellation set (SCS) is the key design element of BICM-ID, especially regarding the asymptotic performance of the BER. We demonstrate that SCSes of PSK (phase shift keying) with optimum bit pattern assignment are significantly superior in the asymptotic error-floor region to the respective square QAM (quadrature amplitude modulation) SCSes. When modifying the PSK SCSes to non-regular SCSes, performance gains of more than 1 dB are realizable in the case of 4 bits-per-channel symbols. The achievable gains are theoretically determined and analyzed, and verified by simulations.

The source codec parameters determined by modern speech, audio, or video encoders are, especially at high compression rates, highly sensitive to transmission errors and they usually exhibit natural residual redundancy due to delay and complexity limitations in the encoding process. Error concealment techniques like softbit source decoding (SBSD) utilize the residual redundancy as a priori knowledge for parameter estimation. Therewith, the annoying artifacts of transmission errors can considerably be concealed. But, the optimal utilization of all terms of residual redundancy requires large computational complexity. In this paper, we derive a new efficient, nearly optimal version to SBSD. Significant complexity savings become possible by exploiting the intra-frame and inter-frame redundancy separately using two nested forward-backward algorithms. Compared to the optimal joint evaluation, the performance loss is negligibly small.

In this paper we propose to combine the exploitation of residual source redundancy and iterative demodulation of multilevel modulations in a turbo-like process to improve error concealment for the source codec parameters. The developed system, denoted as iterative source coded modulation (ISCM), uses the bandwidth efficiently since no channel code increases the transmitted bit rate. Thus, ISCM is suitable for communication systems without channel coding, but it can be enhanced by a channel code to resemble a multiple turbo code. Furthermore, without a complex channel decoder ISCM requires relatively low computational complexity. Nonetheless, simulations demonstrate considerable performance gains by ISCM with few iterations and small interleaver sizes. Additionally, measures are derived to ensure a correct relation between the unequal error protection in multilevel modulation and the different significance of bits of source codec parameters.

The aim of artificial bandwidth extension (BWE) is to convert speech signals with "standard telephone" quality (frequencies up to 3.4 kHz) into 7 kHz wideband speech. The principal key to high quality BWE is the estimation of the spectral envelope of the wideband speech. In general, this estimation of the wideband spectral envelope is based on a number of features that are extracted from the narrowband input speech signal. We investigate potential features and evaluate their suitability for the BWE application. The quality of each feature is quantified in terms of the statistical measures of mutual information and separability. It turns out that the best BWE results are obtained by using a large feature "super-vector" (/spl rarr/ high mutual information) which is subsequently reduced in dimension by a linear discriminant analysis (/spl rarr/ large separability). This solution also helps to reduce the computational complexity of the estimation of the wideband spectral envelope.

In this contribution, a new software environment for an embedded system is presented that is part of a laboratory for teaching students in digital audio signal processing. This environment enables even students with only a limited background in programming embedded technology to achieve remarkable implementations of realtime projects, including runtime user interaction. Hardware and algorithm related programming issues are separated for audio processing and interactive messaging. While the hardware related part is provided along with the software environment, the students can completely focus on algorithmic aspects. A well defined interface builds the bridge between the two programming components. Examples are presented to demonstrate the flexibility and efficiency of the given software platform. The underlying projects were defined, implemented and finally demonstrated by students in an interactive real-time demonstration.

In the competitive environment of cellular radio networks, the speech quality experienced by the end user is of vital importance for the network operator. Therefore, the operator needs reliable measures for monitoring and controlling the speech quality. For GSM mobile radio networks employing the adaptive multi-rate (AMR) speech codec, a new instrumental and non-intrusive speech quality metric is proposed. 'Non-intrusive' means that the metric is based only on received transmission parameters, such as the frame erasure rate, and needs neither the original nor the transmitted speech samples. However, for the validation of a new metric, standardized intrusive speech quality measures like PESQ, link-level simulations and field measurements are used. Furthermore, for an improved control of speech quality, a novel AMR mode switching procedure is introduced which outperforms the recommended method of the GSM standard.

We propose a specific channel encoder to support error concealment of mutually independent source codec parameters. To introduce artificial dependencies a recursive non-systematic convolutional (RNSC) code is applied which exhibits a code rate near r = 1. The channel encoder may be considered as a smearing filter. A TURBO-like exploitation of artificial dependencies and of residual source redundancy according to the iterative source-channel decoding (ISCD) algorithm permits step-wise quality improvements. Previously known error concealment techniques are outperformed in the most interesting range of channel conditions.

Whenever digital communication is subject to transmission errors, the utilization of reliability information at the receiver is beneficial. Soft-channel decoding techniques improve the error correcting capabilities and in addition, softbit source decoding performs error concealment. While the reliability gain of channel decoding is based on artificial diversities explicitly introduced at the transmitter, softbit source decoding exploits implicit redundancies remaining in the bitstream after state-of-the-art source encoding. In an iterative source-channel decoding scheme the reliability gains of both are exchanged iteratively in order to enhance the common error resistance. However, first investigations have shown that the number of profitable iterations is limited to small values. In this paper, the convergence behavior is analyzed using a modern tool called EXIT charts. EXIT charts permit us to predict the convergence behavior by studying the individual components. As a novelty, EXIT characteristics are specified for softbit source decoding. The convergence analysis is confirmed by simulation.

We present a soft-partitioned frequency-domain adaptive filter (SPFDAF) and compare it in the context of other LMS type acoustic echo cancelers. The SPFDAF uses a nonrectangular window to partition the model taps (impulse response) of an echo canceler. Our soft window can be approximated very efficiently in the DFT domain and still attenuates aliasing components of the DFT by 40 dB. This technique results in a very low computational complexity and maintains an excellent signal quality.

We present an algorithm to derive 7 kHz wideband speech from narrowband "telephone speech". A statistical approach is used that is based on a hidden Markov model (HMM) of the speech production process. A new method for the estimation of the wideband spectral envelope is proposed, using nonlinear state-specific techniques to minimize a mean square error criterion. In contrast to common memoryless estimation methods, additional information from adjacent signal frames can be exploited by utilizing the HMM. A consistent advantage of the new estimation rule is obtained compared to previously published HMM-based hard or soft classification methods.

This paper introduces two short-time spectral amplitude estimators for speech enhancement with multiple microphones. Based on joint Gaussian models of speech and noise Fourier coefficients the clean speech amplitudes are estimated with respect to the MMSE or the MAP criterion. The estimators outperform single microphone minimum mean square amplitude estimators when the speech is highly correlated and the noise is sufficiently uncorrelated. Whereas the first MMSE estimator also requires the desired signals to be in phase, the second MAP estimator performs a direction-independent noise reduction. The estimators are generalizations of the well known single channel MMSE estimator derived by Ephraim and Malah (1984) and the MAP estimator derived by Wolfe and Godsill see (Proceedings of the 11th IEEE Workshop on Statistical Signal Processing, p.496-499, August 2001) respectively.

In this paper we present a voicing controlled, speech parameter based frame loss concealment for frames that have been encoded with the Adaptive Multi-Rate (AMR) speech codec. The missing parameters are estimated by interpolation and extrapolation techniques that are chosen in dependence of the voicing state of the speech frames preceding and following the lost frames. The voicing controlled concealment outperforms the conventional extrapolation/muting based approach and it shows a consistent improvement over interpolation techniques that do not distinguish between voiced and unvoiced speech. The quality can be further improved if additional information about the predictor states of predictively encoded parameters is available from a redundant transmission in future packets.

This contribution introduces instrumental speech quality measures for the GSM system based only on transmission parameters. Certain combinations of GSM parameters which quantify the transmission quality in terms of bit error rate, received power level, etc., were shown be suitable for the prediction of the resulting speech quality. Neither the original nor the received speech signal is needed for this kind of prediction. The proposed speech quality measures have been validated by extensive link-level simulations which are based on measurements of transmission parameters collected in a GSM-1800 network. Speech samples were produced by bit-exact transmission simulations using the measured link parameters for channel modelling. The reference speech quality assessments of these samples were carried out with the PESQ algorithm Ref.3. The correlation of the presented parameter measures with the intrusive PESQ measure is remarkable.

Whenever digital communications is subject to constraints in delay and complexity, some residual redundancy remains after source coding. Shannon's separation theorem does not hold strictly and hence joint source-channel decoding is beneficial. In this paper we discuss two competing approaches exploiting residual redundancy in terms of (auto)-correlation. Both will be analyzed with respect to their impact on quantization noise and channel distortion. The first approach exploits redundancy at the receiver for minimum mean squared error estimation of source codec parameters. The second approach exploits correlation at the transmitter by predictive encoding. The prediction gain allows to reduce quantization noise or bitrate. The spare bitrate can be used for specific error protection and hence reduction of channel distortion. But otherwise, less redundancy is utilizable at the receiver for parameter estimation. Simulations will show that the predictive encoding scheme improves quality in a wide range of channel conditions, but in contrast to Shannon's separation theorem the utilization of correlation at the receiver side is beneficial in very bad channel conditions.

Residual echo arises in hands-free telephony equipment due to insufficient echo canceler convergence, but can be suppressed using a postfilter. The most important control parameter for postfilter adaptation is therefore the residual echo power spectral density (PSD). In this contribution we present and compare residual echo PSD estimation techniques. We introduce a new partitioned block-adaptive estimator delivering unbiased residual echo PSD estimates in strongly reverberant and noisy acoustic environments.

In this paper we present a novel speech coding algorithm for transmission over packet networks based on CELP (Code-Excited-Linear-Prediction). While various CELP type speech coders like the ITU-T G.729 or the Adaptive Multi-Rate (AMR) codec are applicable to packet voice communication systems such as the Internet, there are some fundamental reasons that may limit their Performance in case of packet loss. We investigate the alternative of embedded coding in the CELP framework. focussing on an efficient quantization of the excitation signal to generate a hierarchically-structured bit stream by means of pyramid coding [1]. Thus, the receiver can reconstruct speech at a basic quality by decoding only a subset of the entire bit stream. The quality of the decoded speech increases with the amount of received bits. We demonstrate the performance by an experimental hierarchical wideband (0.05–7 kHz) speech coder, applied to an IP (Internet Protocol) channel simulation.

The aim of the artificial bandwidth extension (BWE) of speech signals is to recover wideband speech from bandlimited speech. As the BWE algorithm is supposed to operate without additional side information on the original wideband speech, it has to exploit mutual dependencies between the available and missing frequency bands of the speech signal. In this paper the BWE is examined from an information theoretic perspective. After defining a performance measure, and introducing a few assumptions on a generalized BWE algorithm, a general relationship between mutual information and the maximum achievable estimation performance is formulated, which ensues an upper bound on the performance of BWE algorithms. Finally, some measurements considering a representative BWE scenario are presented.

This paper describes an eigendecomposition based approach to estimate the direction of arrival (DOA) of speech or noise sources in a reverberant room by using a microphone array. Unitary ESPRIT, a DOA estimation algorithm for narrowband signals, is applied to these wideband signals, which results in DOA estimates for each frequency. A single estimate is then obtained by means of a histogram. The estimated DOAs of the speech source are used to enhance a noisy speech signal by means of a superdirective beamformer. Experiments in a noisy reverberant room showed a higher estimation accuracy, and thus a better speech enhancement performance, at low SNRs compared to other DOA estimation techniques.

This paper presents and compares algorithms for combined acoustic echo cancellation and noise reduction for hands-free telephones. A structure is proposed, consisting of a conventional acoustic echo canceler and a frequency domain postfilter in the sending path of the hands-free system. The postfilter applies the spectral weighting technique and attenuates both the background noise and the residual echo which remains after imperfect echo cancellation. Two weighting rules for the postfilter are discussed. The first is a conventional one, known from noise reduction, which is extended to attenuate residual echo as well as noise. The second is a psychoacoustically motivated weighting rule. Both rules are evaluated and compared by instrumental and auditive tests. They succeed about equally well in attenuating the noise and the residual echo. In listening tests, however, the psychoacoustically motivated weighting rule is mostly preferred since it leads to more natural near end speech and to less annoying residual noise.

In digital mobile communications, efficient compression algorithms are needed to encode speech or audio signals. As the determined source parameters are highly sensitive to transmission errors, robust source and channel decoding schemes are required. This contribution deals with an iterative source-channel decoding approach where a simple channel decoder and a softbit-source decoder are concatenated. We mainly focus on softbit-source decoding which can be considered as an error concealment technique. This technique utilizes residual redundancy remaining after source coding. We derive a new formula that shows how the residual redundancy transforms into extrinsic information utilizable for iterative decoding. The derived formula opens several starting points for optimizations, e.g. it helps to find a robust index assignment. Furthermore, it allows the conclusion that softbit-source decoding is the limiting factor if applied to iterative decoding processes. Therefore, no significant gain will be obtainable by more than two iterations. This will be demonstrated by simulation.

This paper describes an adaptive multi-rate wideband (AMR-WB) speech codec proposed for the GSM system and also for the evolving third generation (3G) mobile speech services. The speech codec is based on SB-CELP (splitband-code-excited linear prediction) with five modes operating bit rates from 24 kbit/s down to 9.1 kbit/s. The respective channel coding schemes are based on RSC (recursive systematic code) and UEP (unequal error protection). Both, source and channel codec are designed as homogenous as possible to guarantee robust transmission on current and future mobile radio channels.

This contribution deals with a robust decoding approach of source codec parameters called softbit-source decoding. It can be considered as an error concealment technique which estimates codec parameters at the receiver utilizing the residual redundancy remaining after source coding and the channel decoder reliability information. We apply the turbo-principle to softbit-source decoding. We derive a formula that shows how different terms of residual redundancy can be transformed into independent extrinsic information utilizable in iterative processes. After introducing a new, iterative approach to softbit-source decoding, we analyze the influences of different bit mappings and of the number of iterations. Finally, the new approach is compared to conventional techniques with respect to performance and complexity.

In digital speech communication over noisy channels there is the need for reducing the subjective effects of residual bit errors which have not been eliminated by channel decoding. This task is usually called error concealment. We describe a new and generalizing approach to error concealment as part of a modified robust speech decoder. It can be applied to any speech codec standard and preserves bit exactness in the case of an error free channel. The proposed method requires bit reliability information provided by the demodulator or by the equalizer or specifically by the channel decoder and can exploit additionally a priori knowledge about codec parameters. We apply our algorithms to PCM, ADPCM, and GSM full-rate speech coding using AWGN, fading, and GSM channel models, respectively. It turns out that the speech quality is significantly enhanced, showing the desired inherent muting mechanism or graceful degradation behavior in the case of extreme adverse transmission conditions.

For speech transmission in digital land mobile telephony, effective compression algorithms have to be used to achieve a high bandwidth efficiency. Furthermore, a variety of adverse transmission effects make it necessary to employ powerful error control techniques to keep bit error rates tolerably low and thus to guarantee a high speech duality. Speech compression is designed to remove irrelevancy and redundancy from the speech signal. Yet measuring the statistical properties of speech parameters extracted by practical compression schemes shows that a considerable amount of redundancy still remains, either in terms of non-uniform distribution or due to time-correlation of parameters extracted from subsequent speech segments. In this contribution, we propose a new minimum mean square error (MMSE) decoder for block-oriented trellis codes, that is able to exploit the time-correlation of subsequent parameter sets. The decoder yields non-discrete speech parameter mean square (MS) estimates. Thus it combines two approaches to exploit residual redundancy: source controlled channel decoding (SCCD) (Hagenauer 1995) and soft bit source decoding (SBSD) (Fingscheidt and Vary 1997) in one algorithm.

This paper describes a combined adaptive transform codec (ATC) and code-excited linear prediction (CELP) algorithm, called ATCELP, for the compression of wideband (7 kHz) signals. The CELP algorithm applies mainly to speech, whereas the ATC mode is selected for music and noise signals. We propose a switching scheme between CELP and ATC mode and describe a frame erasure concealment technique. Subjective listening tests have shown that the ATCELP codec at bit rates of 16, 24 and 32 kbit/s achieved performances close to those of the CCITT G.722 at 48, 56 and 64 kbit/s, respectively, at most operating conditions.

We address the problem of acoustic echo cancellation and noise reduction for narrow and wide band telephone applications. We combine a conventional echo canceller with a postfilter implemented in the frequency domain and derive an algorithm for the simultaneous attenuation of residual echo and noise. The main goals are a low level natural sounding background noise without artifacts such as musical tones, and an inaudible residual echo. This is achieved by considering the masking properties of the human auditory system. Simulation results verify that these goals are reached, while the distortion of the near end speech is comparable to conventional algorithms.

We propose a new 6.1 to 13.3-kb/s speech codec called variable rate code-excited linear prediction (VR-CELP) for adaptive multi-rate (AMR) transmission over mobile radio channels such as GSM or UMTS. The AMR concept allows to operate with almost wireline speech quality for poor channel conditions and better quality for good channel conditions. This is achieved by dynamically splitting the gross bit rate of the transmission system between source and channel coding according to the current channel conditions. Thus the source coding scheme must be designed for seamless switching between rates without annoying artifacts. To enhance the transmission quality under very poor channel conditions, a new powerful error concealment strategy based on estimation theory is applied.

Digital speech transmission systems use source coding to reduce the bit rate and channel coding to correct transmission errors. Furthermore, in periods of a very poor channel quality error concealment of residual bit errors becomes necessary as channel decoding fails. However, if the channel is clear, channel coding would not be required at all and the speech quality could be improved by allowing a higher bit rate for source encoding. Usually a compromise is taken between speech quality in case of a clear channel and error robustness in case of poor channel quality. This paper addresses the problem of a joint optimization of error concealment and source/channel coding. Under the premise of a minimum mean square error criterion for signal reconstruction it turns out that error concealment instead of error correction may be the best choice if source coding leaves sufficient residual parameter correlations by less bit rate reduction.

In this paper we propose an algorithm for reduction of noise in audio signals. In contrast to several previous approaches we do not try to achieve a complete removal of the noise, but instead our goal is to preserve a pre-defined amount of the original noise in the processed signal. This is accomplished by exploiting the masking properties of the human auditory system. The speech and noise distortions are considered separately. The spectral weighting rule, adapted by utilizing only estimates of the masking threshold and the noise power spectral density, has been designed to guarantee complete masking of distortions of the residual noise. Simulation results confirm that no audible artifacts are left in the processed signal, while speech distortions are comparable to those caused by conventional noise reduction techniques.

In digital mobile communication systems there is the need for reducing the subjective effects of residual bit errors which have not been eliminated by channel decoding by the use of error concealment techniques. Due to the fact that most standards do not specify these algorithms bit exactly, there is room for new solutions to improve the speech quality. This article develops a new approach for optimum estimation of the speech codec parameters. It can be applied to any speech codec standard if bit reliability information is provided by the demodulator (e.g. DECT), or by the channel decoder (e.g. soft-output Viterbi algorithm-SOVA in GSM). The proposed method includes an inherent muting mechanism leading to a graceful degradation of speech quality in case of adverse transmission conditions. Particularly the additional exploitation of the residual source redundancy, i.e. some a priori knowledge about the codec parameters gives a significant enhancement of the output speech quality. In the case of an error free channel, bit exactness as required by the standards can be preserved.

In 1991 a digital mobile radio system will be introduced in Europe. The speech codec to be used as the standard is presented. The coding scheme which has been selected by the CEPT Groupe-Speciale-Mobile (GSM) as a result of formal subjective listening tests, is based on the regular-pulse excitation LPC technique (RPE-LPC) combined with long-term prediction (LTP). The so-called RPE-LTP codec has a net bit rate of 13 kbit/s. The algorithm and the experimental implementations based on different VLSI signal processors are described and demonstrated.>