Psychoacoustic Model Research Articles

Multiband subspace tracking speech enhancement for in-car human computer speech interaction

In this paper, a new subspace-based speech enhancement algorithm for in-car human computer speech interaction is presented. We first incorporate a perceptual filterbank which is derived from psycho-acoustic model with signal subspace approach to effectively suppress in-car noises of engine. Second, for real-time applications, a new subspace tracking algorithm is derived by modifying PASTd algorithm to solve the data dependent hazard of tacking algorithm. Six different types of in-car noises in TAICAR database are used in our evaluation. The experimental results demonstrate that our approach is superior to conventional subspace and spectral subtraction methods.

Cues for masked amplitude-modulation detection

The ability of psychoacoustic models to predict listeners' performance depends on two key stages: preprocessing and the generation of a decision variable. The goal of the current study was to determine the perceptually relevant decision variables in masked amplitude-modulation detection tasks in which the modulation depth of the masker was systematically varied. Potential cues were made unreliable by roving the overall modulation depth from trial to trial or were reduced in salience by equalizing the envelope energy of the standard and target after the signal was added. Listeners' performance was significantly degraded in both paradigms compared to the baseline (fixed-level modulation masker) condition, which was similar to those used in previous studies of masking in the envelope-frequency domain. Although this observation was broadly consistent with a simple long-term envelope power-spectrum model, there were several aspects of the data that were not. For example, the steep rate of change in threshold with masker depth and the fact that an optimal amount of envelope noise could enhance performance were not predicted by decision variables calculated directly from the stimulus envelope. A physiologically based processing model suggested a realistic nonlinear mechanism that could give rise to these second-order features of the data.

Predicting changes in breathiness resulting from variations in aspiration noise and open quotient using a loudness model

Previous studies have reported that increases in high-frequency aspiration noise (AH) and open quotient (OQ) are correlated with an increase in the perception of breathiness in vowels [Klatt and Klatt, J. Acoust. Soc. Am. 87, 820–857 (1990)]. However, a system based on psychoacoustic loudness models has been shown to predict the level of breathiness in vowels to a greater extent than AH and OQ themselves [Shrivastav and Sapienza, J. Acoust. Soc. Am. 114, 2217–2224 (2003)]. The goal of this experiment was to verify the significance of AH and OQ in the perception of breathiness and to determine how each of these influences the output of the loudness model, individually and in combination. Ten vowels, five male and five female, were synthesized using a Klatt synthesizer. Each vowel was used to generate three continua varying in terms of their OQ and/or AH. Ten listeners rated breathiness for each continuum using a magnitude estimation task. Results show that normalized partial loudness as well as the loudness of aspiration noise predicted changes in breathiness resulting from an increase in AH. Increase in OQ resulted in very small changes in perceived breathiness, resulting in poor correlation with loudness measures. [Research supported by NIH/R21DC006690.]

Robust and high-quality time-domain audio watermarking based on low-frequency amplitude modification

This work proposes a method of embedding digital watermarks into audio signals in the time domain. The proposed algorithm exploits differential average-of-absolute-amplitude relations within each group of audio samples to represent one-bit information. The principle of low-frequency amplitude modification is employed to scale amplitudes in a group manner (unlike the sample-by-sample manner as used in pseudonoise or spread-spectrum techniques) in selected sections of samples so that the time-domain waveform envelope can be almost preserved. Besides, when the frequency-domain characteristics of the watermark signal are controlled by applying absolute hearing thresholds in the psychoacoustic model, the distortion associated with watermarking is hardly perceivable by human ears. The watermark can be blindly extracted without knowledge of the original signal. Subjective and objective tests reveal that the proposed watermarking scheme maintains high audio quality and is simultaneously highly robust to pirate attacks, including MP3 compression, low-pass filtering, amplitude scaling, time scaling, digital-to-analog/analog-to-digital reacquisition, cropping, sampling rate change, and bit resolution transformation. Security of embedded watermarks is enhanced by adopting unequal section lengths determined by a secret key.

Sound intelligibility enhancement using a psychoacoustic model and an oversampled filterbank

A sound intelligibility enhancement (SIE) system is disclosed. The SIE system uses a psychoacoustic model and preferably an oversampled filterbank wherein the level of a signal-of-interest that falls below the environmental noise is selectively amplified as a function of the input level and frequency so that it is audible above the noise but never exceeds a predetermined maximum output level as a function of frequency. The SIE system can be combined with active noise cancellation.

Commentary on Cook & Fujisawa's "The Psychophysics of Harmony Perception: Harmony is a Three-Tone Phenomenon"

Cook & Fujisawa (2006) point out that, contrary to the predictions of psychoacoustic models, the diminished triad is more consonant and prevalent in western tonal music than the augmented. A possible simple explanation is that the diminished triad often functions as an incomplete dominant (major-minor) seventh chord, the most prevalent tetrad in mainstream tonal music.

Application of psychoacoustic filtering for machine fault detection

A method for nondestructive machine fault detection based on the evaluation of acoustic machine signatures is proposed. Various mechanical defects of rotary machines can be reflected in altered acoustic signatures. The proposed method is based on the psychoacoustic modelling of human auditory perception. For the purpose of machine fault detection, the gammatone filterbank is applied in the preprocessing of acoustic signals. Filtered signals are then rectified and features are calculated as mean values of rectified signals. A set of features represents an extracted machine signature. The machine states are evaluated by comparing the extracted signature with the database of previously collected machine signatures. The proposed approach is illustrated by a case study where the quality of commercially produced compressors is estimated. Results show that major faults that occur in a production can be reliably detected.

On the perceptual performance limitations of echo cancellers in wideband telephony

In this paper, standard echo canceller performance measures are evaluated in terms of psychoacoustic aspects of human hearing. The focus is on wideband speech communications systems with long round-trip delays of 200 ms and up present in the transmission path. The results of a simple acoustic echo cancellation experiment are analyzed with a standard psychoacoustic model, revealing that steady-state echo return loss enhancement and mean square error cannot be used to determine whether residual echo is perceivable in the presence of background noise. In addition, a simple modification to the normalized least mean square (NLMS) algorithm is introduced by adding a perceptual preemphasis filter. Simulation results and listening tests show that it is possible to improve the perceived performance of an echo canceller during convergence by placing greater emphasis on frequencies at which the human auditory system is most sensitive.

A Perceptual Model for Sinusoidal Audio Coding Based on Spectral Integration

Psychoacoustical models have been used extensively within audio coding applications over the past decades. Recently, parametric coding techniques have been applied to general audio and this has created the need for a psychoacoustical model that is specifically suited for sinusoidal modelling of audio signals. In this paper, we present a new perceptual model that predicts masked thresholds for sinusoidal distortions. The model relies on signal detection theory and incorporates more recent insights about spectral and temporal integration in auditory masking. As a consequence, the model is able to predict the distortion detectability. In fact, the distortion detectability defines a (perceptually relevant) norm on the underlying signal space which is beneficial for optimisation algorithms such as rate-distortion optimisation or linear predictive coding. We evaluate the merits of the model by combining it with a sinusoidal extraction method and compare the results with those obtained with the ISO MPEG-1 Layer I-II recommended model. Listening tests show a clear preference for the new model. More specifically, the model presented here leads to a reduction of more than 20% in terms of number of sinusoids needed to represent signals at a given quality level.

Multiresolution sinusoidal model with dynamic segmentation for timescale modification of polyphonic audio signals

In this paper, we propose an efficient sinusoidal model of polyphonic audio signals especially good for the application of timescale modification. One of the critical problem of sinusoidal modeling is that the signal is smeared during the synthesis frame, which is a very undesirable effect for transient parts. We solve this problem by introducing multiresolution analysis-synthesis and dynamic segmentation methods. A signal is modeled with a sinusoidal component and a noise component. A multiresolution filter bank is applied to an input signal which splits it into octave-spaced subbands without causing aliasing and then sinusoidal analysis is applied to each subband signal. To alleviate smearing of transients during synthesis, a dynamic segmentation method is applied to the subband signals that determines the optimal analysis-synthesis frame size adaptively to fit its time-frequency characteristics. To extract sinusoidal components and calculate respective parameters, a matching pursuit algorithm is applied to each analysis frame of the subband signal. A psychoacoustic model implementing frequency masking is incorporated with matching pursuit to provide a reasonable stop condition of iteration and reduce the number of sinusoids. The noise component obtained by subtracting the synthesized signal with sinusoidal components from the original signal is modeled by a line-segment model of short time spectrum envelope. For various polyphonic audio signals, the results of simulation shows the proposed sinusoidal modeling can synthesize original signals without loss of perceptual quality and do more robust and high-quality timescale modification for large scale factors.

Perceptually motivated time-frequency analysis

This paper describes the design of a bilinear time-frequency distribution which is a joint model of temporal and spectral masking. The distribution is used to generate temporally evolving excitation patterns of nonstationary signals and systems and is conceived as a tool for acousticians and engineers for perceptual time-frequency analysis. Distribution time and frequency resolutions are controlled by a separable kernel consisting of a set of low-pass time and frequency smoothing windows. These windows are designed by adapting existing psychoacoustic models of auditory resolution, rather than using mathematical window functions. Cross-term interference and windowing clutter are highly suppressed for the distribution, ensuring resolution accuracy over a dynamic range sufficient to encompass that of the auditory system (in excess of 100 dB). Application to the analysis of a synthetic and two real signals are included to demonstrate the approach.

Single-Channel Noise Reduction for Multiple Background Noises Using Perceptual Wavelet Packet Transform and Fuzzy Logic

Background noises interfere with communication devices such as mobile telephone, digital hearing aid, etc. Therefore noise reduction (NR) part for limiting the effect of these noises is important. The paper proposes a noise reduction method based on the soft decision-making by the fuzzy inference system (FIS). The different characteristics of noises frequently occurring are used for creating the fuzzy decision rule base of the FIS. The FIS have two input parameters: the average energy and the difference of the average energy. The analysis of the FIS is done in the domain of the perceptual wavelet packet transform (PWPT) that is the human’s psychoacoustic model. The output of the FIS is used to modify the PWPT coefficients in such a way that it is more likely that the noise components are reduced while the speech signal is enhanced. The enhanced speech signal is the result of the inverse perceptual wavelet packet transform (IPWPT) of the modified coefficients. The experiment results show that the proposed method gives lower distortion than do the conventional methods especially when the input signal-to-noise ratio (SNR) is low; e.g. at SRN at 0dB the proposed method improves the output SNR level up to 4.18dB.

Low Complexity Scalable Perceptual Audio Coder using an Optimum Wavelet Packet basis Representation and Vector Quantization

In this paper, we describe a high quality low complexity scalable audio coding scheme, using an optimum wavelet packet (WP) basis signal representation based on the time varying characteristics of the audio signal. In ISO/MPEG audio coding standards [1–3], resolution of decomposition filterbank (uniform) does not match with the resolution of psychoacoustic model (which requires more resolution and needs to be matched with the critical bands (non uniform) of cochlea). Hence MPEG coder uses a separate high resolution decomposition filterbank for psychoacoustic model implementation, which increases the computational load of the coder. Here, we use a wavelet packet decomposition structure closely matching to the critical bands [4,5] of human auditory system, to transform the data into wavelet domain and then these wavelet packet coefficients are used to drive the psychoacoustic model directly. Hence, psychoacoustic model design is integrated with the design of decomposition filterbank. Other features of the proposed coder are scalability (can support three standard industrial sampling frequencies 11.025 kHz, 22.050 kHz and 44.1 kHz) and optimum wavelet basis selection from a predefined library of wavelet bases, by extracting seven statistical features of the audio signal to be encoded. A new Vector Quantization (VQ) scheme is also proposed here, in which the length of the code book can be varied in accordance with the psychoacoustic model requirement. Experimental results show that the proposed coder yields almost transparent quality with compression ratios in the range of e to 10.

A low complexity design of psycho-acoustic model for MPEG-2/4 advanced audio coding

The paper presents a new low complexity design of psycho-acoustic model (PAM), which is the key technology for MPEG-2/4 advanced audio coding (AAC) encoding. The real-time constraint of MPEG AAC leads to a heavy computational bottleneck on today's portable devices. To overcome this problem, design analysis and optimization of PAM are addressed at the algorithmic level. First, the dominant calculation of spreading function is replaced with small look-up tables. Second, modified-discrete-cosine-transform-based (MDCT-based) PAM with block type decision in frequency domain is proposed to reduce the overall AAC encoder complexity and maintain the quality. This technique is different from the traditional methods. The computational complexity of PAM is reduced by 82% in total. The proposed design could be implemented in a real-time MPEG-2/4 AAC stereo encoder at low complexity profile and bit rate 128 kb/s with computational complexity of less than 20 MOPS while maintaining CD quality.

Perceptual audio modeling with exponentially damped sinusoids

This paper presents the derivation of a new perceptual model that represents speech and audio signals by a sum of exponentially damped sinusoids. Compared to a traditional sinusoidal model, the exponential sinusoidal model (ESM) is better suited to model transient segments that are readily found in audio signals. Total least squares (TLS) algorithms are applied for the automatic extraction of the modeling parameters in the ESM, i.e. the amplitude, phase, frequency and damping factors of a user-defined number of damped sinusoids. In order to turn the SNR optimization criterion of these TLS algorithms into a perceptual modeling strategy, we use the psycho-acoustic model of MPEG-1 Layer 1 in a subband TLS-ESM scheme. This allows us to model each subband signal in accordance with its perceptual relevance, thereby lowering the number of required modeling components for a given modeling quality. Simulations and listening tests confirm that perceptual ESM achieves the same perceived quality as plain ESM while using substantially less components, and provide support for applying the new model in the fields of parametric audio processing and coding.

A new speech compression method

In this paper a new speech compression method is presented. The traditional speech compression method is based on linear prediction. The compression method, proposed in this paper, is based on the use of an orthogonal transform, the discrete cosine packets transform. This method is well suited for the speech processing, taking into account the sine model of this kind of signals and because this transform converges asymptotically to the Karhunen-Lofleve transform. After the computation of the discrete cosine packets transform, the coefficients obtained are processed with a threshold detector, who keeps only the coefficients superior to a given threshold. This way the number of non zero coefficients is reduced doing the compression. The next block of the compression system is the quantization system. This is build following the speech psycho-acoustic model. The proposed compression method is transparent, the compression rate obtained is important and the operations number and the memory volume used are not very high.

Dual channel phase flag determination for coupling bands in a transform coder for high quality audio

A method and apparatus for subband phase flag determination for coupling of channels in a dual channel audio encoder is based on a psychoacoustic model of the human auditory system. The method and apparatus are applicable to audio encoders which utilize a coupling channel to combine certain frequency components of the input audio signals. The method ensures a least square error between the original channel frequency coefficients at the encoder and the estimated coefficients at the decoder by determining the sign of the dot product of the coefficients for one of the channels and the coupling coefficients. No restriction is placed on the strategy utilized for generating the coupling channel coefficients or the coupling coordinates.

The design of bells with harmonic overtones.

Musical bells have had limited application due to the presence of inharmonic partials in the lower part of their acoustic spectra. A series of bells has been designed that contains up to seven partial frequencies in the harmonic series beginning at the fundamental frequency. This was achieved by choosing geometries for finite-element analysis models in which as many purely circumferential bending modes as possible occurred at frequencies below any mode with an axial ring node. The bell models were then fine tuned using gradient projection method shape optimization and the resulting profiles were cast in silicon bronze. A range of bell geometries and timbres is analyzed using psycho-acoustic models and is discussed in relation to European carillon bells.

Digital watermarks for audio signal based on psychoacoustic masking model

AbstractThis paper describes one digital watermarking method for audio signals using the MPEG psychoacoustic model. It is important that a digital watermark be perceptually inaudible to the human auditory system. One watermark algorithm that takes this into account is the digital watermark algorithm using the MPEG psychoacoustic model proposed by Boney and colleagues. However, this method has several drawbacks such as it is susceptible to MPEG encoding and there is no synchronized detection measure. As a result, several improvements were made to increase the robustness of the digital watermark to MPEG encoding, and information from psychoacoustic tests was also used to introduce successive masking. These improvements enabled a watermark to be implemented that is even robust to MPEG encoding. To ensure robustness to D‐A/A‐D conversions or cropping attacks, a whitened cross‐correlation method (cross‐power spectrum phase) was used to implement synchronized detection. This enabled a watermark that is robust to D‐A/A‐D conversions to be implemented. This watermark was also shown to be robust to attacks such as noise addition, filtering, and downsampling. In addition, this paper shows the results of subjective evaluation experiments, which indicate that there is little quality deterioration due to watermarking. © 2003 Wiley Periodicals, Inc. Electron Comm Jpn Pt 3, 86(12): 65–75, 2003; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/ecjc.10143

The SOM-enhanced JukeBox: Organization and Visualization of Music Collections Based on Perceptual Models

The availability of large music repositories calls for new ways of automatically organizing and accessing them. While artist-based listings or title indexes may help in locating a specific piece of music, a more intuitive, genre-based organization is required to allow users to browse an archive and explore its contents. So far, however, these organizations following musical styles have to be designed manually. With the SOM-enhanced JukeBox (SOMeJB) we propose an approach to automatically create an organization of music archives following their perceived sound similarity. More specifically, characteristics of frequency spectra are extracted and transformed according to psychoacoustic models. The resulting psychoacoustic Rhythm Patterns are further organized using the Growing Hierarchical Self-Organizing Map, an unsupervised neural network. On top of this advanced visualizations including Islands of Music (IoM) and Weather Charts offer an interface for interactive exploration of large music repositories.