Renat Vafin

Improved modeling of audio signals by modifying transient locations

We propose a method for obtaining an improved representation of transients in audio signals. The representation is based on a damped sinusoidal model. To improve the representation, transient locations are modified in such a way that a transient can start only at the beginning of a sinusoidal segment. The introduced modifications facilitate a reduction of the number of damped sinusoids needed to model a transient well and eliminate pre-echo artifacts. We verify with a listening test that the modifications do not result in a perceptual difference between the original and modified audio signals.

Flexible sum-difference stereo coding based on time-aligned signal components

A framework for flexible and efficient coding of general stereo audio signals is proposed. Methods based on the framework can be used together with an arbitrary single channel (mono) coder to achieve seamless transition from pure parametric stereo coding to waveform approximating coding as the bitrate is increased. The idea, based on sum-difference encoding of time-aligned signal components, is presented as a general framework. An example implementation is demonstrated to have the desired convergence properties towards transparent quality.

Sinusoidal modeling of audio and speech using psychoacoustic-adaptive matching pursuits

We propose a segment-based matching pursuit algorithm where the psychoacoustical properties of the human auditory system are taken into account. Rather than scaling the dictionary elements according to auditory perception, we define a psychoacoustic-adaptive norm on the signal space which can be used for assigning the dictionary elements to the individual segments in a rate-distortion optimal manner. The new algorithm is asymptotically equal to signal-to-mask ratio based algorithms in the limit of infinite analysis window length. However, the new algorithm provides a significantly improved selection of the dictionary elements for finite window length.

Entropy-constrained polar quantization: theory and an application to audio coding

In this work, we develop entropy-constrained unrestricted polar quantizers, where phase quantization depends on the input amplitude. Formulas for amplitude and phase quantization point densities are derived under high-rate assumptions. It is shown that the mean-squared distortion is decreased considerably as compared to strictly polar quantization and approaches that of scalar rectangular quantization asymptotically with increasing rate. The unrestricted polar quantization is generalized to include a weighted error measure, such that it accounts for masking effects of the human auditory system. Both amplitude and phase quantization depend on the perceptual importance of sinusoids. The new method is applied to a sinusoidal audio coder, and is shown to outperform a conventional sinusoidal quantization method where the number of phase quantization bits is the same for all audible sinusoids.

Jointly optimal quantization of parameters in sinusoidal audio coding

In this paper, we present new methods for jointly optimal scalar quantization of parameters in sinusoidal audio coding. The sinusoidal amplitudes, frequencies, and phases used to represent (the tonal component of) an audio signal are quantized such as to minimize a single-letter perceptual distortion measure under a given bit-rate constraint. The asymptotically optimal scalar quantizers are derived analytically using high-rate theory. We present different practical methods, where quantization accuracy of some parameters is made dependent on quantized values of other parameters. The choice of method depends on the availability of the frequency-masking threshold, defining the perceptual importance of the individual sinusoids, at the decoder. The high performance of our quantizers is demonstrated by means of a listening test

Towards optimal quantization in multistage audio coding

We develop a new method for quantization in multistage audio coding. We consider the case of a two-stage sinusoidal/waveform coder. Given a distortion measure and a bit-rate constraint, we analytically derive the optimal rate distribution between subcoders (stages) and the corresponding optimal quantizers, which allows the coder to adapt easily to changes in bit-rate requirements. We verify that the performance, both in terms of signal-to-noise ratio (SNR) and perceptual quality, is higher if the input to the second stage is obtained by subtracting the quantized first-stage reconstruction from the original signal, as opposed to subtracting the unquantized reconstruction.

Rate-distortion optimal high-resolution differential quantisation for sinusoidal coding of audio and speech

Sinusoidal coding plays an important role in low-rate audio coding. Typically, (time/frequency) differential techniques are employed to reduce the bit rate for representing the sinusoidal components. In this paper we derive optimal entropy-constrained differential quantisers for quantising the sinusoid parameters. More specifically, the quantisers minimise a perceptually relevant distortion measure while the corresponding quantisation indices satisfy an entropy constraint. The quantisers turn out to be flexible and of low complexity. Subjective evaluations with audio signals suggest a bit-rate reduction as high as 20% with the derived quantisers over state-of-the-art (logarithmic) quantisers.

Multi-variate block polar quantization and an application to audio

We introduce multi-variate block polar quantization (MBPQ). MBPQ minimizes a weighted distortion for a set of complex variables representing one block of a signal under a resolution constraint for the entire block. MBPQ allows for different probability distributions in different dimensions of the set of complex variables. It outperforms a block polar quantizer introduced earlier (Pobloth, H. et al., Proc. Eurospeech, p.1097-100, 2003), and unrestricted polar quantization (UPQ), for both Gaussian complex variables and sinusoids found from audio data. In the case of audio data, we found a performance gain of about 2.5 dB over the best performing conventional resolution-constrained polar quantization, UPQ.