Stanislaw Gorlow

Informed source separation: Underdetermined source signal recovery from an instantaneous stereo mixture

The present paper exposes a new technique that aims at solving an ill-posed source separation problem encountered in stereo mixtures. The proposed method is realized in an encoder-decoder framework: On the encoder side, a set of spectral envelopes is extracted from the original tracks, which are known. These envelopes are passed on to the decoder in attachment to the stereo mixture, whereas the frequency resolution of the former is adapted to the critical bands, and their magnitude is logarithmically quantized. On the decoder side, the mixture signal is decomposed by time-frequency selective iterative spatial filtering guided by a source activity index, which is derived from the spectral envelope values. A comparison with a similar algorithm reveals that the novel approach yields a higher perceptual audio quality at a much lower data rate.

Informed Audio Source Separation Using Linearly Constrained Spatial Filters

In this work we readdress the issue of audio source separation in an informed scenario, where certain information about the sound sources is embedded into their mixture as an imperceptible watermark. In doing so, we provide a description of an improved algorithm that follows the linearly constrained minimum-variance filtering approach in the subband domain, in order to obtain perceptually better estimates of the source signals in comparison to other published approaches. Just as its predecessor, the algorithm does not impose any restrictions on the number of simultaneously active sources, neither on their spectral overlap. It rather adapts to a given signal constellation and provides the best possible estimates under given constraints in linearithmic time. The validity of the approach is demonstrated on a stereo mixture with two levels of sound complexity. It is also shown by means of both objective and subjective evaluation that the proposed algorithm outperforms a reference algorithm by at least one grade. Bearing high perceptual resemblance to the original signals at a fairly tolerable data rate of 10-20 kbps per source, the algorithm hence seems well-suited for active listening applications such as re-mixing or re-spatialization in real time.

Model-Based Inversion of Dynamic Range Compression

In this work it is shown how a dynamic nonlinear time-variant operator, such as a dynamic range compressor, can be inverted using an explicit signal model. By knowing the model parameters that were used for compression one is able to recover the original uncompressed signal from a “broadcast” signal with high numerical accuracy and very low computational complexity. A compressor-decompressor scheme is worked out and described in detail. The approach is evaluated on real-world audio material with great success.

Informed separation of spatial images of stereo music recordings using second-order statistics

In this work we address a reverse audio engineering problem, i.e. the separation of stereo tracks of professionally produced music recordings. More precisely, we apply a spatial filtering approach with a quadratic constraint using an explicit source-image-mixture model. The model parameters are “learned” from a given set of original stereo tracks, reduced in size and used afterwards to demix the desired tracks in best possible quality from a preexisting mixture. Our approach implicates a side-information rate of 10 kbps per source or channel and has a low computational complexity. The results obtained for the SiSEC 2013 dataset are intended to be used as reference for comparison with unpublished approaches.

Multichannel object-based audio coding with controllable quality

In this paper a new multichannel object-based audio coding scheme with scalable signal quality is proposed. The novel scheme is based on controlled downmixing and demixing. By means of a dedicated control mechanism, a number of distinct audio objects are mixed into a lower number of channels. The latter is chosen such that the desired quality level is met after demixing. The quality is assessed with two new psychoacoustically motivated metrics. Following the informed source separation approach, the downmix is decomposed via optimum spatial filtering guided by short-time power spectral densities of the audio objects. In an experiment it is shown that the raw data rate of an exemplary 10-track recording can be reduced by at least 30 % using linear pulse-code modulation while maintaining perceptual transparency.

Restoring the dynamics of clipped audio material by inversion of dynamic range compression

In this work, a novel approach for the restoration of clipped audio alias declipping is presented. It is based on the inversion of a nonlinear dynamic system varying over time. The inverse system is parametrized according to a brickwall limiter. The threshold and the makeup gain are then adjusted in such a manner that the desired effect i.e. the accentuation of transients or peaks is observed at the output. The validity of the approach is confirmed in a formal listening test, in which a performance on a par with the state of the art is achieved. The application of the approach is straight forward and the effect can be tuned to meet an objective criterion, such as a sufficiently high peak-to-average power ratio or crest factor.