[PDF] Performance measurement in blind audio source separation | Semantic Scholar (2024)

A system level design for blind source separation to separate mixture automatically by combining the binary time-frequency mask and spectral subtraction method, which shows that the proposed algorithm can provide high performance on audio source separation algorithm.

A new MUSHRA-based subjective test protocol is proposed to assess the perceived quality with respect to each kind of distortion and collect the scores of 20 subjects over 80 sounds, and a family of objective measures aiming to predict the subjective scores based on a decomposition of the estimation error into several distortion components are proposed.

A new adaptive approach is presented to evaluate different blind source separation algorithms with three new evaluation metrics, the minimum number of samples required for a successful separation process, and the time needed to complete the separation process.

Compared to the reference MPEG-4 Spatial Audio Object Coding system, the proposedISS system pro- vides much cleaner separated speech signals (consistently 10- 20 dB higher Signal to Interference Ratios), revealing strong potential for audio conference applications.

A new blind source separation technique, separating more than two sources from a stereo recording mixture based on time delay of arrival (TDOA), and instants of significant excitation (GCI's) of each speaker is presented.

The tasks of speech source localization, source counting and source separation are addressed for an unknown number of sources in a stereo recording scenario and an approach for separating the sources based on Maximum Likelihood (ML) estimation is proposed.

An evaluation of single-channel blind audio source separation (BASS) algorithms for different musical instrument combinations, based on statistical approach, with main difference in the evaluation of average deviations from ground truth, instead of evaluating absolute values of signal-to-distortion ratio.

The goal of this review is to make clear the modeling assumptions and performance limitations of each approach, emphasize the importance of audio-specific issues in the design of BASS algorithms, and clarify the modeling assumptions and performance limitations of each approach.

This paper presents a novel time-frequency domain audio-visual speech separation method, RTFS-Net, which applies its algorithms on the complex time-frequency bins yielded by the Short-Time Fourier Transform and introduces a unique attention-based fusion technique for the efficient integration of audio and visual information.

...

...

When the sources are estimated from a degenerate set of mixtures by applying a demixing matrix, it is proved that there are upper bounds on the achievable Source to Interference Ratio.

A unified methodology of evaluating BSS algorithms along with providing data online such that researches can compare their results is provided.

This work suggests a two-stage separation process: a priori selection of a possibly overcomplete signal dictionary in which the sources are assumed to be sparsely representable, followed by unmixing the sources by exploiting the their sparse representability.

It is shown that Belouchrani and Amin's technique can be interpreted as a practical implementation of the general equations provided in the stochastic case, and a new criterion aimed at selecting more efficiently the time-frequency locations where the spatial matrices should be joint-diagonalized is introduced, introducing single autoterms selection.

A typology of BASS tasks would greatly help the building of an evaluation framework and some qualitative criteria to evaluate separation in each case are proposed.

A novel method for blind separation of any number of sources using only two mixtures when sources are (W-)disjoint orthogonal, that is, when the supports of the (windowed) Fourier transform of any two signals in the mixture are disjoint sets.

The experimental results reveal that the signal separation performance of the proposed SIMO-ICA is the same as that of the conventional ICA-based method, and that the spatial quality of the separated sound in SIMO -ICA is remarkably superior to that ofThe conventional method.

The experimental results reveal that the signal separation performance of the proposed SIMO-ICA is the same as that of the conventional ICA-based method, and the spatial quality of the separated sound is remarkably superior to the conventional method, particularly for the fidelity of the sound reproduction.

A method of sparse decomposition of stereo audio signals is developed, and its application to blind separation of more than two sources from only two linear mixtures is tested.

...

...