Stanisław Kacprzak

Speech/music discrimination via energy density analysis

In this paper we suggest to apply a new feature, called Minimum Energy Density (MED), in discrimination of audio signals between speech and music. Our method is based on the analysis of local energy for 1 or 2.5 seconds audio signals. An elementary analysis of the probability for the power distribution is an effective tool supporting the decision making system. We compare our feature with Percentage of Low Energy Frames (LEF), Modified Low Energy Ratio (MLER) and examine their efficiency for two separate speech/music corpora.

Spoken language clustering in the i-vectors space

This paper presents the results of language clustering in the i-vectors space, a method to determine in an unsupervised manner how many languages are in a data set and which recordings contain the same language. The most dense i-vectors clusters are found using the DBSCAN algorithm in a low dimensional space obtained by the t-SNE method. Quality of clustering for spherical k-means and the proposed method are tested with the data from NIST 2015 i-Vector Challenge. Usefulness of obtained clustering is tested in the challenge evaluation system. The results demonstrate that the proposed method allows to find 109 dense clusters with low impurity for 50 target languages.

Speech/music discrimination for analysis of radio stations

A computationally efficient feature, called Minimum Energy Density (MED) was applied to discriminate audio signals between speech and music in the radio stations programs. The presented binary classifier is based on testing two features: energy distribution and differences between energy in channels. We analyzed 240 hours of signals, from 10 Polish radio stations. Our analysis enables us to provide information about content of particular radio stations.

On the extraction of early reflection signals for automatic speech recognition

Room reverberation caused by multipath sound wave propagation in acoustic enclosures constitutes an unwanted distortion for automatic speech recognition systems. Multichannel speech enhancement methods often aim to enhance the signal impinging at the microphone array from the source direction while reducing late reverberation. In this paper, we investigate the applicability of spatial filters which constructively combine the direct-path signal with distinct early room reflection signals to increase the direct-to-reverberation ratio and to reduce the word error rate (WER) of automatic speech recognition systems. We present suitable filters and compare them with existing approaches. Results for the simulated acoustic environments indicate that an improvement in WER can indeed be achieved by the spatial filters which account for strong early reflections.

Automatic extraction and clustering of phones

The automatic segmentation and parametrization based on the frequency analysis was used to compare with manually annotated phones. The phones boundaries were fixed in places of relatively large changes in the energy distribution between the frequency bands. Frequency parametrization and clustering enabled the division of phones into groups (clusters) according to their acoustic similarities. The results of performed experiments showed that analysis of the frequency properties only, results in correct segmentation but accuracy of recognition was about 20% only.