• View
• Edit
• History
• Print

Exprtest

ANALYSIS OF EMOTIONAL SPEECH USING AN ADAPTIVE SINUSOIDAL MODEL

George P. Kafentzis, Theodora Yakoumaki, Athanasios Mouchtaris, Yannis Stylianou

Abstract - Processing of emotional (or expressive) speech has gained attention over recent years in the speech community due to its numerous applications. In this paper, an adaptive sinusoidal model (aSM), dubbed extended adaptive Quasi-Harmonic Model - eaQHM, is employed to analyze emotional speech in accurate, robust, continuous, timevarying parameters (amplitude, frequency, and phase). It is shown that these parameters can adequately and accurately represent emotional speech content. Using a well known database of narrowband expressive speech (SUSAS) we show that very high Signal-to- Reconstruction-Error Ratio (SRER) values can be obtained, compared to the standard sinusoidal model (SM). Formal listening tests on a smaller wideband speech database show that the eaQHM outperforms SM from a perceptual resynthesis quality point of view. Finally, preliminary emotion classification tests show that the parameters obtained from the adaptive model lead to a higher classification score, compared to the standard SM parameters.

Thank you for your time !

In this test, the goal is to evaluate the perceptual quality between recordings of expressive speech and their reconstruction by several signal processing algorithms.

You will listen to an original and a few reconstructed speech waveforms. You will rate them according to the perceptual quality compared to the original. Your criteria should include presence of artifacts and quality of the sound.

As a guide for the test, at first you can play the original sound as many times as you want (labeled Original). Then, you can play the other sounds (labeled Resynthesis 1,2,3) one at a time, and then again the original (if you want). Then you can rate the quality of the resynthesis, compared to the original.

Please note that the order of appearance is random, that is, for example, the waveform labeled Resynthesis 1 does not necessarily correspond to a specific model.

The rating is the following:

Perfect: No artifacts and/or identical resynthesis quality.

Very good: Very few artifacts and/or very good resynthesis quality but not identical.

Good: Some artifacts and/or good synthesis quality but not identical.

Bad: Annoying artifacts and/or bad resynthesis quality.

Very bad: Very annoying artifacts and/or very bad resynthesis quality.

Recommendations

• If there is any technical problem with one sound, select Prob.
• Absolutely use headphones. Do not use earphones or speakers!!!
• Verify that the sound is loud enough to hear the details properly.
• Do the test in a quiet place.
• Take the time to listen !
• Please, do not stop the sound before it finishes!
• Please, do not play audio files simultaneously !
• Before answering the test, do not hesitate to ask me any question.

The test

Please, evaluate the quality of the resynthesized waveforms according to the Original.