I am trying to assess the degree of degradation that "musical noise" causes in the low frequency bands of the spectrum of speech signals. Perceptually (playing back the treated signal) this artifact is stronger in mid and high frequencies (over 700 Hz), however I need an objective way to confirm or disprove this.
Does anyone have information on this subject or knows a way to evaluate the amount of musical noise present in a signal?
Thank you very much.
It is exactly in the frequencies which are weakened by the algorithm or in the places that noise has strong peaks and more power.
If the noise is white then the musical noise is distributed in all frequencies.
The main reason of the musical noise is over reduction(or cutoff) of noise bands. Most speech enhancemnet algorithms don't touch peaks of the noisy (speech + noise) spectrum, but they often too much notch valleys of the noisy spectrum. If you reduce the amount of enhancement gain, then the musical noise disappears.
Evidently not, because this noise is generated thanks to speech short-spectrum non-uniformity.
"speech short-spectrum non-uniformity" is general in any noisy speech signal. Most speech enhancement algorithms calculate each optimal gain for the consecutive short-spectrum. This gain is usually 0.0~1.0 1.0 gain at any spectrum band is no problem, but 0.0 gain at the valley (not peak) of a specific spectrum causes the musical noise.
Thank you very much for your answers and sorry for the tardiness of my response.
This question originated from some experiments I'm performing on contaminated voice signals, in which I'm using subspace Speech Enhancement techniques in the Wavelet domain, more precisely on the Wavelet Packet decomposition coefficients. After reconstructing the signal from the processed coefficients it had some musical noise.
From your answers and further reading on several sources I can conclude that "musical noise" is related to frequency spectrum characteristics.
I would like to know the relation between the Wavelet domain processing I applied and the occurrence of "musical noise".
I can improvement the musical noise with perceptual frequency masking filter.
- Badges
- Science topic
- Similar topics
- Correction