This paper presents methods for speech spectrum prediction based on Gaussian mixture models. Spectrum prediction may be useful in a packet transmission system where the sensitivity to packet losses is a major problem. Models of speech are trained by the Expectation Maximization algorithm using pairs, triples etc. of consecutive cepstral vectors. The models are used to design first, second etc. order predictors. The prediction schemes are evaluated using the spectral distortion criterion and compared to a simple reference method. The best prediction scheme obtains an average spectral distortion that is 0.46 dB less than for the reference method.

YY model based on spectrum prediction.pdf

• We are an exchange download platform that only provides communication channels. The downloaded content comes from the internet. Except for download issues, please Google on your own.
• The downloaded content is provided for members to upload. If it unintentionally infringes on your copyright, please contact us.
• Please use Winrar for decompression tools
• If download fail, Try it againg or Feedback to us.
• If downloaded content did not match the introduction, Feedback to us，Confirm and will be refund.
• Before downloading, you can inquire through the uploaded person information

Nothing．