Hasil (
Bahasa Indonesia) 1:
[Salinan]Disalin!
Vocoders are a class of speech coding systems that analyze the voice signalat the transmitter, transmit parameters derived from the analysis, and then syn-thesize the voice at the receiver using those parameters. All vocoder systemsattempt to model the speech generation process as a dynamic system and try toquantify certain physical constraints of the system. These physical constraintsare used to provide a parsimonious description of the speech signal. Vocoders arein general much more complex than the waveform coders and achieve very higheconomy in transmission bit rate. However, they are less robust, and their per-formance tends to be talker dependent. The most popular among the vocodingsystems is the linear predictive coder (LPG). The other vocoding schemes includethe channel vocoder, formant vocoder, cepstrum vocoder and voice excitedvocoder.Figure 7.5 shows the traditional speech generation model that is the basisof all vocoding systems [Fla79]. The sound generating mechanism forms thesource and is linearly separated from the intelligence modulating vocal tract fil-ter which forms the system. The speech signal is assumed to be of two types:voiced and unvoiced. Voiced sound ("m”, “n”, “v” pronunciations) are a result ofquasiperiodic vibrations of the vocal chord and unvoiced sounds (“/", “s”, “sh” pro-nunciations) are fricatives produced by turbulent air flow through a constriction.The parameters associated with this model are the voice pitch, the pole frequen-cies of the modulating filter, and the corresponding amplitude parameters. Thepitch frequency for most speakers is below 300 Hz, and extracting this informa-tion from the signal is very difficult. The pole frequencies correspond to the reso-nant frequencies of the vocal tract and are often called the formants of thespeech signal. For adult speakers, the formants are centered around 500 Hz,1500 Hz, 2500 Hz, and 3500 Hz. By meticulously adjusting the parameters of thespeech generation model, good quality speech can be synthesized.
Sedang diterjemahkan, harap tunggu..
