Described herein are analog speech encoder and decoder systems using a plurality of narrow band pass filters with associated rectifiers and ripple filters for spectrum analyzing the speech or other suitable signals, and a corresponding plurality of narrow band pass filters with associated voltage controlled amplifiers for remaking the speech with either an injected carrier or a noise signal being applied to the voltage controlled amplifier inputs, said carrier or noise source signals being selected by a voice controlled circuit, activating the carrier in the presence of voiced sounds or vowels, and the noise source in the presence of unvoiced sounds or consonants; said innovations or the combination thereof consisting of (1) the use of a direct bypass for the high speech frequencies from the voice input to the decoder output, using a high pass filter essentially passing the unvoiced speech sounds, (2) the limitation of the bandpass filter range of the encoder and decoder sections to the cut-off frequency of the high pass filter, (3) switching means to include voice controlled circuitry to activate the high frequency bypass only in the presence of unvoiced speech sounds, (4) switching means to include a noise source for the simulation of unvoiced sounds in the range of the band pass filters below the cut-off frequency of the aforesaid high pass filter, when activated by the voice controlled circuitry, and (5) the inclusion of an amplitude compressing circuit at the voice input and an amplitude expanding circuit at the decoder output for the purpose of increasing the signal-to-noise ratio of the speech encoder and decoder.

In an overlap addition unit, speech waveform data is subjected to overlap addition every period read out from a period storage unit, and in a simple addition unit, the waveform data obtained by the overlap addition and the aperiodic waveform data read out from an aperiodic waveform storage unit are added to each other. Thus, the aperiodic waveform is given to the speech waveform to improve the quality of synthesized speech.

This improved voice-excited speed synthesizer expands the voice-band spectrum to generate high-frequency band excitation signal by using two separate distortion networks: a full-wave rectifier for lower frequencies, and a limiter for higher frequencies. Channel and formant type voice excited embodiments are shown.

A processing device transmits speech signals of constant amplitude during periods of voice activity and interrupts transmission outside these periods, so that noise is not transmitted. To this end the processing device comprises an amplitude-compressing device preceded by a voice activity detector comprising a delay line in series with an electronic switch and a speed - noise discriminator having a common input with the delay line for receiving the voice signals. The switch is controlled by the output signal of the discriminator in such a way that it only permits the passage of the voice signals corresponding to the periods of voice activity.

Improved unvoiced speech performance in low-rate multi-pulse coders is achieved by employing a multi-pulse architecture that is simple in implementation but with an output quality comparable to code excited linear predictive (CELP) coding. A hybrid architecture is provided in which a stochastic excitation model that is used during unvoiced speech is also capable of modeling voiced speech by use of random codebook excitation. A modified method for calculating the gain during stochastic excitation is also provided.