Neural Vocoder

A deep-learning model that synthesises audio waveforms from intermediate acoustic representations such as mel-spectrograms or discrete speech units. Examples include HiFi-GAN, WaveNet, WaveGlow, and SoundStream. Neural vocoders have largely replaced classical signal-processing vocoders in text-to-speech, voice conversion, and speech-restoration pipelines because they produce higher-quality, more natural output. For accessibility, neural vocoders underpin modern assistive speech technologies — screen-reader voices, AAC output, voice-cloning for voice banking, and dysarthric-to-normal speech restoration.