Speech codec compresses the input signal into compact bit stream, which is then decoded
at the receiver to generate the best possible perceptual quality. This compression makes …

GAN vocoders are currently one of the state-of-the-art methods for building high-quality
neural waveform generative models. However, most of their architectures require dozens of …

In order to efficiently transmit and store speech signals, speech codecs create a minimally
redundant representation of the input signal which is then decoded at the receiver with the …

Neural vocoders have recently become popular in text-tospeech synthesis and voice
conversion, increasing the demand for efficient neural vocoders. One successful approach is …

In this paper, we propose a method of perceptually optimizing the deep neural network
(DNN)-based speech coder using multi-time-scale perceptual loss functions. We utilize a …

We present Espresso, an open-source, modular, extensible end-to-end neural automatic
speech recognition (ASR) toolkit based on the deep learning library PyTorch and the …

Although a HiFi-GAN vocoder can synthesize high-fidelity speech waveforms in real time on
CPUs, there is a tradeoff between synthesis quality and inference speed. To increase …

Recent studies have shown that neural vocoders based on generative adversarial network
(GAN) can generate audios with high quality. While GAN based neural vocoders have …

Good speech quality has been achieved using waveform matching and parametric
reconstruction coders. Recently developed very low bit rate generative codecs can …

Recently, GAN based speech synthesis methods, such as MelGAN, have become very
popular. Compared to conventional autoregressive based methods, parallel structures …