Speech information can be roughly decomposed into four components: language content,
timbre, pitch, and rhythm. Obtaining disentangled representations of these components is …

Modern neural text-to-speech (TTS) synthesis can generate speech that is indistinguishable
from natural speech. However, the prosody of generated utterances often represents the …

In this paper we propose a Sequential Representation Quantization AutoEncoder (SeqRQ-
AE) to learn from primarily unpaired audio data and produce sequences of representations …

The rapid spread of media content synthesis technology and the potentially damaging
impact of audio and video deepfakes on people's lives have raised the need to implement …

Spontaneous speech has many affective and pragmatic functions that are interesting and
challenging to model in TTS. However, the presence of reduced articulation, fillers …

This paper describes a method to control prosodic features using phonetic and prosodic
symbols as input of attention-based sequence-to-sequence (seq2seq) acoustic modeling …

The cloning of a speaker's voice using an untranscribed reference sample is one of the great
advances of modern neural text-to-speech (TTS) methods. Approaches for mimicking the …

Neural text-to-speech (TTS) synthesis can generate speech that is indistinguishable from
natural speech. However, the synthetic speech often represents the average prosodic style …

Although Sequence-to-Sequence (S2S) architectures have become state-of-the-art in
speech synthesis, capable of generating outputs that approach the perceptual quality of …

Sequence-to-Sequence Text-to-Speech architectures that directly generate low level
acoustic features from phonetic sequences are known to produce natural and expressive …