Non-Autoregressive Neural Text-to-Speech

By Kainan Peng, Wei Ping, Zhao-long Song, Kexin Zhao. 2019

119 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Natural Language Processing

Citations in the corpus

DiffWave: A Versatile Diffusion Model for Audio Synthesis (2020, methodology): “In contrast to GAN or VAE-based models (Donahue et al., 2019; Peng et al., 2020), it is much easier to train without mode collapse, posterior collapse, or training instability stemmed from the joint training of two networks.”
WaveGrad: Estimating Gradients for Waveform Generation (2020, methodology): “, 2020), variational auto-encoder models (Peng et al., 2020), as well as models inspired by digital signal processing (Ai & Ling, 2020; Engel et al.”
A Survey on Neural Speech Synthesis (2021, methodology): “Some works such as ParaNet [262] and Flow-TTS [228] uses positional attention [264] for the text and speech alignment in parallel generation.”
Conditional Variational Autoencoder with Adversarial Learning for End-to-End Text-to-Speech (2021, methodology): “Parallel TTS models (Ren et al., 2019; Peng et al., 2020; Kim et al., 2020; Ren et al., 2021; Lee et al., 2021), on the other hand, have been relied on deterministic duration prediction.”
AdaSpeech: Adaptive Text to Speech for Custom Voice (2021, methodology): “We build AdaSpeech based on the popular non-autoregressive TTS models (Ren et al., 2019; Peng et al., 2020; Kim et al., 2020; Ren et al., 2021) and further design several techniques to address the challenges in custom voice: • Acoustic condition modeling.”
It's Raw! Audio Generation with State-Space Models (2022, neutral): “(iii) Autoencoders: WaveNet Autoencoder [11] and WaveVAE [31], which use WaveNets in their encoders.”
Glow-TTS: A Generative Flow for Text-to-Speech via Monotonic Alignment Search (2020, background): “In addition, most of the autoregressive models show a lack of robustness in some cases [20, 16].”
WaveFlow: A Compact Flow-based Model for Raw Audio (2019, neutral): “Several neural textto-speech (TTS) systems have been introduced, including WaveNet (van den Oord et al., 2016), Deep Voice 1 & 2 & 3 (Arık et al., 2017a;b; Ping et al., 2018), Tacotron 1 & 2 (Wang et al., 2017; Shen et al., 2018), Char2Wav (Sotelo et al., 2017), VoiceLoop (Taigman et al., 2018), Wav”
Semi-Supervised Speech Recognition via Local Prior Matching (2020, methodology): “Several methods can be used to estimate the text length on an unlabeled utterance, including the average speaking rate (Peng et al., 2019) or a phoneme/syllable segmentation (Adell & Bonafonte, 2004; Scharenborg et al., 2010; Wang et al., 2017).”
SC-GlowTTS: an Efficient Zero-Shot Multi-Speaker Text-To-Speech Model (2021, methodology): “We synthesize 15 different sentences as in [36] 10 times for each of the 11 speakers of the VCTK test set and calculated the average.”
One TTS Alignment to Rule Them All (2021, methodology): “These models generally rely on ex-ternal aligners [4] like the Montreal Forced Aligner (MFA) [10], or on durations extracted from a pre-trained autoregressive model (or forced aligner) [5,7,11] like Tacotron 2 [2].”
RAD-TTS: Parallel Flow-Based TTS with Robust Alignment Learning and Diverse Synthesis (2021, background): “Recent works (Kim et al., 2020; Miao et al., 2020; Peng et al., 2020) have presented increasingly parallel solutions to address these concern.”
DiffGAN-TTS: High-Fidelity and Efficient Text-to-Speech with Denoising Diffusion GANs (2022, background): “They either leverage an external text-to-acoustic alignment module (Ren et al., 2019; 2021a; Peng et al., 2020; Elias et al., 2021) or jointly train one within the TTS model (Zeng et al., 2020; Miao et al., 2021; Badlani et al., 2021).”
Translatotron 2: High-quality direct speech-to-speech translation with voice preservation (2021, result): “However, it caused a signiﬁcant regression on the naturalness of the predicted translation speech, which is consistent with the observation in TTS in Shen et al. (2020); Peng et al. (2020); Hwang et al. (2021), etc., suggesting more exploration is needed on this direction.”
An Overview of Affective Speech Synthesis and Conversion in the Deep Learning Era (2022, background): “DL-based methods can be taxonomized along several categories : 1) autoregressive (AR) [14] versus non-AR (NAR) structures [70]; 2) type of network structure , where we primarily differentiated between convolutional neural networks (CNNs) [66], [67], sequential models (RNNs, gated recurrent unit…”