English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Non-autoregressive text to speech (TTS) models such as FastSpeech cansynthesize speech significantly faster than previous autoregressive models withcomparable quality. The training of FastSpeech model relies on anautoregressive teacher model for duration prediction (to provide moreinformation as input) and knowledge distillation (to simplify the datadistribution in output), which can ease the one-to-many mapping problem (i.e.,multiple speech variations correspond to the same text) in TTS. However,FastSpeech has several disadvantages: 1) the teacher-student distillationpipeline is complicated and time-consuming, 2) the duration extracted from theteacher model is not accurate enough, and the target mel-spectrograms distilledfrom teacher model suffer from information loss due to data simplification,both of which limit the voice quality. In this paper, we propose FastSpeech 2,which addresses the issues in FastSpeech and better solves the one-to-manymapping problem in TTS by 1) directly training the model with ground-truthtarget instead of the simplified output from teacher, and 2) introducing morevariation information of speech (e.g., pitch, energy and more accurateduration) as conditional inputs. Specifically, we extract duration, pitch andenergy from speech waveform and directly take them as conditional inputs intraining and use predicted values in inference. We further design FastSpeech2s, which is the first attempt to directly generate speech waveform from textin parallel, enjoying the benefit of fully end-to-end inference. Experimentalresults show that 1) FastSpeech 2 achieves a 3x training speed-up overFastSpeech, and FastSpeech 2s enjoys even faster inference speed; 2) FastSpeech2 and 2s outperform FastSpeech in voice quality, and FastSpeech 2 can evensurpass autoregressive models. Audio samples are available athttps://speechresearch.github.io/fastspeech2/.

FastSpeech 2: Fast and High-Quality End-to-End Text to Speech