Research breakthrough · curated entry
WaveNet models raw audio waveforms
WaveNet: A Generative Model for Raw Audio
- Date
- September 2016
- Category
- Research breakthrough
- Organization / Model
- DeepMind
- Evidence
- 1 source · 1 evidence family
DeepMind used dilated causal convolution directly on raw audio and clearly beat the concatenative systems of the day in English and Mandarin speech synthesis evaluations. Speech generation moved from acoustic-parameter pipelines to end-to-end neural synthesis, and demonstrated generative models on time-series signals.
Curator's note: Cross-checked against the arXiv preprint.
Curated sources · 1
Cite this page
This page is a curated static snapshot and can be cited directly. Preferred format GB/T 7714—2025 (electronic resource [EB/OL]); the other 5 styles open below, together with the machine-readable files.
GB/T 7714—2025 (electronic resource [EB/OL])
AI 纪年. WaveNet 直接建模原始波形，神经语音合成进入新阶段[EB/OL]. (2016-09)[2026-11-03]. https://aihist.net/e/wavenet-2016.html.
APA 7 · American Psychological Association
AI 纪年. (2016, September). WaveNet 直接建模原始波形，神经语音合成进入新阶段 [Web page]. https://aihist.net/e/wavenet-2016.html
MLA 9 · Modern Language Association
"WaveNet 直接建模原始波形，神经语音合成进入新阶段." AI 纪年, Sept. 2016, aihist.net/e/wavenet-2016.html. Accessed 3 Nov. 2026.
Chicago 17 · notes-bibliography
AI 纪年. "WaveNet 直接建模原始波形，神经语音合成进入新阶段." Published Sept. 2016. Accessed Nov. 3, 2026. https://aihist.net/e/wavenet-2016.html.
IEEE · citation-reference list
"WaveNet 直接建模原始波形，神经语音合成进入新阶段," AI 纪年, Sept. 2016. [Online]. Available: https://aihist.net/e/wavenet-2016.html. [Accessed: 3 Nov. 2026].
ACM · author-date variant
AI 纪年. September 2016. WaveNet 直接建模原始波形，神经语音合成进入新阶段. Retrieved November 3, 2026 from https://aihist.net/e/wavenet-2016.html.
All six styles come from one template and can be compared word for word. [引用日期], Accessed, urldate and Y2 all carry this page's build date, 2026-11-03 — a static page is an immutable snapshot, so substitute the date you actually consulted it. The address in every style is the archive's canonical page: this English page translates the same snapshot, and the work keeps its name, so one record cites both. For one-click copying use the same entry in the timeline, which takes the day you opened it. Expand the other styles before printing.