Categories Companies Research

Loading the SOTA2 catalog…

Explore

Categories Companies Research

Contribute

Join SOTA2 Contact

SOTA2

© 2026 SOTA2

Build in Bellevue, USA

WizWand is now SOTA2 Research

Papers Benchmarks Datasets Tasks

Audio Synthesis research benchmarks · SOTA2 Research

Audio Synthesis

Benchmarks

Dataset NameSOTA MethodSort

Dataset NameSOTA methodMetricTrendResultsLast Updated

Singing Voice MUSHRA (evaluation)

86.78MUSHRA Score

21

Feb 26, 2026

Singing Voice Industrial setting

4.33MOS Prediction

21

Feb 26, 2026

Singing Voice Academic setting

4.18MOS Prediction Score

21

Feb 26, 2026

Groove MIDI Dataset (test)

Target DAC reconstruction

0.016FAD∞

13

May 14, 2026

4.23NMOS

12

Mar 26, 2026

LJSpeech (unseen)

0.11MAE

10

Feb 26, 2026

LJSpeech (test)

4.82GPU Execution Time

6

Feb 26, 2026

MovieGen-Audio-Bench

8.95IS

5

Jul 1, 2026

0.878MR-STFT

5

Mar 13, 2026

20 generated audio samples (test)

1.525RTF

5

Feb 26, 2026

LJSpeech Long Audio

3.53SMOS Score

3

Mar 10, 2026

LJSpeech Short Audio

3.07SMOS

3

Mar 10, 2026

Tri-modality (test)

81.07OVL Score

3

Feb 26, 2026

LJSpeech 44.1kHz (test)

152.58GPU xRT

2

Feb 26, 2026

Birdsong (test)

-0.81SNR (dB)

1

Feb 26, 2026

MUSDB 18 (test)

—

—Primary metric

0

Feb 26, 2026

Popular datasets

LJSpeech Singing Voice MUSDB BirdSong Tri-modality

Follow for updates

Join SOTA2 Contact