Aggregate correlations
Click on the metrics for the detailed view. Legends are clickable, too!
Mean subjective scores by subset
Stimulus-level scores
Correlation by processing method
About the subsets
| Subset | Summary | Reference paper |
|---|---|---|
| 2024-icassp | Mono focus: Movie-like centered dialogue with music and effects + music-only items (including solo instruments). Monaural coding artifacts, namely low-pass (LP), pre-echoes (PE), spectral holes (SH), tonality mismatch (TM), and unmasked noise (UN), plus dialogue-enhancement (DE) systems. MUSHRA with 26 expert listeners from two labs (USA and Germany). | M. Torcoli; C.-W. Wu; et al.; "ODAQ: Open Dataset of Audio Quality;" ICASSP 2024. (arXiv:2401.00197) |
| 2024-aes | Mono focus: The same 2024-icassp stimuli rated again in a MUSHRA in a third lab with 16 trained university students (two cohorts with different training), consistent with the expert ratings. | S. Dick; C. Thompson; et al.; "Expanding and analyzing ODAQ - the open dataset of audio quality;" AES Convention; 2024. (arXiv:2504.00742) |
| 2025-aes | Stereo focus: Solo instruments, pop and R&B mixes with a wide stereo image, and hard-panned speech with music. Stereo coding artifacts, namely quantization noise (QN) and spectral holes (SH), applied in left/right (LR), mid/side (MS), or both within one trial (mix), giving SHLR, SHMS, SHmix, QNLR, QNMS, and QNmix. MUSHRA with 16 trained music-production students, plus 8 experts for extra hard-panned items (not included in the benchmark). | S. Dick; C. Thompson; et al.; "Investigating the impact of stereo processing - a study for extending the Open Dataset of Audio Quality (ODAQ);" AES Convention; 2025. (arXiv:2512.14259) |
| 2025-lanzendorfer | Mono focus, neural speech codecs: Speech (panned to the center) from 2024-icassp, clean and with background, reconstructed by six neural codecs at 2.6-7.7 kbps (each channel separately). MUSHRA-like crowdsourced online test. | L. A. Lanzendoerfer; F. Groetschla; "Evaluating objective speech quality metrics for neural audio codecs;" arXiv preprint; 2025. (arXiv:2511.19734) |
| proposed-addition | Stereo focus: Altered inter-channel coherence (ICC) and level differences (ILD) on speech, music and solo instruments, plus dialogue enhancement of spatial speech with music and effects. MUSHRA in a lab with trained students (23 and 12 listeners). | C.-W. Wu; M. Torcoli; et al.; "ODAQ-Collection: Expanding the Open Dataset of Audio Quality with a Focus on Spatial Distortions;" arXiv preprint; 2026. (arXiv:2610.37959) |