Rate the audibility of musical noise (isolated tonal artifacts) in 26 short speech clips on a 5-point ACR scale. Approximately 15 minutes with headphones. Supports the MN_rate validation paper.
Download the full anonymized dataset for your research.
No personal data included. User IDs are hashed. Export is rate-limited.
This download provides raw, unfiltered data. By downloading you accept full responsibility for its use. The dataset relies on participant good faith. No responses are pre-screened or removed. As the number of participants grows, outliers from misunderstandings or deliberate noise naturally fade in statistical significance.
flagged_anchor: anchor stimulus S25 rated ≥ 3 (should be inaudible; participant likely missed the reference artifact).flagged_retest: same clip presented twice (S10 vs S26) rated ≥ 2 points apart (low intra-rater consistency).flagged_variance: ≥ 24 of 26 trials given the same rating (response stuck on one button).flagged_fast: median reaction time < 1 s (likely clicking through without listening).Flags are informational. Rows are not removed; decide your own QC threshold.
Loading your progress…
Download all personal data linked to your account (LGPD Art. 18 / GDPR Art. 15) or delete your account. Deletion erases your identifiers; your anonymized experiment responses remain in the open-science dataset and cannot be linked back to you. Cannot be undone.