Loading datasets...
Release Date: 7/16/2026
Format: TSV, MP3
Size: 1.94 GB
Share
This dataset is a derived, accent-filtered subset of the Mozilla Common Voice Scripted Speech 26.0 English release (`cv-corpus-26.0-2026-06-12`, locale `en`). It keeps only validated rows (present in the source `validated.tsv`) whose self-declared `accents` field is *exactly* `"Australian English"`. No gender or age filter is applied. The result is split speaker-disjoint into train/dev/test (no `client_id` appears in more than one split).
Restrictions/Special Constraints
N/A
Forbidden Usage
- You agree not to attempt to determine the identity of speakers. - You agree not to rehost this dataset.
Ethical Review
Source recordings and self-declared demographics (age, gender, accent) were volunteered by Common Voice contributors under Common Voice's own consent and contribution terms. This package only filters and repackages a subset; it does not collect new data or alter consent terms.
Intended Use
Research and development of English ASR systems, with emphasis on a clean, single-tag Australian English accent slice.
Splits and speakers: 55,922 validated clips total — train 52,984 clips / 754 speakers, dev 1,540 clips / 16 speakers, test 1,398 clips / 34 speakers. Split is speaker-disjoint (no client_id appears in more than one split), proportioned ~94.75/2.75/2.50 by clip count via a greedy per-speaker balancing algorithm targeting 95/2.5/2.5 (speakers have uneven clip counts, so exact ratios aren't achievable while keeping speakers disjoint).
Clip duration: 78.01 hours total (280,847 seconds) across 55,922 clips. Per-split: train 74.34h, dev 1.86h, test 1.81h. Per-clip: mean 5.02s, median 4.92s, min 0.86s, max 22.49s, std 1.67s.
Clip duration histogram:
0-1s | 1 ( 0.0%)
1-2s | ### 860 ( 1.5%)
2-3s | ##################### 5,356 ( 9.6%)
3-4s | ########################################## 10,507 (18.8%)
4-5s | ################################################## 12,260 (21.9%)
5-6s | ############################################### 11,536 (20.6%)
6-7s | ################################## 8,374 (15.0%)
7-8s | ################# 4,369 ( 7.8%)
8-9s | ####### 1,843 ( 3.3%)
9-10s | ## 682 ( 1.2%)
10-12s | 119 ( 0.2%)
12-15s | 10 ( 0.0%)
15-20s | 3 ( 0.0%)
20-30s | 2 ( 0.0%)
30s+ | 0 ( 0.0%)
Gender distribution: male_masculine 31,526; blank/unspecified 14,682; female_feminine 9,714.
Age distribution (shipped subset, self-declared): twenties 14,069; teens 13,403; fourties 9,015; thirties 8,794; fifties 4,260; blank/unspecified 2,566; sixties 2,465; seventies 1,095; eighties 255.
Closed-form accent matching (QA): the accents field is pipe-delimited multi-select in the source corpus, but this subset was deliberately filtered to the closed-form (exact-match) rule — 100% of shipped clips have accents == exactly "Australian English", no co-occurring tags.
Open-form vs. closed-form comparison: open-form (accents contains "Australian English") = 56,321 clips / 841 speakers / 78.57h; closed-form (accents == exactly "Australian English") = 55,922 clips / 804 speakers / 78.01h. Closed-form thus sacrifices the 399 clips (0.71%).
Full-corpus accent x gender and age x gender cross-tabs are in stats.md.
Derivation methodology: derived from the Mozilla Common Voice Scripted Speech 26.0 English release (cv-corpus-26.0-2026-06-12, locale en). Filter: validated rows only (present in the source validated.tsv), whose self-declared accents field is exactly "Australian English" (closed-form/exact match). No gender or age filter applied. Speaker-disjoint train/dev/test split via a greedy per-speaker assignment balancing toward 95/2.5/2.5 targets. Only the referenced MP3 clips were extracted from the ~2.58M-clip, ~89GB source archive (not the full archive) via a single sequential streaming pass over the source tar.gz.