For most languages, automatic speech recognition (ASR) is limited less by model architecture than by a lack of accurately transcribed spontaneous speech. We present Monsoon, a framework for collecting such data at scale. It elicits unscripted conversations between paired contributors, screens recordings with automated classifiers, corrects transcripts over five levels of human review, and selects segments for lexical and acoustic coverage. It has collected speech in 50 languages; we analyse seven 3,500-hour corpora. At equal volume, its quality stages reduce semantic word error rate (WER) by 10-39% relative to a conventional protocol. Compared with an architecture-matched baseline trained on public corpora, Whisper fine-tuned on Monsoon performs comparably on the baseline's own test sets and reduces semantic WER by 20-44 points on conversational benchmarks. The gains transfer to a second architecture and grow with data volume. We release the framework, its quality-control models, and the fine-tuned models.
Figures & tables
Lang.
Hours
Utt.
Dur.
Spk.
F
Dist.
Dev.
MOS
Bengali
3,500
1.04M
13.4
4,863
62%
127
787
3.66
Gujarati
3,544
1.22M
11.9
5,568
40%
106
806
3.53
Hindi
3,611
1.16M
12.9
13,731
37%
781
1,123
3.61
Kashmiri
3,330
1.62M
6.4
2,936
19%
31
447
3.61
Marathi
3,500
1.08M
13.2
5,329
57%
111
812
3.63
Tamil
3,488
1.14M
12.5
3,452
58%
95
777
3.56
Table 1: Corpus statistics. Dur.: median segment length (s); F: female share of audio; Dist.: districts; Dev.: device models; MOS: median DNSMOS.
Figure 1: Semantic WER (%) of Monsoon-trained Whisper and IndicWhisper. (a) Vistaar mean (sets per language in brackets); (b) IndicVoices (solid) and Voice of India (dotted, OIWER). Radial scales differ.
Figure 2: Semantic WER (%) at 1,000 hours: conventional protocol (grey) vs. Monsoon (blue), same recipe. Labels: relative reduction.
System
hi
bn
gu
mr
ta
te
Avg.
MAI-Transcribe-2
9.5
8.3
11.6
10.5
13.2
17.7
11.8
Gemini 3.1 Pro
13.1
10.3
14.0
14.7
18.3
18.8
14.9
ElevenLabs Scribe v2
13.2
12.0
16.6
16.4
22.7
23.2
17.4
Muse Voice Transcribe
16.2
14.5
31.9
24.8
35.2
33.0
25.9
GPT Realtime Whisper
33.4
45.0
47.9
48.1
64.5
55.0
49.0
Whisper medium
57.3
91.5
91.4
83.5
74.9
92.7
81.9
Table 2: Semantic WER (%) on IndicVoices. Best in bold. Commercial systems evaluated in September 2026.
Figure 3: Semantic WER relative to the 1,000-hour model, on IndicVoices (solid) and the Vistaar mean (dashed). Labels: total relative reduction.
Automatic speech recognition (ASR) performs well for high-resource languages with abundant paired audio-transcript data, but its accuracy degrades sharply for most languages due to limited publicly available aligned data. To this end, we introduce WorldSpeech, a 24 kHz multilingual speech corpus comprising 65k hours of aligned audio-transcript data across 76 languages, collected from diverse public sources including parliamentary proceedings, international broadcasts, and public-domain audiobooks. For 37 languages, WorldSpeech provides more than 200 hours of aligned speech, with 28 exceeding 500 hours and 24 surpassing 1k hours. Fine-tuning existing ASR models on WorldSpeech results in an average relative Word-Error-Rate reduction of 63.5% across 11 typologically diverse languages.
Antonis Asonitis, Luca A. Lanzendörfer, Frédéric Berdoz +1
We introduce BuzzASR, a collection of language-specialized fine-tuned Whisper models adapted for automatic speech recognition (ASR) in 102 languages. Large end-to-end Transformer-based ASR models such as Whisper have revolutionized ASR, but most prominent models are highly multilingual. As a result, these models often perform poorly on languages less well-represented in their training set. While it has long been known that effective language adaptation can be achieved through simple fine-tuning on monolingual data, this strategy has only been applied to a small number of languages. We massively scale up this simple approach to 102 languages covered in the FLEURS dataset, while also implementing a more complex language adaptation strategy that integrates monolingual tokenizer replacement and data augmentation using text-only fine-tuning. BuzzASR models outperform Whisper-large-v3 on 77 out of 102 languages, reducing character error rates (CER) by a factor of over 2.8 on average. Our models achieve state-of-the-art CER among open-source systems on 27 of 102 languages on the combined FLEURS and Common Voice test set. Our tokenizer replacement strategy yields an average 3.3x improvement in compression rate (characters per token) over Whisper's multilingual BPE, with gains of up to 21.7x. We release all models, code, and detailed results: https://lemn-lab.github.io/buzz-asr
This study reports the development of an Automatic Speech Recognition (ASR) system in Mizo, a low-resource language. The development included collecting 17.62 hours of speech data, curating it, and fine-tuning the Mizo ASR system with three Whisper multilingual models and with the SraVaani 1.0 Indic multilingual model. Whisper-large-v3 achieved the lowest conventional WER (18.08%), while morphology-aware evaluation yielded a WER of 7.22%. Zero-shot evaluation of the SraVaani 1.0 Indic multilingual model yielded a WER of 58.27%, while Mizo-specific fine-tuning reduced the conventional WER to 29.45% and the morphology-aware WER to 17.93%. The results demonstrate that the Whisper model can achieve a substantially low WER, even when adapted to an unseen language. In contrast, SraVaani 1.0 supports the Mizo language in its multilingual model; however, fine-tuning with carefully curated Mizo speech data substantially improves its performance.