ASR datasets · मराठी
Marathi ASR Training Data
Transcribed speech corpora built to train and evaluate automatic speech recognition, with verbatim transcription, timestamps and per-token language tagging where code-mixing occurs. This page covers how that works specifically for Marathi, where retains the retroflex lateral ळ, which has no hindi or english equivalent and is frequently substituted with ल by non-native transcribers.
- Language
- Marathi (mr-IN)
- Dialects covered
- 6
- Typical programme
- 500-2,000 hours
- Cities
- Mumbai, Pune, Nagpur
What changes when the language is Marathi
The service specification stays constant across languages; the linguistics do not. For Marathi, three things drive the design of a asr datasets programme.
- Retains the retroflex lateral ळ, which has no Hindi or English equivalent and is frequently substituted with ल by non-native transcribers
- Dialect spread: Standard (Puneri), Varhadi (Vidarbha), Marathwadi, Konkani-influenced coastal Marathi
- Mumbai and Pune speech mixes Marathi, Hindi, and English in the same sentence. Marathi-only recordings collected in Pune under-represent the Mumbai reality of tri-lingual switching.
Technical specification
| Parameter | Standard |
|---|---|
| Audio | 16 kHz or 48 kHz PCM WAV |
| Transcription | Verbatim, including disfluencies, false starts and fillers |
| Timestamps | Utterance level by default; word level on request |
| Tagging | Noise, overlap, unintelligible, foreign-language and code-switch tags |
| Normalisation | Raw and normalised text columns delivered separately |
| Split | Train/dev/test splits with no speaker leakage across splits |
Marathi cohort design
A representative Marathi cohort should be split roughly 40% western Maharashtra, 25% Vidarbha, 20% Marathwada, 15% Konkan rather than concentrated in Pune.
| Dimension | Typical split | Why it matters for Marathi |
|---|---|---|
| Gender | 50 / 50 | Pitch range differences change acoustic model behaviour; unbalanced cohorts bias recognition |
| Age | 18-25: 30%, 26-40: 40%, 41-60: 30% | Older speakers retain conservative Marathi forms that younger urban speakers have lost |
| Region | Maharashtra / Goa / parts of Karnataka and others | Dialect spread across 6 recognised varieties |
| Education | Mixed, including below-graduate | Prompt-reading fluency correlates with education and skews prosody |
| Condition | Studio / quiet room / field | Match the noise profile of your deployment |
Process
- Style guide authored per language, covering numerals, loanwords, script and disfluency rules
- Transcriber calibration round with inter-annotator agreement measurement
- First-pass transcription
- Second-pass native review
- Automated consistency checks against the style guide
- Split generation with speaker-disjoint partitions
Marathi-specific quality rules
- ळ vs ल substitution by Hindi-trained transcribers
- Anusvara placement varies between conservative and modern orthography
- Varhadi verb endings get 'corrected' to standard forms unless the guide explicitly requires verbatim dialect transcription
Agreement is measured, not assumed. We report word-level agreement on a held-out sample so you can judge label quality before training on it.
Deliverables
- Audio plus aligned transcripts (JSON/TSV, or your schema)
- Style guide as delivered documentation
- Inter-annotator agreement report
- Speaker-disjoint train/dev/test splits
Worked example
A representative Marathi asr datasets engagement: 500 hours from 1,000 speakers, 50/50 gender, ages 18-45, spread across Mumbai, Pune, Nagpur, recorded to the specification above and delivered in WAV with a per-utterance manifest.
Timeline: Transcription adds roughly 1-2 weeks per 100 hours after recording, per language.
Where this data is missing today
Available Marathi speech data is dominated by standard Puneri read speech. Vidarbha, Marathwada, and coastal Konkan varieties are severely under-collected, which is exactly where deployed voice products lose accuracy.
Frequently asked
How much does Marathi asr datasets cost?
Priced per delivered hour or unit against a written spec. The cost drivers for Marathi are dialect spread, demographic narrowness and recording condition, in that order.
Which Marathi dialects are included?
By default Standard (Puneri), Varhadi (Vidarbha), Marathwadi, Konkani-influenced coastal Marathi and others, tagged per speaker. You can also commission a single-dialect corpus if you are targeting one region.
Can you deliver Marathi data in our format?
Yes. Audio plus aligned transcripts (JSON/TSV, or your schema) is the default, but naming, schema and directory structure follow your pipeline.
How long does a Marathi programme take?
Transcription adds roughly 1-2 weeks per 100 hours after recording, per language.
Request a Marathi asr datasets quote
Hours, speakers, dialects, deadline. Send what you have.