2,000 hours of Urdu Scripted Speech

Speakers read prompts from a phonetically balanced script shown one line at a time, with retakes on any flubbed or misread line.

A 45-minute booth session yields roughly 300–400 prompts. One file per utterance, cut at the prompt boundary, so alignment is exact before any forced-alignment pass runs.

Yield at this style: 85–90% of recorded time is accepted. Prompt-level retakes catch problems inside the session, so very little is discarded afterwards. This is the highest-yield style we run.

• Script built for triphone coverage rather than word coverage, so rare phone contexts appear often enough to train on
• Digit strings, dates, currency and person names blocked separately, because these are where deployed ASR actually fails
• Domain lexicon injected from your product vocabulary when you supply one
• Sentence-length distribution spread deliberately, since all-short prompts produce a model that cannot handle long utterances
• Built against Urdu specifically: Shares most phonology with Hindi but adds Perso-Arabic phonemes (/q/, /x/, /ɣ/, /z/, /f/) that many speakers merge
• Code-mixing handled explicitly rather than edited out — Spoken Urdu and spoken Hindi are largely mutually intelligible; the distinction is mainly lexical and orthographic. Decide up front whether transcription is in Nastaliq, Devanagari, or both.

Twenty to twenty-eight weeks, run in two or three waves. Wave one establishes the protocol and the transcription convention; later waves scale against a proven baseline rather than an assumed one.

Continuous fortnightly delivery from week four, roughly fourteen batches, each with acceptance testing and a rolling demographic reconciliation against the target quota matrix.

Fix the script decision before fielding; retro-transcribing a Nastaliq dataset into Devanagari after delivery costs as much as the original transcription pass.

At 2,000 hours the quota matrix is enforced per cell, not in aggregate. Every dialect, age and gender combination carries its own target and is signed off individually before final acceptance, because an aggregate 50/50 split can hide a cell that was never filled at all.

Requires literate speakers comfortable reading aloud in the target script, which is the main constraint on cohort breadth and has to be actively counterweighted.

• Shares most phonology with Hindi but adds Perso-Arabic phonemes (/q/, /x/, /ɣ/, /z/, /f/) that many speakers merge
• Spoken Urdu and spoken Hindi are largely mutually intelligible; the distinction is mainly lexical and orthographic. Decide up front whether transcription is in Nastaliq, Devanagari, or both.
• Indian Urdu specifically, and Dakhini in particular, are absent from public data dominated by Pakistani Urdu broadcast speech.

• Prompt-to-audio match verified by a native reviewer; a misread line is a rejected file, not an edited transcript
• Hyperarticulation flagged — a speaker over-enunciating because they are reading produces audio that does not match deployment
• Clipping and truncation checked at both utterance boundaries
• Per-speaker prompt coverage confirmed, so no speaker silently skips a block
• Right-to-left Nastaliq tooling errors and diacritic loss
• Merged phonemes transcribed by sound rather than by etymology, or vice versa, inconsistently

100% technical QA, 5% content QA stratified across every axis, a blind 2% re-transcription audit, and a monthly cross-wave consistency review comparing early and late batches for drift.

• Reading voice: flat prosody and unnatural stress that trains a model on speech nobody actually produces
• Prompt fatigue in the back half of long sessions, where accuracy drops and pace flattens
• Speakers who are not fluent readers, which quietly biases the cohort toward higher education bands
• Script leakage across speakers, producing a corpus that memorises sentences instead of covering sounds

Cost at this band is driven by: Wave structure and the programme governance it requires; Long-tail demographic and dialect quotas, which dominate the final third of the build.

• Drift between waves is the defining risk: audio recorded in month one and month six must be indistinguishable in convention, or the corpus splits into two datasets
• Speaker pool exhaustion is a live constraint in all but the largest languages and shapes which cities are used
• Staff turnover across twenty-eight weeks is a certainty, so handover documentation is part of the deliverable rather than an afterthought
• Storage, transfer and manifest integrity become engineering problems in their own right at this size
• Urdu carries 5 recognised varieties across Uttar Pradesh, Telangana, Bihar, so the quota matrix is wider than the headline volume suggests
• Indian Urdu specifically, and Dakhini in particular, are absent from public data dominated by Pakistani Urdu broadcast speech.

• WAV audio to your naming convention, with the per-file manifest
• Verbatim Perso-Arabic (Nastaliq) transcripts with utterance-level timestamps
• Prompt ID mapped to every utterance
• Verbatim deviation flags where the speaker departed from the script
• Per-utterance SNR and duration in the manifest
• Per-speaker metadata: age band, gender, region, dialect, education band
• Consent records mapped to speaker IDs
• QA report with pass rates, rejection reasons and agreement statistics
• Speaker-disjoint train / dev / test splits on request

• ASR acoustic model baselines
• TTS voice building where a single speaker is recorded at depth
• Pronunciation lexicon and G2P validation
• Forced-alignment and phone-boundary work

Scripted audio contains no disfluencies, no false starts and no natural turn-taking. A model trained on it alone degrades sharply on real spontaneous input, so it is a baseline layer rather than a complete training set.