1,000 speakers of Odia Telephony Speech

Speech captured over a real telephony path — a genuine call through the network, not studio audio downsampled afterwards to imitate one.

Scenario-driven calls between a caller and an agent or IVR flow, each leg recorded separately, across a deliberate spread of handsets and network conditions.

Yield at this style: 50–60% of recorded time is accepted. Dropped calls, network artefacts beyond tolerance and unusable legs are discarded. Real network conditions are the point of the style and also its main cost.

• Call scenarios drawn from real contact-centre intents: balance enquiry, complaint, booking change, escalation
• IVR flows scripted with deliberate mis-entry and barge-in paths, since those are where deployed systems fail
• Handset spread specified up front — low-end Android, feature phone, landline — because handset variance is a real acoustic axis
• Background conditions varied on purpose: street, vehicle, indoor, since real callers are rarely in quiet rooms
• Built against Odia specifically: Retains a distinct retroflex ଳ and a full retroflex series
• Code-mixing handled explicitly rather than edited out — Urban Odia mixes Hindi and English; western Odisha mixes Chhattisgarhi and Sambalpuri forms.

Nine to twelve weeks, with recruitment running continuously rather than in a front-loaded phase.

Six to eight batches, speaker-disjoint by construction so held-out speakers are genuinely held out.

Western Odisha recruitment requires local field partners; remote-only recruitment yields an all-coastal cohort.

At 1,000 speakers 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.

Recruitment must be stratified by handset and carrier as well as by dialect, which adds a screening axis the other styles do not have.

• Retains a distinct retroflex ଳ and a full retroflex series
• Urban Odia mixes Hindi and English; western Odisha mixes Chhattisgarhi and Sambalpuri forms.
• Odia is one of the least-resourced major Indian languages. Sambalpuri and Ganjami are effectively absent from public data.

• Codec and sample rate verified per file, rejecting any studio audio that has been downsampled to fake a telephony path
• Echo and double-talk checked on both legs
• DTMF events verified against the call log
• Level normalisation applied per leg, since handset output levels vary far more than studio microphones do
• Sambalpuri normalised into coastal Odia
• Unicode confusables between Odia and Bengali characters when transcribers reuse tooling

100% technical QA, 10% content QA stratified by city and coordinator, with voice-based duplicate detection across the whole cohort.

• Studio audio downsampled to 8 kHz and passed off as telephony, which has none of the codec or packet-loss characteristics that matter
• Echo and double-talk that make the agent leg unusable
• Carrier and handset monoculture, producing a corpus that only represents one acoustic path
• Over-clean recordings from participants who move somewhere quiet to take the call, defeating the purpose

Cost at this band is driven by: Number of distinct demographic cells rather than total speakers.

• Speaker-disjoint splits are only trustworthy if duplicate detection is voice-based rather than ID-based, since one person can present two identities
• Coordinator-level demographic bias becomes measurable at this size and should be reported, not smoothed over
• Odia carries 5 recognised varieties across Odisha, parts of Jharkhand, West Bengal, Chhattisgarh and Andhra Pradesh, so the quota matrix is wider than the headline volume suggests
• Odia is one of the least-resourced major Indian languages. Sambalpuri and Ganjami are effectively absent from public data.

• WAV audio to your naming convention, with the per-file manifest
• Verbatim Odia transcripts with utterance-level timestamps
• Call metadata: duration, codec, handset class, carrier, packet-loss events
• Intent label per call and per turn
• DTMF and hold, transfer and barge-in events
• Agent and caller leg identifiers
• 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

• Contact-centre and IVR ASR
• Voice bots operating over the phone network
• Intent classification on narrowband audio
• Robustness to codec and packet loss

Narrowband telephony audio is the wrong input for TTS or any wideband model — the frequency content simply is not there. Use it for models that will be deployed on a phone line and nothing else.