Texadata — Catalogue
Sample catalogue

Verified, in hand, and playable.

12 motion assets and 8 speech assets. This is not the inventory — it is the standard every asset we license is held to. Every asset plays here in full; masters are delivered in their original containers.

12
video assets
10:45
motion runtime
29:04
speech runtime
2.89GB
source volume
106.3Mbps
peak master rate
2
with PCM sync audio
Motion

Aerial and ground capture, Nigeria.

Hover any card for a moving preview; select one to play the complete asset with its full technical and rights record. Filter by scene taxonomy.

Speech

Unscripted, in real acoustic conditions.

Complete recordings at 48 kHz — 29:04 across 8 assets, playing in full. Nothing here is trimmed. Each carries the result of running open-weight ASR against it.

tx-s001 · en-NG
Radio call-in — mental health phone-in
Releases pending
0:00 / 4:43

Live broadcast call-in. Host monologue on emotional load and help-seeking, station number read out repeatedly, then sequential live callers exchanging greetings with the host. Natural turn-taking, overlap on hand-offs, telephone-band caller audio against studio-band host audio.

ASR probe whisper base identified English at p=0.79 — hallucination loop — one sentence repeated 60x across 80 lines
broadcastcall-inmulti-speakernigerian-englishtelephonyspontaneous48 kHz complete · 4:43
tx-s002 · pcm-NG
Conversational Pidgin — trading exchange
Releases pending
0:00 / 5:26

Unscripted multi-party conversation in Nigerian Pidgin with English code-switching, recorded in a live acoustic environment. Negotiation and family-affairs register, long ambient stretches between exchanges, door and footfall transients, background chatter.

ASR probe whisper base identified Yoruba at p=0.52 — 28 of 29 segments returned as ambient noise
conversationalnigerian-pidgincode-switchedmulti-speakerambientspontaneous48 kHz complete · 5:26
tx-s003 · fr/en
Religious broadcast — altar-call register
Releases pending
0:00 / 1:12

Broadcast religious address in a call-to-response register, code-switched across French and English. Sustained single-speaker delivery with formulaic repetition, suited to prosody and register modelling.

ASR probe whisper base identified French at p=0.72 — decoded into French that was never spoken
broadcastreligiouscode-switchedfrancophonesingle-speaker48 kHz complete · 1:12
tx-s004 · und-NG
Extended in-language session
Releases pending
0:00 / 5:54

A near-six-minute continuous recording in a Nigerian language, captured in a live acoustic environment. Language and content are not yet human-verified. Open-weight ASR assigns Yoruba at p=0.45 and returns an unbroken hallucination loop — no usable text is produced from the entire recording.

ASR probe whisper base identified Yoruba at p=0.45 — hallucination loop across full duration
conversationalin-languageunverified-languagespontaneousasr-unresolved48 kHz complete · 5:54
tx-s005 · und-NG
In-language exchange
Releases pending
0:00 / 4:57

Just under five minutes of speech in a Nigerian language recorded in a live acoustic environment. ASR reports high confidence in Yoruba (p=0.90) yet emits a single nonsense token repeated some seventy times in one line — confident and wrong at the same time.

ASR probe whisper base identified Yoruba at p=0.90 — high-confidence language ID with nonsense output
conversationalin-languageunverified-languagespontaneousasr-unresolved48 kHz complete · 4:57
tx-s006 · und-NG
Continuous in-language delivery
Releases pending
0:00 / 4:04

Four minutes of continuous in-language speech. The decoder identifies the recording as Swahili at p=0.32 — an East African language, some three thousand kilometres from where this was recorded — then returns filler syllables repeated for the length of the file.

ASR probe whisper base identified Swahili at p=0.32 — language identified as Swahili; output is filler repetition
conversationalin-languageunverified-languagespontaneousasr-unresolved48 kHz complete · 4:04
tx-s007 · en-NG
Church service — closing announcements
Releases pending
0:00 / 1:06

Closing announcements from a church service: next week's order of service, an upcoming Thanksgiving service, and a transition into a closing hymn. Delivered in Nigerian English from a raised voice in a large reverberant room with congregation response.

ASR probe whisper base identified Yoruba at p=0.54 — language identified as Yoruba while the speech is English
broadcastreligiousannouncementnigerian-englishsingle-speakerreverberant48 kHz complete · 1:06
tx-s008 · und-NG
Ceremonial greeting exchange
Releases pending
0:00 / 1:43

A rapid exchange of greetings, acclamation and call-and-response around a ceremonial gift presentation, with reference to livestock and to Hausa address forms. Multiple speakers over crowd noise. Content is partially resolved and pending human verification.

ASR probe whisper base identified Yoruba at p=0.55 — partial decode; speakers and language not resolved
ceremonygreetingmulti-speakercode-switchedcrowdspontaneous48 kHz complete · 1:43
Access

Full-resolution masters are released under NDA.

Including metadata sidecars and the per-asset rights record. The machine-readable catalogue for this page is public: catalog.json.