Skip to main content

Speech and sign-language collection

Ticks are recorded live

Work these steps on the in-app checklist — account menu → My testing checklist — where every Pass / Fail / Skip is recorded the moment you press it, and Fail files the report for you. This page (and its PDF) is the same list for reading, printing, or working offline; if you used paper, transfer your ticks into the app afterwards.

AudioRecord / VideoRecord projects, QC, prompts, Listen validation. Fixtures: 35, 36 (speech), 37 (sign).

Each step: Do this, then check You should see.

A. Recording​

  • SP-01 Open the read-aloud project as a contributor and record the first prompt. You should see: a big record button; the timer counts; playback works; Re-record discards.

  • SP-02 Record while offline on the phone, then reconnect. You should see: the clip queues and uploads later; no data loss.

  • SP-03 Record on the sign-language project (VideoRecord) — allow the camera. You should see: camera preview, recording, playback with duration/resolution; optional skeleton overlay draws on top.

B. QC​

  • SP-04 Submit a near-silent clip. You should see: it is flagged (silence / low speech ratio) — a toast explains; the manager sees the flag.

  • SP-05 Settings → Audio QC appears for the VIDEO project too. You should see: the tab is there; 'Run automated checks' toggles.

C. Prompts and speakers​

  • SP-06 Create a speaker profile and submit prompts via the prompt collector. You should see: one prompt per speaker rule enforced; cohort/leaderboard updates.

D. Listen validation​

  • SP-07 Set validation mode to Crowd; as another contributor open Listen. You should see: clips to vote Yes/No/Report; the vote threshold resolves clips.

  • SP-08 Set mode to Hybrid; as a reviewer open Listen. You should see: a 'Re-check crowd verdicts' queue and Validate/Reject override buttons; overriding closes the crowd report and credits the speaker.

  • SP-09 Empty the vote queue. You should see: the page still offers the queue switch so reviewers can reach the re-check queue.

E. Export​

  • SP-10 Export the speech project as ASR_MANIFEST and as a media ZIP. You should see: manifest lines per clip; ZIP contains annotations.json + media/.