Lexical stress recording lab
Prepare one complete two-syllable word for the checkpoint test.
This page does not run the model or upload audio. Recording, trimming and WAV conversion happen in this browser. The exported clip must be tested using the private command-line kit.
1. Record or open audio
Record at a natural pace. The recorder stops after eight seconds. For a useful comparison, record the same word twice with different stress, in separate clips.
Ready.
2. Select the whole word
Remove surrounding silence, but keep both syllables intact. The selection must be 0.12–1.00 seconds for this prototype. Do not speed up a naturally longer word; use a shorter word for the first trial.
3. Export the selected WAV
Export format: 16,000 Hz, mono, 16-bit PCM WAV. The text label is not an acoustic judgement. Do not use student names in filenames.
What to do with the WAV
Upload the downloaded file through cPanel File Manager into the private directory
/home2/englishlab/stress-checkpoint-prototype/recordings, not into
public_html. Then follow START_HERE.html in the ZIP.
Your previous stress staging page is not replaced. This recorder has no analytics, external libraries, network uploads, or AI scoring. Reloading the page discards unsaved recordings.