Leave room to answer in a listening exercise

Decide what the learner should hear and when

Suppose a French tutor records twelve practice items in one sitting. In item four she asks « Où est la gare ? », leaves four seconds for a reply, then models « Elle est à gauche. » During that pause the learner is doing the exercise. A cleaned-up recording that removes the quiet, or a caption that shows both sentences at once, changes the lesson.

Ask the tutor which pauses belong to the exercise, whether a prompt should remain visible, and when the model may appear. For this example she wants a clear screen during the reply. A separate, immediately available reading transcript can show the whole exchange; label it so learners can choose it. This is a practice activity, not guidance on withholding captions or accommodations in an assessment.

Use recordings you own or have permission to process and distribute to this class. A teacher’s master may include retakes, a publisher’s licensed dialogue or students speaking between items. Get a cleared edit if those voices or materials are outside your permission.

In Audacity, mark a region for each practice file

Work on a copy of the approved master. In Audacity, place a region label over each complete exercise and give it a stable name such as station-04. End the region after the model’s last syllable; leave the next take and the tutor’s direction outside it. A point label only marks a start. Explicit regions are useful when the parts between exercises should be discarded.

In Export Audio choose Multiple Files, split by Labels and use the label names. Audacity’s manual says only the uppermost label track supplies the export boundaries; highlighting a waveform does not set them. Place that label track below the intended audio tracks and mute any unwanted track above it. Keep the option to include audio before the first label off if the introduction is private. Export to an empty revision folder. Audition the resulting files.

An Audacity label file stores boundaries and names, not subtitles. A label for the entire drill cannot say when its two lines should appear. Keep the region list with the master and give the caption reviewer the final audio files. Polyglot does not split Audacity projects or attach text to a course player.

Put the model answer on the clip’s clock

Open station-04’s exported audio in Polyglot on an Apple Silicon Mac running macOS 14.4 or later. Transcribe the clip locally in a supported speech language. Compare the words against playback, including small words such as « à » that matter to the teaching point. If you already have a reviewed SRT for this exact clip, use that instead. A full-session track still uses full-session times.

Here is a fictional timing sheet. The region begins 72 seconds into the master. The model starts at 78.5 seconds on that clock, but only 6.5 seconds into the file the learner opens. Listen for the actual first word after export; trimming or adding a lead-in changes the calculation.

station-04.wav · 9 s
Master region: 01:12–01:21

On the clip clock:
00:00.5–00:02.5  prompt
00:02.5–00:06.5  reply pause
00:06.5–00:08.5  model answer

Keep the pause in the exported text

If recognition puts the prompt and answer in a single cue, correcting the sentence alone will still reveal the answer early. Use separate timed cues. Adjust existing boundaries in the Polyglot timeline; if a cue needs splitting, use a subtitle editor that supports the split and recheck the exported track. Do not invent words for the silence. If a bell signals when to speak, describe that audible signal where it occurs.

Export a WebVTT for a player that accepts it, or SRT if the course author requests it. Give each file the clip identifier, language and revision. Provide reviewed plain text alongside the audio when timed text is unsupported or learners want a reading copy. An audio player does not necessarily display a VTT just because it sits beside the recording.

Try item four without playing items one through three

In the learner view, open item four directly with its captions selected. Listen for « Où est la gare ? », watch the entire response interval, then compare the start of « Elle est à gauche » with the first audible syllable. Replay after seeking into the gap. If the phrase appears early, check its cue start; if the voice itself starts early, return to the audio edit. Sliding captions cannot restore a missing pause.

Finally open the adjacent item and its reading transcript. Check that both belong to that item, and that the files contain no coach’s direction or discarded take. If a tutor re-records a model at a different pace, replace and recheck that audio/text pair together. A correct first item tells you little about a later item loaded on its own.

Keep exploring.