Arabic quiz voiceovers: keep shuffled answers and feedback matched

The recording is correct. The option beside it is not.

A course producer records “B: five hundred millilitres.” The learning platform later moves that answer to position A. A learner now sees one label and hears another. An even less obvious fault is feedback that still says “B is correct” after the options move. Good pronunciation will not catch either mistake.

For Saudi training agencies or in-house course teams, settle the playback design before recording. Check what your authoring tool actually supports. Some can bind a separate clip to each answer; others play one track for the whole question. Do not assume that a platform’s shuffle setting also remaps audio.

A worked example: the ID follows the answer

This fictional product-training question is a production exercise, not an actual customer course: «أي سعة تطابق عبارة نصف لتر؟» The choices are 250, 500 and 1,000 millilitres. Use consistent spoken units in all three clips. The correct-answer rule points to the 500-millilitre option regardless of where it appears.

The IDs and filenames below are a suggested handoff convention in your own project, not a FasihTTS export format. Generate and review the approved lines, download the recordings, and map them in the authoring tool. Keep the text alternatives beside the audio.

Keep the correct choice out of the performance

Use the same voice and a similar instruction for every option. Do not ask for a confident, cheerful reading of the correct answer and a doubtful reading of the distractors. Listen to the options without viewing the answer key: does one have an extra pause, emphasis or explanatory phrase that gives it away?

A direction such as “read each option calmly and with similar emphasis” is a starting instruction, not a guarantee of identical delivery. Review the actual outputs. If you revise the wording or performance of one option, listen to the full set again. Keep the approved script and voice settings with the files.

Separate retry feedback from the final explanation

Feedback has two jobs that should not be mixed: help someone try again, or reveal and explain the answer. Decide when each is allowed in the assessment. Neither the audio generator nor a filename enforces that rule; configure and test it in the course.

Test the exported course, not just the MP3 files

• Preview several answer orders in the actual learner player. Select every answer at least once and check its visible text, recording, scoring result and feedback. A fixed preview order is not evidence that shuffled order works.
• Use replay, retry and Back/Next controls. Check that moving to another question stops the previous clip and that final-reveal feedback does not play during a permitted retry. These controls depend on your player.
• Test keyboard operation and visible text alternatives in the target course. Do not make hearing the recording the only way to understand a question. An audio file by itself does not establish accessibility compliance.
• After an edit, replace the intended file and verify the learner-facing version. Record the question ID, option ID, approved text and file version in the handoff. Check that the player has not retained an old cached recording.

Audition a voice before recording the question set

This is an existing Modern Standard Arabic recording of Omar, with its original transcript shown below. It is not a recording of the quiz example and does not demonstrate an LMS interaction. Use it to assess the voice, then test your own approved question and option lines.