Control temporal information and resource use.
1. Audio preparation
Preserve the required sample format and distinguish transcription from reasoning over audio. Show capture state and allow the user to inspect or correct recognized text where appropriate.
2. Video sampling
Select frames or segments according to the task and supported API. Repeated nearly identical frames add cost without necessarily adding information.
3. Timestamps and freshness
Associate media with capture time and avoid implying that an old frame reflects the current scene. Synchronize supporting text with the media it describes.
Worked scenario
A user dictates a note after moving to another object. The app must associate the note with the intended photo, not simply the most recent camera frame.
Apply it
Design a media bundle with explicit IDs and timestamps, then test reordered upload completion.
Check your understanding
Asynchronous media processing cannot attach evidence to the wrong note. Explain the decision and show evidence from your implementation or design. If you cannot demonstrate it yet, revisit the relevant section before continuing.