01
Choose a source worth preserving
Start from the original recording when possible. Converting a compressed copy again adds another encoding step, so keep that original for future exports. The current MP3 page uses the LAME encoder's quality mode with a fixed setting; it does not expose an exact bitrate selector.
For a meeting or interview, make sure the source has the intended mix before adding it. There is no track picker or speech-enhancement control. If one channel contains an interpreter or a second language, prepare a source with the correct track rather than relying on automatic stream selection.
02
Judge the listening copy by listening
Play a quiet sentence, a loud passage and the last few seconds. Check that the duration is complete and that background sounds have not become distracting. A smaller file is not useful if its speech is harder to understand.
Source descriptive metadata is removed, so use a meaningful downloaded filename or add library tags afterward. Several selected recordings are converted as separate jobs; the page is not an audio merger. If a lengthy video fails while a short one works, browser memory or conversion time may be the limiting factor rather than MP3 support.