Make speech or music smaller by choosing a level that fits the type of content.
EditFull practical documentation · edition reviewed: 26 October 2026
When this task helps
Speech recordings usually tolerate more reduction than music with complex detail and a wide frequency range.
Worked example
Twenty minutes at 192 kbit/s takes roughly 28.8 decimal MB; at 96 kbit/s, about 14.4 MB before small overhead differences. Trimming silence reduces duration; lowering bitrate reduces data per second.
What to check before starting
For messaging or email, start with moderate compression and listen before reducing the size further.
Choosing the settings
For speech, compare consonant clarity and noise. For music, check stereo and high frequencies. Leave room for metadata with a size target, and allow a larger result if speech quality suffers.
Limits that affect the result
WAV is not necessarily smaller. MP3 reduces data lossily and may offer little benefit when the source is already heavily compressed.
Mistake to avoid
Avoid forcing a very small target size on long recordings. The result can become harsh or lose clarity.
Illustrative example: size and quality depend on the source file.