MP3, WAV, M4A, or OGG? The Best Audio Format for Transcription
By the Speakmi team · Updated November 3, 2026 · 2 min read
You do not need studio-quality audio to get a good transcript. Speech recognition works best with clear voices, and the file format mostly affects size. Here is how the common formats compare.
The formats
- MP3: works everywhere. At 128 kbps it uses about 1 MB per minute.
- M4A (AAC): the default on iPhones and many Android recorders, with similar size and quality to MP3.
- WAV: uncompressed and large. Mono speech at 16 kHz is about 2 MB per minute, while stereo music quality at 44.1 kHz is about 10 MB per minute.
- OGG: efficient and open, but browser and app support varies.
What fits in the limit
This transcriber accepts files up to 25 MB and 20 minutes. A 128 kbps MP3 stays within the size limit for the full 20 minutes. A 16 kHz mono WAV fits roughly 13 minutes, and a CD-quality stereo WAV fits only about 2 minutes. That is why compressing a large file often solves upload problems.
Best settings for speech
- Mono instead of stereo, unless each person is on a separate channel.
- 16 kHz or higher sample rate.
- 64 to 128 kbps for MP3 or M4A.
Going below about 32 kbps makes voices sound thin and can reduce accuracy.
How to convert
Free tools such as Audacity and VLC can export MP3. Command-line users can run ffmpeg -i input.wav -ac 1 -ar 16000 -b:a 64k output.mp3. If your file is a video, see our guide to converting MP4 to text.
Quality beats format
A clear 64 kbps MP3 will transcribe better than a noisy WAV. Spend your effort on a quiet room and a close microphone first, and choose the format second.
Frequently asked questions
Which format is best for phone recordings?
Keep the original M4A or MP3. Converting between compressed formats does not add quality, so only convert when you need a smaller file.
Does a higher bitrate improve accuracy?
Only up to a point. For speech, beyond about 64 to 128 kbps a clearer recording matters far more than a higher bitrate.