Audio to text: MP3, M4A, WAV and more
Interviews, discussions or dictation: Themera converts English and German audio into text with speakers and timestamps. Here is which formats work and how to make large files small.
Transcribe 30 minutes free. No card needed.
Supported formats
Video files such as MP4 or MOV are not accepted. Export the audio first, as described below.
| Format | Typical source | Note |
|---|---|---|
| MP3 | Dictaphones, recorder apps, exports | Small and widely readable, good for long recordings |
| M4A | iPhone voice memos, Zoom audio file | Small, good speech quality |
| WAV | Field recorders, audio software | Uncompressed, so very large |
| FLAC | Audio software, archive copies | Lossless and smaller than WAV |
| OGG | Browser and open-source tools | Compact |
| WebM | Recordings made in a browser | Compact |
How big is an hour of audio?
The figures are bitrate times duration. For speech, a mono MP3 or M4A at 64 kbps is plenty; higher quality barely improves the transcript but slows the upload.
| Setting | Size per hour, rounded |
|---|---|
| WAV, 44.1 kHz, 16-bit, stereo | about 635 MB |
| WAV, 16 kHz, 16-bit, mono | about 115 MB |
| MP3 at 128 kbps | about 58 MB |
| MP3 or M4A at 64 kbps, mono | about 29 MB |
Converting video and large files
Audio editors such as Audacity can also export a recording straight to mono MP3.
- Open the file in the free VLC media player via Media, then Convert / Save.
- Choose the "Audio - MP3" profile and, if needed, a lower bitrate and one channel in the profile settings.
- Save the new file and upload it.
What decides transcript quality
- Clear speech and little background noise: noise, music and echo do the most damage.
- Distance to the microphone: quiet, distant voices are misheard more often.
- Overlapping speech: when several people talk at once, usually only one voice is transcribed.
- Strong accents and dense jargon: expect more corrections.
Frequently asked questions
Which languages are supported?
English and German. Choose the language of the recording when you upload.
How accurate is it?
With a clear recording the text reads well, but it is never error-free. Check every transcript against the audio before you quote or analyse it.
How long does it take?
Usually a few minutes. You can close the page; we email you when the transcript is ready.
Related
Further reading
Transcribe 30 minutes free
No card needed. Upload a recording, check the transcript, download it as Word, text or for MAXQDA.