File formats
Audio and video formats we transcribe
Scriptorium reads the audio in all the common recording and video formats directly, with no conversion first. Pick yours below for what to expect: where that kind of file comes from, how big it gets, and what goes wrong.
Updated
Audio formats
An audio file is just a container plus a way of squeezing the sound. For transcription the squeeze rarely matters, because the recogniser works on speech-band audio. The things that matter are how clear the voice is and whether the file fits the upload limits.
| Format | Usually comes from | Size |
|---|---|---|
| M4A to text | iPhone Voice Memos, Zoom audio-only, QuickTime | Compressed (AAC), or lossless if chosen |
| MP3 to text | Podcasts, downloads, dictation, voicemail exports | Compressed; 128 kbps is about 1.0 MB a minute |
| WAV to text | Audio editors, field recorders | Uncompressed; CD-quality stereo is 10.6 MB a minute |
| FLAC to text | Archives, digitised interviews | Lossless, smaller than WAV |
| OGG to text | Telegram, Audacity, open-source tools | Compressed (Vorbis or Opus) |
| Opus to text | WhatsApp voice notes on Android and desktop | Compressed, tiny |
| AAC to text | Streams, plain recorders | Compressed |
| AMR to text | Old Android recorders, call-recording apps | Narrowband speech; at most 0.09 MB a minute |
Video formats
From a video, only the sound is transcribed. That is why a huge 4K file and a small phone clip give the same text if the audio is the same, and why extracting the audio first is the practical way past the size cap.
| Format | Usually comes from | As subtitles |
|---|---|---|
| MP4 to text | Zoom, Teams, Meet, phones, editors | MP4 to SRT |
| MOV to text | iPhone camera, Mac screen recordings | MOV to SRT |
| WebM to text | Browser screen recorders, web tools | WebM to SRT |
| MKV to text | OBS, downloads | MKV to SRT |
| 3GP to text | Older phones | Export SRT from the transcript |
Subtitles from a recording
If the goal is a subtitle file rather than a script, each of these makes an SRT (one cue per transcript segment, with start and end times): MP4 to SRT, MOV to SRT, WebM to SRT, MKV to SRT, M4A to SRT, WAV to SRT. The general audio and video to SRT and audio and video to VTT pages cover everything else.
Which format should I record in?
If you are choosing: M4A or MP3 for voice is plenty; WAV or FLAC only if you want a master to keep; never use AMR if you can avoid it, because its narrow bandwidth costs accuracy. Place the microphone close, avoid speakerphone, and record in a quiet room; those choices matter more than any format. The accuracy guide has the details.
Limits that apply to every format
Without an account the preview takes one file at a time, up to 15 minutes and 25 MB, 3 tries a day. A free account lifts the preview limits; see pricing. If a file is over the size cap, the audio converter and video to audio tool make a smaller copy first.
Formats you can get out
The transcript exports as plain text, timestamped text, SRT, VTT, Markdown, JSON with word timings and CSV. If you need to convert between subtitle formats or tidy them, the free SRT to VTT and SRT to text tools run in your browser.
If your file type is not here
The list above is what we describe, not a hard limit. Other video and audio containers the server can decode (AVI, WMV, FLV, MPEG, AIFF, CAF, M4V, TS and more) are accepted; if one is rejected, convert it to M4A or MP4 first.
Questions
Do I need to convert my file before uploading?
Which format is the best for transcription?
Can I upload a video to get audio transcribed?
Is it free?
Try it with your own file
Upload a recording and read the transcript.