Check an audio file's length, size and bitrate
Choose an audio or video file to see how long it is, how big, and what it is made of: format, sample rate, channels, bit depth and average bitrate. It also tells you whether the file is within the limits for transcribing without an account.
Loading the tool…
Processed in your browser. Your file is never uploaded.
What you get
The length comes from your browser's own player, so it is the real duration for any file the browser can open, video as well as audio. The rest comes from the first few kilobytes of the file, where most audio formats keep a header describing themselves:
- WAV: format, sample rate, channels, bit depth and exact length from the header.
- MP3: MPEG version, sample rate, channels and bitrate from the first audio frame, and whether it is a variable-bitrate file.
- FLAC: sample rate, channels, bit depth and total length from the stream information block.
- Ogg (Opus or Vorbis), M4A/MP4, Matroska/WebM, AMR: the container and, where the header allows, the codec, channels and sample rate.
When a number cannot be read it is simply left out. The page does not guess.
How bitrate relates to file size
Average bitrate is size times eight, divided by length. A 128 kbit/s MP3 is about 0.96 MB per minute. An uncompressed CD-quality WAV (44.1 kHz, 16-bit, two channels) is 1,411 kbit/s, about 10.6 MB per minute, which is why a short WAV can be bigger than an hour of compressed speech. Speech does not need that detail. A mono recording at 16 kHz is plenty for recognising words, and an M4A or MP3 of a voice recording is a small fraction of the WAV it came from.
Will it fit the free limits?
For a file used without an account the limits are 15 minutes and 25 MB, whichever you reach first. The checker at the top reads those numbers from the server, so it always uses the current ones.
If the file is too long, the choices are a free account or cutting the recording into parts. If it is only too big, a smaller copy usually solves it without losing speech clarity: the audio converter makes an M4A or MP3 in your browser, and video to audio pulls the sound out of a video, which is much smaller than the video itself.
Sample rate, channels and what matters for transcription
A higher sample rate or a stereo recording does not make a transcript better by itself. What matters is that the voice is close, loud enough and not buried under music or room echo. Two things in the details are worth noticing, though. A file with a very low bitrate (for example a heavily compressed voice message) has lost detail that cannot be brought back. And a stereo file with one speaker on each side is mixed down for transcription, so overlapping speech stays overlapped.
When the length is missing
Browsers play a limited list of formats. If yours cannot open a file (AMR phone recordings and WMA are the usual cases), the page still shows the size, the container and what the header says, and compares the size with the limit. Uploading the file to the transcriber works regardless, because the server decodes it with its own tools.
Questions
Is my file uploaded?
No. The page asks your browser for the length and reads the first 64 KB of the file to see its format. Nothing is sent anywhere, and the page reports only that the tool was used, with no file name or size.
Why does it say the bitrate is an average?
Because it is size times eight divided by length, which includes everything in the file: the audio, tags, cover art and container overhead. Compressed formats with a variable bitrate spend more bits on busy passages and fewer on silence, so there is no single true figure. For an uncompressed WAV the number is exact, because the header states it.
Why can the browser not read the length of my file?
It uses its own media player to get the length, and each browser plays a different set of formats. AMR, WMA and some old or unusual files are not playable in a browser, so no length comes back. That does not mean the file is broken or that it cannot be transcribed: the transcriber decodes on the server with its own tools.
Which limits does it compare with?
The no-account limits of the free preview, read from the server when the page loads. A free account has different limits, which are on the pricing page.
Have the recording instead of the text?
Scriptorium listens to an audio or video file and writes the transcript, with timestamps and an editor, then exports TXT, SRT, VTT or JSON. You can try it free.