Get a text with timestamps from MP3 files

By the Potok teamUpdated ~2 min read

An MP3 is cut into many tiny frames, so its own clock is not made of readable minutes. To get a text with timestamps from MP3 files, send the recording to Potok in Telegram as 1 file. You receive the transcript with times in minutes and seconds; compare 2 lines with the audio.

Turn a recording into text

In Telegram, write to @potok_now_bot and add your file. The finished text appears in the conversation when recognition ends.

Open @potok_now_bot in Telegram

About 30 languages · speaker labels and timestamps · files up to 2 GB · export to TXT, MD, DOCX, PDF, SRT and VTT

Key facts

  • Telegram Bot API accepts MP3 as an audio file and as a voice message.
  • The audio/mpeg media type covers MPEG audio layers I, II and III elementary streams and lists .mp3 among its file extensions.
  • An MPEG-1 Layer III frame covers 1152 samples (about 26 ms at 44.1 kHz), so time in MP3 is counted in frames.
  • Send .mp3 files up to 2 GB as they are: Potok decodes them without a separate conversion.

What a MP3 file is

PropertyValue
KindAudio format (MPEG audio, Layer III) with perceptual coding
File extensions.mp3
What it holdsAn MPEG audio elementary stream; layers I, II and III share the audio/mpeg media type

How to get timestamps for MP3 files

  1. Send the MP3 file to the bot. In Telegram, open @potok_now_bot and attach the .mp3 file to the chat. Larger recordings, up to 2 GB, go through the Upload button in the bot's mini app.
  2. Mind this about the format. An MPEG-1 Layer III frame covers 1152 samples (about 26 ms at 44.1 kHz), so time in MP3 is counted in frames.
  3. Switch on timestamps. Open the other-format options under the result, turn on timestamps (speaker labels can be added there too) and check three marks against the sound.
  4. Before sending. Send the MP3 as it is: re-saving it at a lower bitrate removes detail the recognizer may need.

What time marks look like

Example of a transcript with timestamps

[00:00] Speaker 1: Welcome everyone to the parents' evening.

[00:12] Speaker 1: First, the teacher will explain the new reading plan.

[00:31] Speaker 1: Questions are welcome after the short presentation.

[00:58] Speaker 1: The new library opens on Monday with a children's reading hour.

Lines written by us to show the layout. They are not output from a real MP3 file.

Potok at a glance
FeatureDetails
Where it worksIn Telegram, through @potok_now_bot; the bot's mini app takes large files
Upload limitFiles up to 2 GB
InputVoice messages, audio and video
Other languagesThe text keeps the language of the audio and is not translated
DigestA summary button under the transcript gives a short digest

Common mistakes with MP3 files

Where things go wrong
MistakeWhat happensFix
Expecting MP3 frame counts instead of readable timesYou get numbers that are hard to read in a transcriptUse the minutes and seconds shown in the bot text
Cutting the start of the MP3 after getting the transcriptEvery timestamp shifts and no longer matches the audioTrim the MP3 first, then request the transcript
Joining several MP3 files before sending themTimes run from one start, so finding a source file is hardSend each MP3 separately to keep its own times
Copying timestamps into notes without keeping the file nameLater you cannot tell which recording each time belongs toWrite the file name above every group of timestamps

Where to look next

Need a converter or a calculator? Open the tools. Unsure about a term? Check the glossary. Apps are on the download page.

Common questions

Does Potok accept MP3 files?

Yes. The bot accepts .mp3 files up to 2 GB, so no conversion is needed first.

Do the time marks depend on the MP3 format?

The marks point to a place in the recording so you can find it again. Check a few against the sound before you quote them.

Which apps and devices create MP3 files?

Telegram sendVoice accepts audio in MP3 format (Telegram sendVoice input), so MP3 reaches a bot as a voice message too.

Sources

Every source below was checked by hand on the date given. Sites change, so open the source if the exact wording matters.

  1. RFC 3003: The audio/mpeg media type, IETF · checked
  2. MPEG-1 Audio Layer III / MP3 (Library of Congress), Library of Congress · checked
  3. Telegram Bot API, Telegram · checked
  4. RFC 2250: RTP payload format for MPEG1/MPEG2 video and audio, IETF · checked

Try it on your own recording

In Telegram, write to @potok_now_bot and add your file. The finished text appears in the conversation when recognition ends.

Write to @potok_now_bot in Telegram