How to Transcribe a 1-Hour (or Longer) Audio to Text Without Getting Stuck
Transcribing a two-minute clip is easy with any tool. The trouble starts when the recording runs an hour, an hour and a half, sometimes three: a full lecture, a conference talk, a long interview, an internal training session.
Long audio breaks in three different places: the file is too big to upload, the plan doesn’t cover the full length, and when the transcript finally comes out, nobody can read 9,000 words in one sitting.
This guide solves all three, in that order, using Sintesy.
What changes when audio runs past 1 hour
An hour of continuous speech turns into a lot of text. The most cited benchmark, from the National Center for Voice and Speech (US), is about 150 words per minute in conversation. Over an hour, that comes to roughly 9,000 words, something like 30 pages of running text.
Doing it by hand isn’t a realistic option either. Qualitative research guides usually estimate 4 to 6 hours of typing for every hour of audio, and that’s for people with practice.
So the challenge with long audio isn’t just “transcribing.” It’s:
- Being able to upload the file.
- Having the right plan for the length.
- Having a way to use the text afterward without rereading all of it.
How big is 1 hour of audio
File size depends on the bitrate, not on duration alone. The math is simple: kbps × 0.45 = megabytes per hour of audio.
| Format and quality | Bitrate | Size of 1 hour |
|---|---|---|
| Uncompressed WAV (CD quality, stereo) | 1,411 kbps | ~635 MB |
| High-quality MP3 | 320 kbps | ~144 MB |
| Standard MP3 | 128 kbps | ~58 MB |
| Voice M4A/AAC | 64 kbps | ~29 MB |
| Voice Opus/OGG | 32 kbps | ~14 MB |
Two numbers matter in Sintesy:
- Free plan: up to 20MB per file. In practice, long recordings only fit in heavily compressed formats, like Opus or voice AAC at a low bitrate.
- Pro plan: up to 2GB per file. That fits a WAV of more than three hours, or the MP4 of a meeting recorded on video.
If the file goes over the limit, the upload tells you: “Upgrade your plan to upload files larger than 20MB!”.
Tip: for transcription, music-grade quality doesn’t help. Voice at 64 kbps mono sounds the same to the transcription and takes up a tenth of a WAV. If you control the recording, record in M4A or MP3 instead of WAV.
Step by step: transcribing long audio in Sintesy
1. Upload the file
In the dashboard, click New and choose Upload file. Or just drag the file anywhere onto the dashboard screen: the upload opens on its own.
Sintesy accepts the most common audio and video formats, such as MP3, M4A, WAV, OGG, Opus, AAC, FLAC, MP4, MOV, and WEBM. The full list is in the guide on how to transcribe an audio file to text.
2. Wait for the upload to reach 100%
With a large file, this is the only part that depends on your connection. Use Wi-Fi, keep the tab open, and wait for the progress bar to finish. Once the upload is done, Sintesy opens the page for your new Sintesy.
3. Let the processing run
From there, the work happens on the server. The file is optimized (recompressed when that makes it smaller) and moves on to transcription and summary. You don’t need to watch it: the Sintesy shows up ready in your list when it’s done.
4. Check the language before you upload
Transcription uses the language set in Settings → Preferences → Sintesys language. If the recording is in Spanish or Portuguese and the setting is English, the text comes out wrong. For that case, see the guide on transcribing Spanish audio into English.
How to read a 9,000-word transcript
A long transcript with no structure becomes a file nobody opens again. The Sintesy page has four views, and each one solves part of the problem:
| View | What it’s for with long audio |
|---|---|
| Summary | Knowing in 2 minutes what the recording covers and what the main points were |
| Mind Map | Seeing the structure: which topic blocks exist and how they connect |
| Checklist | Pulling out tasks and next steps, when the content includes that kind of agreement |
| Transcript | Going to the exact passage when you need the literal quote |
The order that works: start with the summary, use the map to get your bearings, and only open the transcript to look for something specific.
For that search, open the Sintesy’s ”…” menu and choose Search. The in-document search highlights every occurrence of the word in the text, so you jump straight to the passage where someone mentioned “deadline,” “exam,” or a client’s name.
And when the question is more open-ended, use the chat: mention the Sintesy with @ and ask, for example, “what were the three objections the client raised?” or “what kind of question did the professor say will be on the exam?”. The answer comes from the content of the recording, without you rereading 30 pages. More on the views in the guide on turning a transcript into a summary, mind map, and checklist.
Free or Pro plan for long audio?
It’s worth being honest here: audio of an hour or more is where the free plan feels tightest.
| Free | Pro | |
|---|---|---|
| Size per file | up to 20MB | up to 2GB |
| Content per day | up to 1 hour | no daily limit |
| What you see of the Sintesy | preview (the start of the summary and text) | full content |
The free plan works for testing the flow with a short clip and checking whether the transcription holds up on your kind of audio. For full-length recordings with the complete text, the path is the Pro plan; see current pricing on the plans page. If you want to compare against human transcription and other tools, the numbers are in how much audio transcription costs in 2026.
When it’s worth splitting the recording
Splitting isn’t required, but it helps in three situations:
- Unstable connection. A 1.5GB file that drops halfway through the upload is more frustrating than three 500MB files.
- Different content in the same file. A full day of an event with four talks works better as four Sintesys, each with its own summary and map.
- Parts you’ll look up separately. Course modules, research sessions, days of a training program.
Cut at natural breaks (coffee break, speaker change, topic shift), not into fixed-size chunks. Each part becomes its own Sintesy, and in the chat you can mention several of them with @ to ask about the whole set.
Common mistakes with long audio
- Recording in WAV “to be safe on quality.” It only makes the file bigger. For voice, M4A or MP3 does the job.
- Recorder far from the speaker. In an hour-long talk, the problem is almost never the tool: it’s the microphone at the back of the room. See why the transcript comes out wrong.
- Wrong language in preferences. Check it before uploading, not after.
- Opening the transcript first. Start with the summary and the map. The transcript is for reference, not for reading straight through.
If the long recording is a YouTube live stream or video, you don’t need to download anything: see how to transcribe and summarize a YouTube live stream.
Quick recap
- 1 hour of speech ≈ 9,000 words. By hand, that’s 4 to 6 hours of typing.
- File size: kbps × 0.45 = MB per hour. Record voice in M4A or MP3, not WAV.
- In Sintesy: New → Upload file (or drag and drop), wait for the upload to reach 100%, and let the server do the work.
- Read in the order summary → mind map → transcript search, and use the chat with @ for questions.
- A full-length recording with the complete text calls for the Pro plan.
