File size comes from one number: the bitrate. Multiply the bits per second by the number of seconds, divide by eight, and you have bytes. An hour is 3,600 seconds, so a 128 kbps MP3 hour is 128,000 × 3,600 ÷ 8 = 57,600,000 bytes, or 57.6 MB. Every figure below uses that arithmetic, with 1 MB = 1,000,000 bytes and 1 GB = 1,000 MB.
Compressed audio: MP3, AAC, OGG and Opus
Lossy formats store a fixed or average bitrate you choose when you export. The format matters less than the bitrate: an MP3, an M4A and an OGG file at 128 kbps are all about 57.6 MB per hour. Voice recorders, phone memo apps and most podcasts sit between 64 and 128 kbps.
| Format | Bitrate | Per hour |
|---|---|---|
| Opus (voice) | 32 kbps | 14.4 MB |
| MP3 or Opus | 64 kbps | 28.8 MB |
| MP3 | 128 kbps | 57.6 MB |
| AAC / M4A | 128 kbps | 57.6 MB |
| OGG Vorbis | 128 kbps | 57.6 MB |
| MP3 | 192 kbps | 86.4 MB |
| AAC / M4A | 256 kbps | 115.2 MB |
| MP3 | 320 kbps | 144 MB |
Measured on production, 30 days to 2026-10-16: the median MP3, M4A and AAC file uploaded to WhipScribe came to 58 MB per hour, and the median OGG file to 40 MB per hour.
Uncompressed audio: WAV
WAV stores every sample, so its size is fixed by three settings: sample rate, bit depth and channels. Multiply them together for the bitrate. CD quality is 44,100 samples × 16 bits × 2 channels = 1,411.2 kbps. A mono file is half the size of the same file in stereo, and speech recorded for transcription is often 16 kHz mono.
| WAV setting | Bitrate | Per hour |
|---|---|---|
| 16-bit, 16 kHz, mono | 256 kbps | 115.2 MB |
| 16-bit, 44.1 kHz, mono | 705.6 kbps | 317.5 MB |
| 16-bit, 44.1 kHz, stereo | 1,411.2 kbps | 635 MB |
| 24-bit, 48 kHz, stereo | 2,304 kbps | 1.04 GB |
| 32-bit float, 48 kHz, stereo | 3,072 kbps | 1.38 GB |
Measured on production, 30 days to 2026-10-16: the median WAV uploaded to WhipScribe came to 115 MB per hour, which matches 16-bit 16 kHz mono.
Lossless compressed audio: FLAC
FLAC keeps every sample of the WAV but packs it smaller. How much smaller depends on the sound: music with a lot going on compresses less, and speech with pauses compresses more. A fair planning figure is about 55% of the WAV, so roughly 349 MB per hour for 16-bit 44.1 kHz stereo. Treat it as an estimate and check the real file.
Video: 720p, 1080p and 4K
Video bitrates are far higher than audio, and they vary by camera, frame rate and how much moves in the picture. The rows below use common bitrates and add a 128 kbps AAC audio track.
| Video | Bitrate (video + audio) | Per hour |
|---|---|---|
| 720p | 5 + 0.128 Mbps | 2.31 GB |
| 1080p | 8 + 0.128 Mbps | 3.66 GB |
| 1080p, phone or camera | 16 + 0.128 Mbps | 7.26 GB |
| 4K | 40 + 0.128 Mbps | 18.06 GB |
Measured on production, 30 days to 2026-10-16: the median MP4 uploaded to WhipScribe came to 192 MB per hour and the median MOV to 326 MB per hour. Most of those are meeting and screen recordings, which use a fraction of a camera's bitrate.
What that means for a 2, 5 and 10 hour recording
Long recordings are where the format starts to matter. Here are the same formats at the lengths people actually record: a long interview, a full workshop day, an all-day hearing or event.
| Format | 2 hours | 5 hours | 10 hours |
|---|---|---|---|
| MP3 / AAC, 128 kbps | 115 MB | 288 MB | 576 MB |
| MP3, 320 kbps | 288 MB | 720 MB | 1.44 GB |
| WAV 16-bit 16 kHz mono | 230 MB | 576 MB | 1.15 GB |
| FLAC (about 55% of WAV) | about 0.7 GB | about 1.75 GB | about 3.49 GB |
| WAV 16-bit 44.1 kHz stereo | 1.27 GB | 3.18 GB | 6.35 GB |
| WAV 24-bit 48 kHz stereo | 2.07 GB | 5.18 GB | 10.37 GB |
| Video 1080p, 8 Mbps | 7.32 GB | 18.29 GB | 36.58 GB |
Highlighted cells are over 5 GB.
For transcription with WhipScribe, none of this needs compressing or splitting. One file up to 10 hours or 5 GB uploads as it is, so every compressed audio format fits a full 10 hours in one file, and so does FLAC. A file over 2 GB uploads the same way, up to 5 GB, and transcription starts with a free preview of the opening minutes. The highlighted cells are the cases that need one extra step. For a recording longer than 10 hours, send two files. There is more on both on the pages for transcribing large files and transcription for long recordings.
When to export the audio track
Export the audio when a video is over 5 GB and is not somewhere you can link to. The audio track is the same size whatever the resolution: about 57.6 MB per hour as AAC at 128 kbps. Ten hours of 4K video, about 180 GB, has an audio track of about 576 MB. Transcription only reads the sound, so nothing is lost.
If the video is already online, skip the export and paste the link. WhipScribe fetches YouTube, Vimeo, Twitch and Kick VODs, Google Drive files and podcast or RSS episodes on its side, so there is nothing to upload. For a file on your computer, most video editors export audio on its own. If the video is an MP4 or MOV with AAC sound, ffmpeg copies the track out in seconds without re-encoding it:
ffmpeg -i video.mp4 -vn -c:a copy audio.m4a
Below 5 GB, do not bother. Upload the video as it is.
Work out your own file
For any length and bitrate, use the audio file size calculator or the video file size calculator. To see whether a specific recording uploads as is, how long transcription takes and which pack covers it, use the large recording calculator. Processing runs at about 2 minutes per hour of audio. The opening minutes are a free preview; packs are $4 for 500 minutes, $12 for 2,000 minutes and $24 for 5,000 minutes, and credits never expire.
Recordings are transcribed on our own servers and are never used for training.