A finished episode or chapter usually exists in two forms: a lossless master you keep and a compressed file you deliver. Choosing the right format for each avoids rejected uploads, needless quality loss and files larger than they need to be. This guide covers the formats, encoding settings and sample rates that podcast and audiobook platforms actually document.
Lossless and lossy formats
Lossless formats store every sample exactly. Lossy formats leave out detail the ear is unlikely to notice so the file is much smaller, and whatever they leave out is gone for good.
WAV
WAV holds uncompressed PCM audio. Every editor reads it, which makes it the safest format for masters and for moving files between tools. Its drawback is size: 24-bit stereo at 48 kHz takes about 1 GB per hour.
FLAC
FLAC is lossless compression. It decodes to exactly the same samples as the WAV it came from but takes less space; how much less depends on the material. It is a good archive format.
MP3
MP3 is the most widely supported lossy format, and the format ACX requires for audiobooks.
AAC
AAC is the codec Apple strongly recommends over MP3 for podcasts, ideally in an MP4 (.m4a) container rather than the raw ADTS format. Spotify accepts M4A files, and YouTube recommends AAC-LC or Opus for uploads.
What each platform asks for
| Platform | Formats | Sample rate | Bit rate | Notes |
|---|---|---|---|---|
| ACX audiobooks | MP3 | 44.1 kHz | 192 kbps or higher, CBR | No joint stereo; all files the same channel format |
| Apple Podcasts (RSS) | AAC recommended, MP3 accepted | 22.05/24 or 44.1/48 kHz | Mono 64–128 kbps, stereo 128–256 kbps at 44.1/48 kHz | MP4 container preferred over ADTS |
| Apple Podcasts Connect | WAV, FLAC or MP3 | WAV and FLAC at least 44.1 kHz | MP3 recommended: mono 96–128 kbps, stereo 128–256 kbps | WAV and FLAC: 16 or 24-bit, stereo only |
| Spotify episodes | MP3, M4A or WAV | Not stated | Not stated | Mono or stereo; no file-size limit |
| YouTube | AAC-LC or Opus recommended | 48 kHz | Mono 128 kbps, stereo 384 kbps | Recommended upload settings |
CBR or VBR
A constant-bit-rate (CBR) file spends the same number of bits on every second. A variable-bit-rate (VBR) file spends more on complex passages and fewer on simple ones, which can give better quality per megabyte. CBR is the more predictable of the two: the data rate never changes, so durations and seek positions follow directly from the file size.
ACX requires CBR MP3 so its "encoding process works without error". Apple and Spotify do not state a bit-rate mode for podcast episodes.
Choosing a bit rate
- ACX audiobooks: 192 kbps CBR MP3 is the minimum, and enough. ACX allows 256 or 320 kbps but says the difference heard by listeners will be negligible.
- Podcasts: stay within Apple's ranges, 64–128 kbps for mono and 128–256 kbps for stereo at 44.1 or 48 kHz. Going higher mainly makes larger downloads.
- Video: YouTube recommends 128 kbps for mono and 384 kbps for stereo audio in uploads.
Apple's mono ranges are half its stereo ranges, so channel count matters as much as codec for file size. Mono vs stereo podcasts explains that trade-off.
Sample rate and bit depth
44.1 kHz and 48 kHz both cover the full range of human hearing. ACX requires 44.1 kHz, YouTube recommends 48 kHz, and Apple accepts either. Record at the rate your main deliverable needs; if you have to convert, do it once, with a good resampler, before the final encode.
Bit depth applies only to uncompressed and lossless files. Record and edit at 24-bit for headroom. Apple Podcasts Connect accepts 16 or 24-bit WAV and FLAC and recommends 24-bit. ACX's current requirements page does not state a bit depth.
Joint stereo and ACX
MP3 can encode two channels in more than one way. Plain stereo encodes left and right separately. Joint stereo lets the encoder switch to a mid and side representation when that saves bits, which helps when the channels are similar. The ACX glossary is explicit: "ACX does not accept Joint Stereo files."
The trap is that joint stereo is often the default. ffmpeg's LAME-based MP3 encoder enables it for stereo unless you turn it off with -joint_stereo 0. An ACX-ready encode of a stereo chapter with ffmpeg looks like this:
ffmpeg -i chapter01.wav -ar 44100 -c:a libmp3lame -b:a 192k -joint_stereo 0 chapter01.mp3
Setting -b:a without a quality option such as -q:a produces a constant bit rate. Mono files have no joint stereo setting to worry about. Check the result with a media-info tool before uploading, and see ACX audio requirements explained for the full rule set.
Quick guide by project
- Talk podcast in an RSS feed: AAC (or MP3) within Apple's ranges, mono 64–128 kbps or stereo 128–256 kbps at 44.1 or 48 kHz.
- Audiobook for ACX: 192 kbps CBR MP3 at 44.1 kHz, mono or plain stereo, one chapter per file.
- Video voice track: AAC-LC at 48 kHz, at YouTube's recommended bit rates.
- Apple Podcasts Connect: 24-bit stereo WAV or FLAC at 44.1 kHz or higher, duplicating a mono master to two channels if needed.
- Your archive: 24-bit WAV or FLAC of every stage you might need again.
What to archive and what to deliver
- Archive the raw recordings, the edited mix and the final master as 24-bit WAV or FLAC. Platform requirements change, and a lossless master lets you deliver again without re-encoding a lossy file.
- Deliver the format each platform asks for, encoded once from the lossless master.
- Avoid lossy-to-lossy chains. Converting an MP3 to AAC, or re-encoding an MP3 at another bit rate, adds a second round of losses on top of the first.
Measure after encoding. Encoding can raise peak levels, so the true peak of a delivered MP3 or AAC file can be higher than the master's. See true peak vs sample peak.
LufsReady accepts WAV, MP3, M4A, FLAC, OGG, Opus, AIFF and video files with an audio track, and delivers MP3, AAC (M4A), WAV or FLAC. Its ACX preset always delivers 192 kbps CBR MP3 at 44.1 kHz without joint stereo, and the features page lists the other output options.
Sources
- ACX: Audio Submission Requirements
- ACX: Audio terminology glossary
- Apple Podcasts for Creators: Audio requirements
- Spotify for Creators: Publishing audio episodes
- YouTube Help: Recommended upload encoding settings
Check your file in seconds
Measure loudness, true peak and noise floor for free in your browser, or master a file to the exact target with a report.