What sample rate should I use for a podcast? 48 kHz in, 44.1 kHz out

Apple will not accept a WAV or FLAC file for subscriber audio below 44.1 kHz, and YouTube asks for 48 kHz on every upload it processes. That gap is the entire argument, and it has a short answer.

What sample rate should I use for a podcast? Record at 48 kHz, 24-bit, and export the file that goes in your feed at either 44.1 or 48 kHz. Recording at 48 means nothing gets resampled on the way to YouTube, and the figure sits inside Apple's recommended specification either way. If your show has run for three years at 44.1 kHz and nobody has complained, leave it alone. None of this is worth re-recording an episode over.

Any show can be listed on PodAnswer for free, and a listed show gets its own page that can surface when somebody searches the show's name, so you can claim yours at podanswer.com/signup.

The two tables Apple publishes, and the gap between them

Apple's audio requirements document splits into two paths, and hosts conflate them constantly.

The first path is subscriber audio uploaded through Apple Podcasts Connect. Apple accepts WAV, FLAC or MP3 there. For WAV and FLAC it prints a minimum of 44.1 kHz at 16-bit or 24-bit, and a recommended specification of 44.1, 48, 88.2, 96, 176.4 or 192 kHz at 24-bit. It also refuses single-channel WAV and FLAC outright: where no stereo source exists, Apple tells you to send two identical channels rather than one. For MP3 on that path, mono has a floor of 44.1 kHz at 32 kbps and a recommendation of 44.1 or 48 kHz at 96 to 128 kbps, while stereo has a floor of 44.1 kHz at 64 kbps and a recommendation of 44.1 or 48 kHz at 128 to 256 kbps.

The second path is the file sitting at the end of the URL in your RSS feed, which is what almost every independent show actually ships. Apple accepts MP3 or AAC there and publishes bit rates against two sample-rate columns. At 22.05 or 24 kHz, mono wants 40 to 80 kbps and stereo wants 80 to 160. At 44.1 or 48 kHz, mono wants 64 to 128 kbps and stereo wants 128 to 256. Apple recommends AAC over MP3 for feeds, and within AAC it recommends the MP4 container over ADTS, on the grounds that MP4 streams more efficiently and seeks accurately.

Read the two tables together and the practical range narrows fast. The feed table treats 44.1 and 48 kHz as one bucket with identical bit-rate guidance, which is Apple saying, in the only way a specification can, that it does not care which you pick.

48 kHz is the number video chose, and video now sets the floor

YouTube's recommended upload encoding settings name a single audio sample rate: 48 kHz. Not a range, not a minimum. The same document asks for AAC-LC, Opus or Eclipsa Audio as the codec, stereo channels, and audio bit rates of 128 kbps for mono, 384 kbps for stereo and 512 kbps for 5.1.

That one line decides the question for most shows, because most shows now put some version of the episode on YouTube. Record at 44.1 kHz and upload there, and something in the chain resamples your audio to 48 kHz without telling you which algorithm it used. Record at 48 and the encoder takes the file as it is. The audio in your feed can still be 44.1 kHz if you prefer, because that conversion happens once, offline, in your editor, where you can hear it before anyone else does. Our walkthrough of the upload path is in how to put your podcast on YouTube.

The 44.1 figure comes from the compact disc; 48 comes from professional video and film, where it has been the production standard for decades. Podcasting spent fifteen years at 44.1 because music software defaulted to it. Video pulled the default the other way.

What the extra 3,900 samples a second actually buy

The arithmetic is not in dispute. A sample rate captures frequencies up to half its own value, so 44.1 kHz carries content to 22.05 kHz and 48 kHz carries content to 24 kHz. Both sit above the top of adult human hearing. Almost all speech intelligibility lives between roughly 100 Hz and 8 kHz, so the difference between those ceilings is inaudible in a spoken-word mix.

So the case for 48 kHz is not fidelity. It is three mundane things. No resampling, because every rate conversion is a filtering operation and the one inside a video platform's ingest pipeline cannot be inspected. Slightly more room for frequency-domain processing like de-noising and time stretching, which matters more for music than for speech. And one setting everywhere, so a session, a camera and an upload all agree, which removes a whole category of mistake.

Capture decides more than any of it, which is the point engineers keep making on shows like Recording Studio Rockstars. The answer page on getting a deep, punchy snare sound is a good worked example of that order of operations, and the show itself is listed at Recording Studio Rockstars.

Bit depth does the work people credit to sample rate

Apple recommends 24-bit for WAV and FLAC and accepts 16-bit as the floor. That recommendation pulls a different lever.

Sample rate sets the frequency ceiling. Bit depth sets the distance between the quietest sound you can record and the point where the signal clips. Sixteen bits gives you about 96 dB of that range; 24 bits gives you about 144. For a podcast the difference is not about hearing quiet detail. It is about recording a guest conservatively, with peaks at -18 dBFS instead of -6, then lifting the level in the edit without dragging the noise floor up into audibility with it.

At 24-bit the safe, quiet take costs you nothing. At 16-bit it forces a choice between clipping risk and hiss. Where hiss or room noise is already baked into recordings you have, the repair options and the limits on them are in how to remove background noise from podcast audio. Interview shows should default to 24-bit, then deliver at whatever the destination requires.

Two microphones, two clocks, one episode

Sample rate stops being an abstraction the moment two people record separately. A double-ender produces two files from two machines, and if one is set to 44.1 kHz and the other to 48, importing both into a single timeline means one of them gets converted by your editor, which is fine, or played back at the wrong rate, which is not. The second failure sounds like a voice shifted slightly in pitch and running slightly fast, and it drifts further out of sync as the episode goes on.

Settle it before the session rather than after. Agree the rate in writing, send it with the rest of the recording instructions, and have both parties confirm it in their software before anyone starts talking. The full procedure for remote sessions, including the local-recording setup that makes this necessary, is in how to record a podcast with a remote guest.

A back catalog recorded at mixed rates needs no retroactive fix. Set the standard now.

The settings, in order

1. Set your recording software's project rate to 48 kHz and its bit depth to 24-bit, before you create the session rather than after.
2. Set every interface, recorder and camera in the chain to the same 48 kHz. A mismatch here is where most drift problems are born.
3. Record and edit at that rate. Convert nothing mid-project.
4. Master to Apple's published loudness target: overall loudness around -16 dB LKFS with a tolerance of plus or minus 1 dB, and a true peak no higher than -1 dB FS, both calculated per ITU-R BS.1770-5. Apple is explicit that this conditioning happens before encoding, because compression algorithms typically do not modify loudness and may clip the signal if the true-peak figure is not respected. The longer treatment is in how loud should a podcast be.
5. Export the feed file as AAC in an MP4 container at 44.1 or 48 kHz. Mono at 64 to 128 kbps, stereo at 128 to 256 kbps.
6. Export the YouTube file separately at 48 kHz, AAC-LC, stereo, 384 kbps.
7. Keep the 24-bit session master. It is the only version you can go back to.

The case for going lower

Apple's own guidance includes an instruction that reads strangely until you have shipped a long archive: if the bit rate is constrained, precondition the signal by reducing the sample rate, for example from 48 to 24 kHz, or by reducing stereo to mono, or both.

That path has a use. A talk show with no music, hundreds of episodes and listeners on metered connections can ship mono AAC at 22.05 or 24 kHz inside Apple's 40 to 80 kbps band and sound acceptable, at roughly a third the file size of a 128 kbps stereo file. Smaller files also change how downloads get counted, since the measurement standards are written against bytes delivered rather than minutes played, which we worked through in what counts as a podcast download.

The trade is real too. Cymbals, music beds and sibilance all degrade noticeably below 44.1 kHz. If the show carries music, stay at 44.1 or 48 and cut bit rate instead.

Key takeaways

Record at 48 kHz and 24-bit, and make every device in the chain agree on it. Deliver the feed file at 44.1 or 48 kHz as AAC in an MP4 container, since Apple's published bit-rate guidance treats those two rates identically. Deliver the YouTube file at 48 kHz, the only rate that document names. Match rates across both sides of a double-ender before the session starts. Drop to 22.05 or 24 kHz mono only for a speech-only archive where file size is the binding constraint.

Getting your episodes in front of people searching Google is the business PodAnswer is in: we build answer pages from your transcripts, each one linking back to the episode it came from. Growth covers four articles a month at $149 and Network covers twelve at $349, with the details at podanswer.com/for-podcasters and the plan comparison at podanswer.com/pricing.

FAQ

What sample rate should I use for podcast recording? 48 kHz at 24-bit for the session. It matches the only rate YouTube's encoding document names, it sits inside Apple's recommended specification, and it means no conversion happens anywhere in the chain without your knowledge.

What sample rate should I use if I only publish to an RSS feed? Either 44.1 or 48 kHz. Apple's feed table puts both in the same column with the same bit-rate guidance, so the choice has no consequence for that destination. Record at 48 anyway in case a video version happens later.

What sample rate is best for recording speech? 48 kHz, for workflow reasons rather than audible ones. A spoken voice carries almost nothing above 8 kHz, so both common rates capture it completely.

What sample rate do professional producers use? Film and television post-production runs at 48 kHz and has for decades, which is why YouTube specifies it. Music production splits between 44.1, 48 and higher rates depending on the delivery format. Spoken-word podcast production has largely converged on 48.

Is 44.1 kHz bad for a podcast? No. It is Apple's published minimum for WAV and FLAC and it shares a column with 48 kHz in the feed table. An existing show at 44.1 kHz has no reason to change.

Does a higher sample rate make a podcast sound better? Not for speech. Apple's recommended list runs up to 192 kHz, but those rates exist for music and archival work. Above 48 kHz you pay in file size and processing time for frequency content nobody can hear.

Sources: Apple's audio requirements for Apple Podcasts, published at podcasters.apple.com/support/893-audio-requirements, and Google's recommended upload encoding settings for YouTube, published at support.google.com/youtube/answer/1722171. Loudness and true-peak figures are calculated per ITU-R BS.1770-5 as cited in Apple's document.