DistroKid, TuneCore, CD Baby: Audio Formats for Streaming Distribution

You finally decided to get the old EP onto Spotify and Apple Music, opened a distributor's upload page, and hit a wall: they want lossless audio, and all you can find is the folder of MP3s from 2014. The WAV masters lived on a hard drive that died, a laptop that got sold, or a studio that closed. This situation is extremely common, and the way through it depends on understanding two things: what each distributor actually accepts, and what is genuinely possible when the lossless masters are gone.

Why distributors insist on lossless

DistroKid, TuneCore, and CD Baby do not stream your file to listeners directly. They deliver your audio to dozens of stores and streaming platforms, and each platform transcodes it again into its own delivery formats — Ogg Vorbis for Spotify, AAC for Apple Music, and so on. Every lossy-to-lossy transcode stacks generation loss: artifacts that were inaudible in your MP3 become audible after a second encode. Starting from lossless means every platform's version is a first-generation encode. That is the entire logic behind the requirement — the FAQ on lossy vs lossless audio covers the mechanics.

What each service accepts

DistroKid

The most permissive of the three. DistroKid accepts WAV and FLAC, and will also take M4A and even MP3 uploads — but uploading an MP3 means every store gets a transcode of a transcode. Treat the MP3 door as an emergency exit, not an entrance. WAV at 16- or 24-bit, 44.1 kHz, stereo is the safe default.

TuneCore

The strictest. TuneCore wants stereo WAV, 16- or 24-bit, at 44.1 kHz. Files at other sample rates or in compressed formats get rejected or flagged at upload validation, so resolve format questions before you start filling in release metadata.

CD Baby

Sits in the middle: WAV is the recommended upload, FLAC is accepted, and the platform has historically tolerated MP3 uploads with a quality warning. As with DistroKid, "accepted" and "advisable" are different words.

Specs shift over time, so check the current requirements page before a release — but the pattern has been stable for years: lossless in, or accept degraded delivery.

Getting your files into shape

Assuming you have lossless sources, the job is just container translation:

• FLAC masters from a mastering engineer or a Bandcamp backup: upload directly where accepted, or decompress to WAV — FLAC is lossless, so the round trip is bit-perfect. (For your own listening copies, a FLAC to MP3 conversion handles it; the FLAC FAQ explains why the archive copy should stay FLAC.)
• ALAC (Apple Lossless) files ripped into an old iTunes library are also bit-perfect sources — convert them to WAV for upload, or make ALAC to MP3 listening copies for the road. Do not confuse ALAC .m4a files with lossy AAC .m4a files: same extension, very different contents. A lossless rip of a 4-minute song is typically 20-40 MB; a lossy one is 4-10 MB.
• WAV masters at 48 or 96 kHz: fine for DistroKid; for TuneCore, resample to 44.1 kHz before upload.

When only the MP3s survive

Here is the honest part. If the masters are gone, no conversion can bring back the data the original MP3 encode discarded. Converting a 320 kbps MP3 to WAV produces a perfectly valid file that sounds identical to the MP3 — a lossless wrapper around lossy audio. Your realistic options, in order of preference:

• Exhaust the search first. Mastering engineers keep archives for years, as do pressing plants and old bandmates' hard drives. Bandcamp stores the lossless file you originally uploaded and lets you retrieve it. An email costs nothing; a found master is worth everything.
• Use the highest-bitrate MP3 you have. A 320 kbps or high-quality VBR encode of a well-mastered track transcodes surprisingly gracefully; a 128 kbps file will get audibly rougher on every platform. The MP3 vs FLAC FAQ explains what is and is not recoverable.
• Convert to WAV for strict uploaders. Where a distributor's validator only takes WAV, run the MP3 through an MP3 to WAV conversion so the upload passes — while understanding this changes the container, not the quality. Never present an upsampled file as a new master.
• Consider remastering from the MP3. For a release that matters commercially, an engineer working from your best MP3 can often deliver something that sits acceptably next to modern releases. It is restoration, not resurrection, but it can be the difference between "unlistenable next to the algorithm's next track" and "fine."

Loudness and metadata, briefly

Two adjacent gotchas surface during distribution. First, streaming platforms loudness-normalize playback, so a master slammed to maximum loudness gains nothing and loses dynamics — mastering to a sensible level matters more than winning a loudness war the platforms already ended. Second, distributors read title and artist from their web forms, not from your file's ID3 tags, but embedded tags still matter for the promo copies you send around; keep them accurate and consistent with the release metadata.

Protecting the next release

The 2014 hard-drive story only has to happen once. For every release from now on:

• Archive the mastered WAV and a FLAC copy in two places — one local, one cloud. FLAC roughly halves the storage cost with zero quality cost.
• Keep the distributor upload exactly as delivered. Not "the mix, probably" — the actual files that went to stores, with release name and date in the folder name.
• Make your MP3 listening copies from the master via a WAV to MP3 conversion at 320 kbps, and treat them as disposable. Masters are for keeping; MP3s are for using.

Distribution format rules feel like bureaucracy, but they exist to protect how your music sounds on the platforms where people will actually hear it. Feed the pipeline lossless whenever you can, be honest about the ceiling when you cannot, and never let the only copy of a master live on one drive again.