AV1 vs H.264: Which Should You Upload?
Your phone now records AV1 by default. YouTube serves it. The file is visibly smaller than the H.264 equivalent. So why would anyone still upload H.264?
Because smaller is only half of the question, and the other half is who can open it.
What AV1 actually buys
AV1 reaches roughly the same visual quality as H.264 at about half the bitrate. That is a real saving and it is why every streaming platform has moved to it.
The published equivalences across the common codecs run roughly:
| Codec | Bitrate for the same quality |
|---|---|
| H.264 | baseline |
| HEVC / VP9 | about 60% |
| AV1 | about 50% |
Those numbers are the industry’s own figures rather than something measured here, and they should be treated as approximate — the real ratio depends heavily on the content, with grain and fast motion closing the gap considerably.
Why a bitrate check judges it unfairly
This is the trap. Most “is my bitrate high enough” guidance is quoted for H.264, because that is what the recommendations were written for. Judge an AV1 file against those numbers and it fails for being efficient.
A real 1080x1920 AV1 clip was banded Poor by exactly that mistake: its bitrate was perfectly adequate for AV1 and well below what the same quality would need in H.264. The fix is to know the codec before reading the bitrate, and an unlisted codec is treated as H.264 — which can only ever under-praise a file, never flatter one.
So if a tool tells you an AV1 export is under-bitrated, check whether the tool knows it is looking at AV1.
The cost nobody mentions
AV1 is expensive to decode, and not every piece of software can do it at all.
Plenty of tools built on older media libraries will open an AV1 file, report its duration and resolution correctly, and then produce no frames. The file is not broken; the decoder is missing. That failure mode is particularly unhelpful because nothing errors — you get an empty result rather than a message.
It also costs more memory. Measured at the same resolution, an AV1 capture used 127MB against H.264’s 39MB — more than three times as much, for the same five seconds of 720p.
What to actually upload
Upload AV1 when the destination is a platform that will re-encode anyway (YouTube, Instagram, TikTok) and your connection is the bottleneck. You save the upload time and they transcode regardless.
Upload H.264 when the file is going to a person rather than a platform: a client review, an email attachment, an editor who will cut with it, a CMS you do not control. H.264 opens everywhere, and “it plays for me” is not evidence that it plays for them.
Keep the master in neither. An intermediate codec — ProRes, DNxHR — is what an archive wants. Both AV1 and H.264 are delivery formats, and re-encoding between delivery formats loses a little every time.
No original to compare against? DiffALL measures one file on its own and names what is wrong with it.
Check a file — free