What Is AV1, and Is It Worth Using Yet?

AV1 is the newest of the widely deployed video codecs, finalised in 2018 by a group including most of the companies that actually run large video services. It compresses better than H.265, which compresses better than H.264. It is also royalty-free, which turns out to matter more than the technical comparison.

Why it exists

Not primarily because H.265 was not good enough. H.265 is very good. The problem was its licensing: multiple groups held patents, they did not agree on a single set of terms, and the resulting uncertainty made large distributors reluctant to commit to it.

AV1 was built by an alliance of those distributors and the browser and chip makers, specifically to be free to use. That origin explains its adoption pattern: it appeared on the largest streaming services first, because they were the ones paying the bill that motivated it.

How much smaller

Roughly 30 per cent below H.265 and around half of H.264 for comparable quality, though the range is wide. The gains are largest at high resolutions and on complex footage, and modest at low resolutions where there is less redundancy to exploit.

Treat any single figure with suspicion, including that one. Codec comparisons depend heavily on the encoder used, the settings, and the content. A number quoted without those three is close to meaningless.

The cost: encoding time

This is the real barrier. AV1 achieves its efficiency by searching a far larger space of ways to describe each frame, and that search is expensive. Early encoders were slow enough to be impractical. Modern ones are much faster, and hardware encoders in recent graphics cards are faster again, but software AV1 encoding remains substantially slower than H.264 at comparable quality.

For a streaming service encoding once and serving millions of times, that trade is obviously worth it. For someone exporting a video to send to one person this afternoon, it obviously is not.

Where support stands

• Browsers: broadly good. Decoding is widely available on current desktop browsers.
• Phones: improving. Recent flagship chips include hardware decoding; older and cheaper devices do not, and software decoding drains battery visibly.
• Televisions and streaming boxes: mixed, and heavily dependent on age.
• Editing software: decoding is common, encoding less so, and hardware-accelerated encoding needs a recent graphics card.

The pattern is the same one H.265 went through, and H.264 before it. New codecs arrive on new hardware and take years to become safe defaults.

When to use it now

• You are serving video from your own website and can provide an H.264 fallback for anything that cannot play it.
• Storage matters more than encoding time, such as archiving a large library you control.
• You are encoding once and it will be watched many times, which is exactly the case AV1 was designed for.

When not to

• You are handing a file to someone whose device you do not know. H.264 in an MP4 remains the answer that always works.
• You are uploading to a platform that will re-encode everything anyway, which discards your saving entirely.
• You need the export finished quickly.
• The recipient is on an older phone, where software decoding will drain their battery and stutter.

A sensible position

AV1 is the direction video is going, and the licensing question that held H.265 back does not apply to it. That makes it likely to become the default eventually in a way H.265 never quite did.

Eventually is not today for most people. The practical policy is unchanged: keep archives in whatever is most efficient for your own storage, and deliver in H.264 unless you know what will be playing it. A file that will not open is worth nothing however small it is.