Jean-Baptiste Kempf

And if you were really lucky, you would find, like, if you were unlucky, you'd get nothing or you'd get one or two.

And then sometimes you'd find a gold mine.

It's like, yeah, my company has 100,000 of these files because we depended on it for some reason.

And so those are kind of the best, because then they can test bit exactness across the huge range of coding tools

Bit exactness.

So most but not all video codecs, certainly from about the 2000s onwards, have a bit exact definition.

So every implementation must produce exactly the same bits, bit for bits, in exactly the same data that comes out of a decoder.

for a given sample.

So Lex's implementation, JB's implementation, and my implementation of H.264 must match bit exactly.

That wasn't the case in the 90s of MPEG-2, probably fair to say one of the biggest mistakes the video industry made.

And I think people who were in the room in 92, most of both of us were in diapers, I suspect, but have acknowledged, I would give a shout out to Yuri Reznik.

He's acknowledged that was one of the big mistakes of the era.

I would point out, actually, FFmpeg is unique in the sense that it's...

It has been a winner-takes-all scenario.

Browsers is a good analogy because it has to parse a lot of different content and render it in a particular way, like a decoder.

But there still are multiple browser engines.

There's Firefox's one, there's Chrome's one, there's a few Japanese ones that are pretty decent.

That's not been the case in multimedia in general across a wide range of codecs.

FFmpeg is kind of...

one at all, I suppose, in a sense, because of the fact that you can get every new codec added is actually worth more than the value of that codec itself, because it makes the whole thing better.