x264 algorithms?

[zhou_heng]

where do algorithms of x264, such as rate control, motion estimation, mode decision, error concealment, etc., come from? Are they from latest IEEE CSVT paper or VCEG meeting contributions?

[Dark Shikari]

The brains of the people who write it, obviously.

[Esurnir]

The brains of the people who write it, obviously.

Healthy diet ? Do you advise eating a lot of fish ? What would be the most productive programmer food of choice  ?

[zhou_heng]

I'm just curious about the strategy of development. Do you actively adopt latest algorithms efficient in RD performance and/or complexity to improve the project? If so, how do you handle the patent problem -- carefully avoid those algorithms that are patent-protected, or just use them regardless of the patent?

[Dark Shikari]

I'm just curious about the strategy of development. Do you actively adopt latest algorithms efficient in RD performance and/or complexity to improve the project? If so, how do you handle the patent problem -- carefully avoid those algorithms that are patent-protected, or just use them regardless of the patent?

Yes, we adopt the "latest algorithms" that we came up with, which are probably not patented, given that we came up with them.  If you're suggesting we read IEEE papers or similar, I could probably get more useful insights by putting a cat on my keyboard with a text editor open.

[DarkZell666]

Yes, we adopt the "latest algorithms" that we came up with, which are probably not patented, given that we came up with them.  If you're suggesting we read IEEE papers or similar, I could probably get more useful insights by putting a cat on my keyboard with a text editor open.

+1

[Quarkboy]

Yes, we adopt the "latest algorithms" that we came up with, which are probably not patented, given that we came up with them.  If you're suggesting we read IEEE papers or similar, I could probably get more useful insights by putting a cat on my keyboard with a text editor open.

Considering what you say, do you think it a worthy endeavor for someone to go through and break down the algorithms used within x264 into a scientific paper to be published in one of the journals or simply documented online in some form other than the code itself?

[pengvado]

Sure. Dark Shikari never seems to get tired of tutoring potential developers, but a written documentation would probably reach more people than actually show up on IRC.
Submitting such to a journal would of course increase the average quality of the journal; whether you care about that is a separate matter from whether you want x264 to be documented.

[Quarkboy]

Sure. Dark Shikari never seems to get tired of tutoring potential developers, but a written documentation would probably reach more people than actually show up on IRC.
Submitting such to a journal would of course increase the average quality of the journal; whether you care about that is a separate matter from whether you want x264 to be documented.

Hell I'd just post it to the arxiv.  Not like I have time to do such a thing, but part of my brain is pining for the days I was reading string theory papers and this is the next most interesting thing to me