X264 settings explanations.

[kakomu]

I've been using X264 for a few months. However, I'm still dumbfounded by all of the settings and many of the explanations that MeGUI gives are rather vague or downright poor. Is there an explanation of what all of the different settings mean that goes in depth?

[foxyshadis]

There's a lot of threads where they're hashed out in here, but you might check DeathTheSheep's guide.

[kakomu]

There's a lot of threads where they're hashed out in here, but you might check DeathTheSheep's guide.
I found his guide, but it looks to be based on the VFW ecnoder. From what I've read, the VFW and GUI/commandline encoders are rather different. Is there something a little more...complete?

[check]

yep, they are rather different, but the settings it does cover it covers reasonably well. The mencoder readme also has good information on x264, although it's similarly incomplete.

[bond]

:readguid: :readfaq: :readrule:

[bob0r]

http://x264.nl/jab.gif

[unskinnyboy]

http://x264.nl/jab.gif
haha, I must save this smiley somewhere..:D

[lazyn00b]

Not so fast with those "Read the FAQ" smilies! I'd say it's kind of a mixed bag. Certainly, all the x264 settings that involve tradeoffs between speed and quality have been extensively covered by discussions and guides here at Doom9. The Mencoder documentation is also extremely useful for good descriptions of many of these settings. Even the MEGui tooltips have been helpful.

However, there are a handful of settings that remain somewhat mysterious. Although it's clear enough what the IP and BP ratios do, it's never really been explained what the implications are of actually changing them from their defaults. Why do some people change IP ratio from 1.4 to 1.1 or 1.25? I don't know - I asked once but never got an answer. Also, the VBV settings are kind of a mystery. It's not even clear to me how the VBV buffer size is measured: is it bytes, kilobytes, or what? To tell you the truth, it's not even totally clear to me when the VBV settings are respected by the encoder, or whether they even work at all. I've also seen people change Quantizer Compression from 0.6 to something else, but I've never seen a good explanation of why you would want to do that.

EDIT: Oh, and one more thing: I've never seen a good explanation why the default minimum GOP size isn't simply "1", like a lot of other codecs.

[DeathTheSheep]

Boy how I wish that every "lazy n00b" would ask questions like that... :D

Some of these settings are changed just 'cuz people want to make others say "Wow, look how advanced this guy is! Yaaay!"
And other times the reasons arise from source-specific problems or personal preference. For instance, I've changed the IP ratio before a few times when I had movies (Pxl comes to mind, j/k :D) which loved changing scenes so much my head hurt...but each time the scene changed, there would be too much random motion and artifacting for the beauty of a 1.4'd I frame to even register.... Man I had some problems with Elfen Lied at Q34 :D
just li'l problems like the keyframes looked great but everything else sucked. Psh. Only happened once, and in a re-encode.

PB ratio is a similar story... I feel inclined to give this rather exclusive example: B-frames severely decrease quality at insanely low bitrates. Why? Because people who use a 60% b-frame reduction don't realize their B-frames will look like sludge way down there.

edit: Wow, that's the last time I make a post so late at night. Don't read it if you know what's good for ya... :D