What is x264 --filter option, and min bitrate?

[73ChargerFan]

Two questions -

What is the --filter option? I can't find in at mewiki, and it is impossible to search for.

Second, is there an option to specify a minimum bitrate?

I'm encoding cartoons recorded off tv, std def, at --CRF 20. Most episodes come in at 850-950 but one came in at 650 (source wasn't the best quality.) I like the simplicity of CRF but want a minimum bitrate of 800. I don't want to use 2-pass.

Thanks.

[LoRd_MuldeR]

1. The "--filter" option controls the H.264 "In-Loop Deblocking" filter. The default value is "0:0". Negative values decrease the filter strength, which preserves more sharpness, but increases the danger of "blocking" artifacts at the same time. The other way around positive values increase the filter strength, which better prevents "blocking" artifacts, but kills more details/sharpness at the same time. Usually you shouldn't need to adjust the "--filter" option directly. Instead use "--tune film" or "--tune animation", whatever suits your type of source (for cartoon "--tune animation" would be the option of choice).

2. Why do you want to specify a minimum bitrate? CRF mode (and 2-Pass mode as well) tries to keep the perceived quality constant throughout the entire video and it will assign each scene the bitrate it requires to retain the desired quality level. If the resulting quality for a certain CRF value doesn't satisfy you, then simply use a lower CRF value until you are satisfied with the result. Otherwise keep it as-is.

[Audionut]

All other things being equal, at the same crf, the smaller encode was easier to encode. Hence why it came out smaller.

[73ChargerFan]

1. The "--filter" option controls the H.264 "In-Loop Deblocking" filter.
... Instead use "--tune film" or "--tune anime", whatever suits your type of source (for cartoon "--tune anime" would be the option of choice).

Thank you! I'm using "--tune animation" with "--filter 1,1".

The reason for the filter option is because it is specified in the scripts created by RipBot264, and in all the presets I found on the internet. Everyone else used it, so I figured I should also until I know otherwise. I'm learning.

I'll remove the "--filter 1,1" now.


2. Why do you want to specify a minimum bitrate?

Most of my encodes come out 180MB, then 2 out of 15 are 120MB, which I just don't trust. Placebo effect? :rolleyes:

So, I'll just re-encode those with a lower CRF manually.
:thanks:

Edit:
If I use 2 pass to specify a bitrate, does x264 use the first pass to calculate a CRF, which is then used on the second pass? I thought CRF would give more bits to more complex scenes, fewer bits to easy scenes, whereas 2 pass is constant bitrate, not adaptive. If it is adaptive, then I'll start using 2 pass.

[73ChargerFan]

All other things being equal, at the same crf, the smaller encode was easier to encode. Hence why it came out smaller.
Do you mean like a cartoon showing outer space - 80% of the screen is black, so it compresses that much better.
Ah...

[J_Darnley]

Edit:
If I use 2 pass to specify a bitrate, does x264 use the first pass to calculate a CRF, which is then used on the second pass? I thought CRF would give more bits to more complex scenes, fewer bits to easy scenes, whereas 2 pass is constant bitrate, not adaptive. If it is adaptive, then I'll start using 2 pass.
Yes. 2-pass and crf, at the same bitrate, are the same quality. 2-pass is most certainly not constant bitrate. You would have to be very silly to tell x264 to do that.

Do you mean like a cartoon showing outer space - 80% of the screen is black, so it compresses that much better.
Yes, something like this. Blackness is easy to compress but so is any flat colour. Still scenes are also easy.