I’ve always thought that setting -crf [x] determines the output quality, while (in the same line) -preset [x] affected the output file size.
(Where faster speeds result in bigger output files at the same quality.
On Sat, May 02, 2020 at 14:30:26 +0200, Bouke / VideoToolShed wrote:
> https://trac.ffmpeg.org/wiki/Encode/H.265 <https://trac.ffmpeg.org/wiki/Encode/H.265>
> Under the Constant Rate Factor (CRF):
> -Put differently, if you choose ultrafast, the encoding process is going to run fast, and the file size will be smaller when compared to medium.
> WTF? I would think the output file size will be bigger? (I take it this is just a stupid mistake…)
Yes, that must be a mistake.
> Then, the line continues:
> -The visual quality will not be as good.
> Now it gets confusing. This sorta kinda suggests that -crf and -preset are not to be used together?
I believe that is also wrong, at least from x264 experience. If you
don't explicitly give either a bitrate or a CRF, the mode will be CRF.
They combine just fine: The preset retains the targeted quality (more
or less), but compromises conversion speed for bandwidth/size/bitrate.