[FFmpeg-user] formula for transcoding to ogv, shrinking an mkv

James Miller gajs-f0el at dea.spamcon.org
Fri Jan 18 01:16:41 CET 2013

On Thur, 17 Jan 2013, Carl wrote:

> If you have a target filesize, don't use a constant
> quantiser, but specify a bit-rate (calculated by you)
> and use 2-pass encoding.

Thank you for your response, which I am trying to interpret. According to 
some research I've done in that effort, the -crf option in my formula is 
probably what you're referring to as a "constant quantizer"; is that 
correct? I've also gathered that, instead of using this, I should use the 
-b option followed by some numerals in order to get a certain bit-rate. 
Correct? Can you recommend any basic documentation on 2-pass encoding? I 
don't know that I've ever tried that before.

> Please note that for all questions on this mailing
> list, you should add complete, uncut console output
> together with your command line.

Hmmm. I'm a little puzzled by this stipulation. When I run my command in 
the console, I get several lines that give specifics about the 
characteristics of the file being processed and about what ffmpeg is 
preparing to perform on it. That, on successful attempts (like the one I 
posted about), is followed by a sort of progress meter that runs until the 
transcoding is finished. Are you saying I should copy and paste all that 
into any query I make to this list? Not that I have any aversion to it; 
rather, I just want to make sure I understand the stipulation.

> (And of course: Please do not forget that the
> theora encoding quality is very limited, meaning
> it may not be easy to get the target filesize
> and acceptable quality.)

I have not found it easy. I did come up with an incantation that produced 
a file very close to the size I needed and that was of acceptable quality, 
although, judging by your comments, it seems not a very elegant, or 
perhaps efficient, solution. So now I will experiment further to see if 
can improve on it.


More information about the ffmpeg-user mailing list