[FFmpeg-user] ffmpeg architecture question

Mark Filipak markfilipak.windows+ffmpeg at gmail.com
Sun Apr 19 05:15:35 EEST 2020


Oops. "(n+1)%5=3" should have been "(n+1)%5==3".

"On 04/18/2020 10:02 PM, Carl Eugen Hoyos wrote:
> Am So., 19. Apr. 2020 um 03:43 Uhr schrieb Mark Filipak
> <markfilipak.windows+ffmpeg at gmail.com>:
> 
>> My experience is that regarding "decombing" frames 2 7 12 17 ...,
>> 'pp=linblenddeint' (whatever it is) does a better job than 'yadif'.
> 
> (Funny that while I always strongly disagreed some people also
> said this when yadif was new - this doesn't make it more "true"
> in any useful sense though.)

I am splitting out solely frames (n+1)%5=3 and applying "deinterlace" solely to them as single 
frames. For that application, 'pp=linblenddeint' appears to do a better job (visually) than does 
'yadif'.

>> "lb/linblenddeint
>> "Linear blend deinterlacing filter that deinterlaces the given block by
>> filtering all lines with a (1 2 1) filter."
>>
>> I don't know what a "(1 2 1) filter" is -- I don't know to what "1 2 1" refers.
> 
> To the fact that no other filter uses as little information to deinterlace.

To me, deinterlace just means weaving the odd & even lines. To me, a frame that is already woven 
doesn't need deinterlacing. I know that the deinterlace filters do additional processing, but none 
of them go into sufficient detail for me to know, in advance, what they do.


More information about the ffmpeg-user mailing list