Since mplayer(2) itself does not detect if a source is interlaced or not I started a small project to analyse a source with mplayers pullup filter.
basically it collects the output of :
and:
which both output stuff like:
per frame.
Depending if there's a + or a ++ between the numbers in affinity it counts light (+) and strong (++) affinities, whether or not there is a break and how the duration changed.
Looking at the affinity, breaks and duration changed it than decides if the content is:
a. normal interlaced
b. telecine
c. field-blended
d. mixed progressive/interlaced
content.
I know that this can't replace detection with your own eyes using (avisynth and separate fields&co), but it might be a nice gimmick for those that don't know how to use avisynth.
Since I normally only have telecine or interlaced content I play with (+ I'm a n00b when it comes to interlacing&co) this decision isn't really tested well.
-> Would be nice if some of you could test it (drag&drop a file into it, than press Analyse) with some of your content and if you think/know that the tool is wrong, please post a small sample of the content which shows the problem and tell me what it should have detected.
Cu Selur
Ps.: I attached a Windows version of MPlayerInterlaceDetection and mplayer2 so people can test. If this works and I get some feedback I plan to put the tool (and it's source code) up on googlecode or something similar. (btw. since it's only dependency is mplayer2, it's also possible to compile Linux&Mac versions)
basically it collects the output of :
Code:
mplayer2 -noframedrop -lavdopts threads=8 -speed 100 -v "path to input" -nosound -vo null -ao null -nosub -vf pullupCode:
mplayer2 -noframedrop -lavdopts threads=8 -speed 100 -v "path to input" -nosound -vo null -ao null -nosub -vf pp=lb,pullupCode:
V: 17.9 537/537 156% 513% 0.0% 0 0 100.00x
affinity: .0++1..2+.3.
breaks: .0..1..2|.3.
duration: 2Depending if there's a + or a ++ between the numbers in affinity it counts light (+) and strong (++) affinities, whether or not there is a break and how the duration changed.
Looking at the affinity, breaks and duration changed it than decides if the content is:
a. normal interlaced
b. telecine
c. field-blended
d. mixed progressive/interlaced
content.
I know that this can't replace detection with your own eyes using (avisynth and separate fields&co), but it might be a nice gimmick for those that don't know how to use avisynth.
Since I normally only have telecine or interlaced content I play with (+ I'm a n00b when it comes to interlacing&co) this decision isn't really tested well.
-> Would be nice if some of you could test it (drag&drop a file into it, than press Analyse) with some of your content and if you think/know that the tool is wrong, please post a small sample of the content which shows the problem and tell me what it should have detected.
Cu Selur
Ps.: I attached a Windows version of MPlayerInterlaceDetection and mplayer2 so people can test. If this works and I get some feedback I plan to put the tool (and it's source code) up on googlecode or something similar. (btw. since it's only dependency is mplayer2, it's also possible to compile Linux&Mac versions)