> I know that this is going to be a strange question ;)
(I would expect no less.)
> The first thought I had was to simply take lines 0, 2, 4, ... 720
> from the first frame and call that field 1, then take the corresponding
> lines (0,2,4,...) from the next frame and call that field 2 and call
> y4m_write_fields() with the two field buffers (and of course set the
> rate code to "-F 4" (30000:1001)). Is this correct, or should I use
> lines 1,3, 5 ... from the 2nd frame? Or will it make much difference?
You want to take the 'other' lines from the 2nd frame....
If you go (0,2,4,...) followed by (1,3,5,...), you will create a
Top-Field-First stream. Vice-versa, it's Bottom-Field-First.
Be careful: this doesn't work on the chroma channels of 4:2:0 streams.
One line of a chroma channel of non-interlaced 4:2:0 video corresponds
to a pair of (un-subsampled) lines from different fields. You need to
get rid of the vertical subsampling first before splitting each frame
up into fields (e.g. convert to 4:2:2).
(And, simply "unzipping" each frame into two fields may induce some kind
of aliasing. Technically, it's a 2:1 subsampling operation, which should
be done with a real filter. But it might not have a big effect; I dunno.
Doing it the real way would take care of the 4:2:0 problem, though.)
-matt m.
ps: Hmm, I should add something like this to y4mscaler; right now it
just bails when facing such a task.
pps: I should also check-in the "YUV4MPEG2, version 3" code I wrote
two/three weeks ago; hopefully by Sunday.
-------------------------------------------------------
This SF.Net email is sponsored by: IBM Linux Tutorials
Free Linux tutorial presented by Daniel Robbins, President and CEO of
GenToo technologies. Learn everything from fundamentals to system
administration.http://ads.osdn.com/?ad_id=1470&alloc_id=3638&op=click
_______________________________________________
Mjpeg-users mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/mjpeg-users
