[FFmpeg-devel] [RFC/PATCH] Add mode to enable disable deinterlacing based on input interlaced flag to yadif

Michael Niedermayer michaelni at gmx.at
Wed Jun 29 05:34:10 CEST 2011


On Wed, Jun 29, 2011 at 12:03:03AM +0200, elupus wrote:
> On Tue, 28 Jun 2011 22:45:07 +0200, Michael Niedermayer wrote:
> 
> > On Tue, Jun 28, 2011 at 09:04:57PM +0200, elupus wrote:
> >> [6 quoted lines suppressed]
> > Content-Description: Attached file: 0001-Add-mode-to-enable-disable-based-on-input-interlaced.patch
> >> [25 quoted lines suppressed]
> > 
> > why dont you just use parity=-1 ?
> 
> That could probably be used I suppose. Seem logical that if there is
> no interlace flag, top_field_first will be invalid anyway. If you
> prefer that feels like an okey change. 
> 
> >> [58 quoted lines suppressed]
> > 
> > This doesnt look entirely correct.
> > consider that every second frame is interlaced. these frames should
> > be deinterlaced using the non interlaced frames as prev/next
> 
> I couldn't decide if that was the best approach, but it probably is.
> The point was to release early to avoid holding on to buffers. But
> it should be enough to just release prev and do a ref copy on
> cur.

what i was thinking about was some interlace detect filter that sets
interlace / no interlace flags a bit unstable over some distance
and i think NULL-ing the references is not a good idea in this case


> 
> >> [26 quoted lines suppressed]
> > 
> > please elaborate on why this change is needed
> > 
> 
> Due to the releas and nulling of cur above. If not
> then this hunk is not needed.

thought so ...



> 
> >> [6 quoted lines suppressed]
> > 
> > please elaborate on why this change is needed
> > 
> > [...]
> 
> Otherwise poll would return 2. But can probably be written like:
> if(yadif->next && !yadif->next->video->interlaced && yadif->parity == -1)

maybe ok, id have to see the complete patch

[...]
-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

Opposition brings concord. Out of discord comes the fairest harmony.
-- Heraclitus
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <http://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20110629/f5ab4b14/attachment.asc>


More information about the ffmpeg-devel mailing list