[FFmpeg-devel] [PATCH] avfilter/tinterlace: merge code with interlace

Clément Bœsch u at pkh.me
Thu Dec 4 13:46:25 CET 2014


On Thu, Dec 04, 2014 at 12:41:47PM +0000, tim nicholson wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> On 04/12/14 09:59, Clément Bœsch wrote:
> > On Thu, Dec 04, 2014 at 09:31:47AM +0000, tim nicholson wrote:
> >> On 02/12/14 18:43, Clément Bœsch wrote:
> >>> ---
> >>> After this commit, interlace doesn't behave the same if the input frames
> >>> are already interlaced. Does anyone wants me to keep this behaviour?
> >>>
> >>> Also, the PTS adjustments might be different, but I didn't follow the
> >>> current discussions about it so I can't tell what people want.
> >>> ---
> >>>  libavfilter/Makefile                |   2 +-
> >>>  libavfilter/interlace.h             |  58 ---------
> >>>  libavfilter/tinterlace.h            |   5 +-
> >>>  libavfilter/vf_interlace.c          | 250 ------------------------------------
> >>>  libavfilter/vf_tinterlace.c         |  44 +++++++
> >>>  libavfilter/x86/Makefile            |   2 +-
> >>>  libavfilter/x86/vf_interlace_init.c |  47 -------
> >>>  7 files changed, 49 insertions(+), 359 deletions(-)
> >>>  delete mode 100644 libavfilter/interlace.h
> >>>  delete mode 100644 libavfilter/vf_interlace.c
> >>>  delete mode 100644 libavfilter/x86/vf_interlace_init.c
> >>>
> >>> [..]
> >>
> >> The docs will need updating to remove the vf_interlace entry and add the
> >> lp filter options. Maybe that is the time to merge the ascii art
> >> graphics and add Michael's additional mode ones as well?
> >>
> > 
> > None of the filter disappear and their options don't change. It's just
> > code refactoring.
> 
> Sorry misread it only looking at the diff. I thought there was a
> funtionality merger as well as per the irc discussion.
> 

There are 2 functional changes, which is why I haven't pushed yet:
 - the "do nothing with already interlaced frame" behaviour from
   vf_interlace is not kept
 - the PTS adjustments might not be the same so I might need to set the
   exact_tb flag internally if it's vf_interlace to copy the behaviour
   (but I need confirmation from the people who are using this filter)

[...]

-- 
Clément B.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 473 bytes
Desc: not available
URL: <https://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20141204/85ca5097/attachment.asc>


More information about the ffmpeg-devel mailing list