Now I'm trying to get rid of the last small artifacts. It seems that mcdeint
helped to reduce the ones I was talking about in my first message. But
mcdeint is designed to work with deinterlacers filters, am I wrong? Do you
know a proper way to fix it?
On Sat, May 10, 2008 at 8:39 PM, Justin Randall <jrrandall@gmail.com> wrote:
> Hi Benoit,
>
> Looking back, your original post with the switching back and forth
> between "30000/1001 fps NTSC" and "24000/1001 fps" does suggest a
> telecline pattern. Sorry I wasn't much help there.
>
> On Sat, May 10, 2008 at 1:33 PM, Benoit <kvasarnov@gmail.com> wrote:
> > Hi,
> >
> > Sorry for answering this late but I had some exams lately.
> >
> > I searched a little more about interlaced and telecined movies. From what
> I
> > read, animes are usually not interlaced but telecined. Plus, I got better
> > results using the filmdint filter associated with mcdeint (I suppose
> mcdeint
> > is made to work with deinterlacers filters but the fact is that I had a
> > better result...). I still have some small artifacts with high motions
> > scenes but they are not visible while playing.
> >
> >
> > On Sun, May 4, 2008 at 11:31 PM, Justin Randall <jrrandall@gmail.com>
> wrote:
> >
> >> You could also try it without the mcdeint (motion componsation
> >> de-interlacer) first if you want, to see if that gives you good
> >> results.
> >>
> >> On Sun, May 4, 2008 at 5:16 PM, Justin Randall <jrrandall@gmail.com>
> >> wrote:
> >> > Ok the best way to convert NTSC pure interlaced 30 fps to progressive
> >> > is to double the frame rate to 60 fps. Otherwise you loose a lot of
> >> > information. I would suggest trying something like this. Be careful
> >> > of frame order in the source. MPEG2 can be bootom frame first or top
> >> > frame first. (Disclaimer: Depending on your CPU, this will take a
> >> > LONG time to process, but that is the price for converting interlaced
> >> > content properly). The reason I only do one pass using constant
> >> > quantizer for the encode is because the de-interlacing takes so long.
> >> > Good luck and again let me know how it comes out.
> >> > -Justin
> >> >
> >> > For top frame first (TFF) MPEG2 content:
> >> >
> >> > >> mencoder dvd://1 -nosound -vf
> >> > yadif=3:1,mcdeint=2:1:10,scale=640:480 -sws 9 -ofps 60000/1001 -ovc
> >> >
> >> > x264 -x264encopts
> >> >
> >>
> threads=auto:qp=17:qpmax=28:subq=6:partitions=all:8x8dct:me=umh:subme=6:frameref=5:bframes=3:b_pyramid:b_rdo:weight_b:keyint=250
> >> > -o Output.avi
> >> >
> >> > For bottom frame first (BFF) MPEG2 content (notice the 0's for the
> >> > second options of yadif and mcdeint):
> >> >
> >> > >> mencoder dvd://1 -nosound -vf
> >> > yadif=3:0,mcdeint=2:0:10,scale=640:480 -sws 9 -ofps 60000/1001 -ovc
> >> >
> >> > x264 -x264encopts
> >> >
> >>
> threads=auto:qp=17:qpmax=28:subq=6:partitions=all:8x8dct:me=umh:subme=6:frameref=5:bframes=3:b_pyramid:b_rdo:weight_b:keyint=250
> >> > -o Output.avi
> >> >
> >> >
> >> >
> >> >
> >> > On Sun, May 4, 2008 at 4:39 PM, Benoit <kvasarnov@gmail.com> wrote:
> >> > > Hi,
> >> > >
> >> > > You were right, it seems like it is an interlaced video but yadif
> did
> >> not
> >> > > work well (more artifacts than before. I used the pp=lb filter
> >> instead.
> >> > >
> >> > > But I get some weird frames:
> >> > >
http://img509.imageshack.us/img509/9369/screenshot3yj0.png
> >> > >
> >> > >
> >> > > As you can see there's no more artifact but it seems like two
> frames
> >> are
> >> > > merged together.
> >> > >
> >> > > Do you know how to fix it?
> >> > >
> >> > >
> >> > >
> >> > >
> >> > >
> >> > > On Sat, May 3, 2008 at 1:16 AM, Justin Randall <
> jrrandall@gmail.com>
> >> wrote:
> >> > >
> >> > > > Hi there,
> >> > > >
> >> > > > If your source is truely 100% teleclined, then this should work
> >> well for
> >> > > > you..
> >> > > >
> >> > > > >> mencoder dvd://1 -nosound -vf
> >> > > > pullup,hqdn3d=3:2:3:3,softskip,harddup -sws 9 -ofps 24000/1001
> -ovc
> >> > > > x264 -x264encopts
> >> > > >
> >> > > >
> >>
> threads=auto:qp=17:qpmax=28:subq=6:partitions=all:8x8dct:me=umh:subme=6:frameref=5:bframes=3:b_pyramid:b_rdo:weight_b:keyint=250
> >> > > > -o Output.avi
> >> > > >
> >> > > > Otherwise you're dealing with interlaced and this will probably
> >> work
> >> > > > better (WARNING: should take a LONG time to encode but the
> results
> >> > > > should be good).
> >> > > > >> mencoder dvd://1 -nosound -vf yadif,hqdn3d=3:2:3:3,harddup
> -sws
> >> 9
> >> > > > -ovc x264 -x264encopts
> >> > > >
> >> > > >
> >>
> threads=auto:qp=17:qpmax=28:subq=6:partitions=all:8x8dct:me=umh:subme=6:frameref=5:bframes=3:b_pyramid:b_rdo:weight_b:keyint=250
> >> > > > -o Output.avi
> >> > > >
> >> > > > Good luck and please let me know how these work for you.
> >> > > >
> >> > > > Thanks,
> >> > > > Justin
> >> > > >
> >> > > > On Fri, May 2, 2008 at 4:48 PM, Benoit <kvasarnov@gmail.com>
> >> wrote:
> >> > > > > Hello,
> >> > > > >
> >> > > > > I bought some DVDs of my favorite anime but I experience
> >> problems
> >> > > > encoding
> >> > > > > them.
> >> > > > >
> >> > > > > It is an NTSC Mixed progressive and telecine video (mplayer
> >> switches
> >> > > > back
> >> > > > > and forth between "30000/1001 fps NTSC" and "24000/1001
> fps").
> >> > > > > I used the 'pullup' filter associated with 'softskip', which
> >> produce
> >> > > > these
> >> > > > > results with a few frames:
> >> > > > >
http://img136.imageshack.us/img136/1602/mencoderkg0.jpg
> >> > > > >
> >> > > > > Here is the telecined frame that causes problems:
> >> > > > >
http://img178.imageshack.us/img178/4061/screenshot1bf7.png
> >> > > > >
> >> > > > > I downloaded the same episode from a torrent to compare the
> >> quality.
> >> > > > The
> >> > > > > encoder did a great job and got this result:
> >> > > > >
http://img353.imageshack.us/img353/3430/mencoder2lx7.jpg
> >> > > > >
> >> > > > > I don't think this encoder used the same source as mine, but
> I'd
> >> like
> >> > > > to get
> >> > > > > some advice from experienced encoders to help me fix my
> problem
> >> and get
> >> > > > a
> >> > > > > similar quality.
> >> > > > >
> >> > > > > The command line I used:
> >> > > > > mencoder dvd://1 -nosound -ovc x264 -x264encopts
> >> > > > >
> >> > > >
> >>
> bitrate=1200:subq=4:partitions=all:me=umh:frameref=3:bframes=2:trellis=2:b_pyramid:weight_b:threads=auto
> >> > > > > -ofps 24000/1001 -vf pullup,softskip,scale=640:480 -fps 29.97
> -o
> >> > > > Output.avi
> >> > > > >
> >> > > > >
> >> > > > > Thanks a lot!
> >> > > > >
> >> > > > >
> >> > > > > Regards
> >> > > > > _______________________________________________
> >> > > > > MEncoder-users mailing list
> >> > > > > MEncoder-users@mplayerhq.hu
> >> > > > >
https://lists.mplayerhq.hu/mailman/listinfo/mencoder-users
> >> > > > >
> >> > > >
> >> > > >
> >> > > >
> >> > > > --
> >> > > > Justin Randall
> >> > > > 617.733.4107
> >> > > > _______________________________________________
> >> > > > MEncoder-users mailing list
> >> > > > MEncoder-users@mplayerhq.hu
> >> > > > https://lists.mplayerhq.hu/mailman/listinfo/mencoder-users
> >> > > >
> >> > > _______________________________________________
> >> > > MEncoder-users mailing list
> >> > > MEncoder-users@mplayerhq.hu
> >> > >
https://lists.mplayerhq.hu/mailman/listinfo/mencoder-users
> >> > >
> >> >
> >> >
> >> >
> >> > --
> >> >
> >> >
> >> > Justin Randall
> >> > 617.733.4107
> >> >
> >>
> >>
> >>
> >> --
> >> Justin Randall
> >> 617.733.4107
> >> _______________________________________________
> >> MEncoder-users mailing list
> >> MEncoder-users@mplayerhq.hu
> >> https://lists.mplayerhq.hu/mailman/listinfo/mencoder-users
> >>
> > _______________________________________________
> > MEncoder-users mailing list
> > MEncoder-users@mplayerhq.hu
> > https://lists.mplayerhq.hu/mailman/listinfo/mencoder-users
> >
>
>
>
> --
> Justin Randall
> 617.733.4107
> _______________________________________________
> MEncoder-users mailing list
> MEncoder-users@mplayerhq.hu
> https://lists.mplayerhq.hu/mailman/listinfo/mencoder-users
>
_______________________________________________
MEncoder-users mailing list
MEncoder-users@mplayerhq.hu
https://lists.mplayerhq.hu/mailman/listinfo/mencoder-users
No comments:
Post a Comment