Ghosting in a NTSC Interlaced source

[Macarse]

Hi trying to make an XviD of a R4 DVD.
Here is a sample:
http://rapidshare.com/files/19592121/sample.vob.html

In DGIndex:
Aspect Ratio: 4:3
Frame rate: 29.970
Video Type: NTSC
Frame Type: Interlaced

I am following the DecombTutorial.html guide to do the IVTC.
I reached this values for the Decomb.dll plugin:


AssumeBFF()
Telecide(guide=0,post=2,vthresh=60)
Decimate()


I used guide=0 because I think it doesn't follow any pattern.
The high value in vthresh was choosen because it reduced significatelly the amount of ghosting, but I am not very pleased with the results.

Would be right to use TomsMoComp with Decomb?
Should I use IVTC in this source or just Deinterlace?

Thanks for reading!

[manono]

Hi-

It's PAL2NTSC and no IVTC will work. One way to attack it is to use something like this to remove the blends and make it progressive again:

LeakKernelBob(Order=0)#the sample was BFF, adjust if the whole clip is TFF
RePAL()

[Macarse]

Thanks for the info. I've been reading about PAL2NTSC.
I didn't try anything yet but I have some doubts.

If my source is NTSC (althought it's PAL=>NTSC) shouldn't I use:

telecide(guide=3)
decimate(6)


In the forum it's said that after using repal I would get 24.975fps
using AssumeFPS(25) would ruin sync with audio?

And about the comment:

#the sample was BFF, adjust if the whole clip is TFF


If a sample is BFF, shouldn't the whole clip be BFF? Speaking of not hybrid sources.

Thanks for your time manono.

[manono]

Hi-

In the forum it's said that after using repal I would get 24.975fps

That's correct, you will. So will Telecide().Decimate(6) give you 24.975. But an IVTC will also give you a ton of blended frames and jerky playback.

using AssumeFPS(25) would ruin sync with audio?

Yes, unless you adjust the audio. I suggest keeping it at 24.975fps, unless you have some reason to change it. And if I were to change it, I'd change it to 23.976fps to get rid of the higher pitched (usually) speeded up PAL audio.

If a sample is BFF, shouldn't the whole clip be BFF? Speaking of not hybrid sources.

Not necessarily. It partly depends on how you cut it. It's easy to switch the field order around. Most, but by no means all, DVDs are TFF.

[Macarse]

I just did some tests:
Tried both,
1)
Telecide(guide=3)
Decimate(6)

2)
LeakKernelBob(order=0)
RePAL()

RePAL() gives me better results. Much less ghosting.


I will post screenshots when I do the whole clip.

[Macarse]

Just finished the encoding process:
Here are the results.
The full clip was encoded to fit in a cd.

Both versions have this attributes:
Lenght: 01:15:32 (113,187 fr)
Resolution: 672x368 (1.83:1) [=42:23]
Bitrate: 1150 kb/s
FPS: 24.975
Qf: 0.186 bits/pixel
---------------------------
Screenshots:
First showing Repal, and afterwards IVTC.
---------------------------
http://img179.imageshack.us/img179/8185/screen1repalyb0.th.png (http://img179.imageshack.us/my.php?image=screen1repalyb0.png)

http://img180.imageshack.us/img180/6057/screen1ivtcla1.th.png (http://img180.imageshack.us/my.php?image=screen1ivtcla1.png)

As you can see, IVTC version has a lot of ghosting compared to repal.
---------------------------
http://img227.imageshack.us/img227/7440/screen2repalzl5.th.png (http://img227.imageshack.us/my.php?image=screen2repalzl5.png)

http://img220.imageshack.us/img220/2006/screen2ivtclw8.th.png (http://img220.imageshack.us/my.php?image=screen2ivtclw8.png)

Changing scenes...
---------------------------
http://img401.imageshack.us/img401/372/screen3repalbt8.th.png (http://img401.imageshack.us/my.php?image=screen3repalbt8.png)

http://img220.imageshack.us/img220/667/screen3ivtcjk3.th.png (http://img220.imageshack.us/my.php?image=screen3ivtcjk3.png)

Once again, IVTC version has lots of ghosting.

------------------------------------------------------
I'd conclude Repal it's the best way to treat this kind of source.
Thanks for your help manono.