Need help deinterlacing a field-blended Bollywood DVD

[Mr.India]

Hi,

I am trying to rip a Indian DVD and the deinterlacing is messy and megui cant solve it

can someone suggest me on what to do.

Sample of SOURCE: http://www.sendspace.com/file/9uyprn

Thanks,

[n0an]

I remember seeing a similar source a few months ago. Was a Bollywood movie as well. Manono would suggest you to get it back to 24.975 or 25p.

TFM().TDecimate(Mode=0,Cycle=6,CycleR=1)

You could even throw in Vinverse() to get rid of residual interlacing. Nothing more can be done to it. As Didée says, garbage in = garbage out :rolleyes:. You could read that thread here (http://forum.doom9.org/showthread.php?t=152834&page=2).

[EuropeanMan]

love didee's quotes

garbage in = garbage out

thank god i recycle

[manono]

It's just standard PAL2NTSC field-blended garbage:

Yadif(Mode=1)#or your favorite bobber
SRestore()
...the deinterlacing is messy and megui cant solve it

MeGUI can't solve anything that isn't simple. Learn how to figure these things out for yourself. Your own eyes are your best tools.

[n0an]

Timecode: 4.605

Yadif(Mode=1).SRestore()
http://i44.tinypic.com/2a5lths.jpg

There is noticeable field overlapping with Yadif.

- edited out IVTC comment.

[manono]

Don't be absurd. You don't use an IVTC on field-blended material. If you want a better looking unblended result, use a better quality bobber.

[n0an]

I was actually bringing up the topic of tdecimate(hybird=3) in that post, but for some reason I missed it :p. I read our earlier discussion on such a source and apparently, you had asked me to use a simple Yadif().Repal() or Srestore().

Anyway, I feel TempGaussMC_beta1mod().SRestore() works best, but is really slow for me. What else could one use for a "better" unblended output?

[Mr.India]

Thanks guys i will try them both on other frames and see Thanks Again,

[EuropeanMan]

Solution goes to 23.976

NOT 25/24.975

with OF COURSE a very good bobber

[n0an]

Solution goes to 23.976

NOT 25/24.975

with OF COURSE a very good bobber

And what would that be? TempGaussMC is really good imo, but too slow for my processor :eek:.

[Vitaliy Gorbatenko]

Humble yourself! And you will have happiness! I have a filtration generally is ~ 1FPS and encoding in x264 (mode Placebo) 1.12-1.35 FPS (and then more and 2 pass!)
In total, a series of 24 minutes - almost two days.

[EuropeanMan]

And what would that be? TempGaussMC is really good imo, but too slow for my processor :eek:.

my point was that this source goes to 23.976...nothing else; although i'm still trying for 24fps :)

[n0an]

lol @ both :)

[manono]

my point was that this source goes to 23.976...nothing else
That's plainly incorrect. It's 24.975/25fps. I'm assuming we're talking about the Sample.vob that Mr.India uploaded earlier

[EuropeanMan]

^ sorry buddy...it goes to 23.976. i've done work on this movie as well...and i have NO ghosting/blends, dropped frames, etc etc...

[manono]

I can prove it to you. Use some variation of this script:

Yadif(Mode=1)#or your favorite bobber
SRestore(Frate=29.97)

and then look for the pattern of duplicate frames. Are they every 5th frame or every 6th frame? Again, I'm basing this on the uploaded sample. If you're using a different DVD version, then anything might happen.

[n0an]

I can prove it to you. Use some variation of this script:

Yadif(Mode=1)#or your favorite bobber
SRestore(Frate=29.97)

and then look for the pattern of duplicate frames. Are they every 5th frame or every 6th frame? Again, I'm basing this on the uploaded sample. If you're using a different DVD version, then anything might happen.

Does the "piiiii" format imply its a PAL2NTSC conversion? Where 'p' stands for progressive and 'i' is interlaced. IIRC, thats what we talked about in another thread. When I try Yadifmod() on the sample, it says "args 2 and 4 must be clips!". What does that mean?

Also, when you say 5th or 6th are the duplicate frames, could it be implied that those duplicate frames are there for the PAL speedup to 25.000? I am not aware if a film can be shot at 25p besides the typical 24p.

[manono]

Does the "piiiii" format imply its a PAL2NTSC conversion?
Yes, but doing as I suggested is more guaranteed, because then you get all progressive and 000001 where 0 is a unique frame and 1 is a duplicate frame. Same thing, really, as what you discovered on your own.
Also, when you say 5th or 6th are the duplicate frames, could it be implied that those duplicate frames are there for the PAL speedup to 25.000?
An already speeded up PAL source was converted to NTSC by field-blending, After being unblended, that dupe frame is what converts it from 24.975fps to 29.97fps, keeping the same PAL speedup.

I can't help with your YadifMod problem

[n0an]

Oh okay. I figured Yadifmod out. Thanks!

[krosswindz]

I dont have the entire source for this DVD but from this sample http://www.sendspace.com/file/0kik4n I do believe it is a PAL2NTSC transfer. The pattern clearly points to PAL2NTSC transfer (thanks to manono for this education). Source is garbage as ever.