Pompeii project (NTSC)

[Morbo]

Ok,my goal is to return Pink Floyd's Pompeii to 24 fps from 29.97.

The whole thing is a problem I;ve worked on and off for years(previously with the laserdisc).

The thing was filmed,and it's PAL(I'm sure I'd be better off with PAL source,but....) sourced,which means that the pitch of the music is WRONG(I've already fixed that,sounds alot better and is still synched).

The DVD we have is badly interlaced,....the progressive player I have(on a regular TV even) artifacts BAD on the amps during Echos enough to see huge blocks(Laserdisc didnt even do that!).

I've fiddled with every IVTC method I know of inside of AVS,and was wondering if I posted a short clip.....maybe I could get some insight.

Say a minutes worth..

I'm eventually gonna get a plasma,and this film is great.....but the artifacts look like defects....:angry:

Thanks!

[Guest]

Yes, post *unprocessed* source (a VOB fragment), as large as you can showing a problematic section.

[Morbo]

http://www.megaupload.com/?d=Q7UT5OV9

Thanks!

[Pookie]

I love that scene where Gilmour plays those cool licks over "The Lunatic is on the Grass".

[Guest]

It's a field blended conversion. Do a search on "blended fields" and prepare for major hassles. You cannot IVTC that by normal means.

[Morbo]

It's a field blended conversion. Do a search on "blended fields" and prepare for major hassles. You cannot IVTC that by normal means.
I've been at it for long time already,..:D ..but thanks..

If I had known what I know now,I would have ordered the PAL version somewhere.....and still might.

Shame,this is a classic.

[Pookie]

Have you gone the Restore24 route (although I'm not certain it works on 29.97) ? FixBlendIVTC and Cdeblend, perhaps?

[Morbo]

Nah,just the basic telecide,etc....

Also,Tsmpeg's IVTC methods......etc,etc...

Even looked at Dynapel's demo..(for this and other
nasty DVDs..if it works OK.)

[G_M_C]

Ooohhh .... i know we arent supposed to talk about stuff like this, but ....

But we in PAL-land have those discs in PAL offcourse. Ive seen that disc for something like ca. 10 euro's or so.

So I might be able to get that disc in PAL version for you. The only problem beeing, that it has to get to you somehow ;-)

Provided that you've got a player that handles PAL offcourse (many of our european players handle both NTSC and PAL, my denon DVR-1920 does anyway). And offcourse you're shure that a PAL source would be much better than the sopurces you've got now :)

Having a better source would help you out a lot, and since you seem to be a big fan of Pink FLoyd (as I am) a couple of bucks should not make a difference ?

[Morbo]

Yeah, I use a Philips DVP-642......plays everything(but mp4 with GMC and QP).

I've already got a few places earmarked for a purchase though.....thanks.

Sad part is,the laserdisc was 35 bucks,the DVD was 20,.....now...I may have to get it again...

But oh how I love that film....;)

[jmac698]

repal(), or unblend()

[Pookie]

'Twas reading this thread and found an easy way to identify and display blends -> http://forum.doom9.org/showthread.php?p=726948#post726948

Also provides a method to visually compare some of the various blend removal scripts out there. The bottom portion is the original image, the top was processed with

Cdeblend(dmode=5, thresh=130) #This removed around 80% of the blends

http://fileserver1.jpghosting.com/images/tn_1_ca8261b51f569762a5e92851c259583c.png (http://fileserver1.jpghosting.com/images/1_ca8261b51f569762a5e92851c259583c.png)

http://fileserver1.jpghosting.com/images/tn_7_aa5fd9f293ddccd9b0d162e547d885fd.png (http://fileserver1.jpghosting.com/images/7_aa5fd9f293ddccd9b0d162e547d885fd.png)

LeakKernelBob(1 or 0 depending on parity)

#Place your blend removal script here

source = last
Anglecriterion = blendangle(source).isabove(100)
weight = blendweight(source)
weightcriterion = And(weight.isabove(0.1),weight.isbelow(0.9))
criterion = And (Anglecriterion, weightcriterion)
if (criterion,\
then = source.subtitle("[-BLENDED-]"), \
else = source.subtitle("[ ]"))

[scharfis_brain]

I think that manual pattern deblending would be the better way here. It was a progressive PAL 25fps Production so phase shifts are unlikely. This means the blending pattern "should" be constant throughout the whole movie. This makes adaptive methods unnecessary.

My manual deblending appraoch is the following:

MPEG2Source"Pompeii.d2v")
TDeint(mode=1, tryweave=true, full=false) #bob-deinterlace with fieldmatching where possible
changefps(25.0*10) #change to ten times the original framerate (25fps)
selectevery(10, ?) here the deblending takes place.
# adjust the ? within a range of 0..9
# until motion is fluent an no blends remain.


with this sample I used the value 1 .