THX 1138 "preservations" + the 'THX 1138 Italian Cut' project (Released) — Page 34

The IVTC process was not the reason behind the frame difference, now I have to scan through the captures and see where they differ. :(

Are you sure you're not trying to mix captures of the same LD, but different player? That's the only possible variation I can think of at the moment...

I'm trying to do a median of the three captures from the US LD, for some reason they differ in frame counts when I line them up, don't know yet if there are dropped frames or duplicates. I suspected it was the IVTC that caused this, but the same problem remains when I now line them up before doing IVTC.

Did I do a capture from the 704? That has some internal video processing that can't be switched off. I have music video discs that play fine on my 8210, but "smear" on the other two machines.

Even with everything switched "off" on the 704, I could see some ghosting going with hologram guy's head when he walks up to THX and SEN in the white void. It was even worse with the noise reduction circuits turned on with a test capture I scrapped.

Yes, you did. :) Of the three captures, the 8210 is definitely the best but it do have those rolling bars and artifacts on bright objects, THX first conversation with SEN also have what appears to be laser rot which is not present on the other captures.

Haven't noticed the smearing you describe on the 704, but I think this one is the second best out of the three from what I can see, must take a look at what you described. The V4400 is very soft and blurry compared to the other two, but all have their advantages over the other.

A median of all three captures should make the best out of it if I only can locate all the frame differences...

Wow, that sounds like a hell of a tedious work. I wish you strength and good patience with it, my friend :-) Oh, and I can't wait to see the result, the samples looked awesome :-)

Harmy said:

Wow, that sounds like a hell of a tedious work. I wish you strength and good patience with it, my friend :-) Oh, and I can't wait to see the result, the samples looked awesome :-)

Thanks, I will need it. :) I will test and see if it's really worth the trouble going through these frame by frame. It's nothing I am looking forward to...

Yeah, I believe you, stuff like that happens to me all the time, as I'm dealing with putting together footage from two different transfers.

Ok, some good and bad news. The good news is that I don't need to scan through these and look for frame differences because I am not gonna be able to do a multi cap median on these even if I wanted, which is the bad news.  In order to do this they need to match up perfectly and these don't, which you can see below.

The US LD from three different players:

PR-8210

LD-V4400

CLD-D704

The Japanese LD:

PR-8210

LD-V4400

CLD-D704

I am still waiting for the PAL transfer from Jaitea, hopefully I'll be able to use that on the badly framed and dirty 15 minutes segment... either way it's going to take some time because I plan to get rid of the distracting dirt once and for all. And no response regarding the re-release trailer.

I'm really disappointed the 4400 is that soft looking. I expected better from an industrial grade player. (It's built like a tank.) I was also about to dive into capturing a bilingual disc that only has English on the digital tracks, so the 8210 is less than ideal for that.

I may have to keep my 704 around now that it's fixed.

This is also one of those scenes that looking slightly worse than the rest of the transfer, you can clearly see that the US transfer is the best looking, the Jap have a different black level but it also have lots of chroma noise which is almost totally absent on the US disc. Is the chroma noise something that gets introduced because it's a Japanese NTSC signal?

I wish I knew. It would have to be something done to the video master since it's obviously the same print. If it's not a digital video master, would they lose a generation adding in the subtitles?

SilverWook said:

I wish I knew. It would have to be something done to the video master since it's obviously the same print. If it's not a digital video master, would they lose a generation adding in the subtitles?

Good question, I'm certainly no expert in the field but I doubt adding in subs electronically would degrade the transfer, but I don't know exactly how they do/did it either, so I guess it's possible. The JSC LD's of Star Wars also seems to display more chroma noise/rainbowing than their US equivalent.

Anyway, I'm starting to like what I see when I'm doing my small test encodes, dirty or not, I think this is gonna be a damn fine transfer. Here's another sample for you: http://www.sendspace.com/file/roskh7

Again, no audio, I'm too lazy for it ;)

I have also found an annoying "flashing white artifact" in the left of the frame which is there largely throughout the film, (you can spot it in the sample I posted) I have no idea what it could be, never seen anything like it before but I guess he is a good friend of this guy:

Here's a few pictures of it:

Doesn't look like regular dirt either.

msycamore said:

No, I had problem register on that site for some reason, maybe you could give it a shot? It would be really nice to have it in better quality.

SilverWook, can you do me one last favor, how does the color look on this computer screen...

...on that LD/Beta transfer? Don't necessarily need a screenshot of it, I'm just a little curious about this one as the colors are very unstable on this particular screen. (it's bright green in the 2004 DVD) Want to rule out any possible flaw in the LD transfer.

Here is the '86 LD shot.

[/IMG]

Uploaded with ImageShack.us

Thanks! Blue it is then.

It's bad enough that thing floats around the edges of my vision when I'm not watching the movie! ;)

erri_wan said:

The white thing has become your avatar? LOL

SilverWook said:

It's bad enough that thing floats around the edges of my vision when I'm not watching the movie! ;)

Sorry, I wanted to keep my enemy close. ;) I'll see what I can come up with, was tired of seeing my old one.

No comment on the sample and the weird artifact I pointed out? Here's an even better pic of it:

Any idea what could cause such a strange artifact? It dances around in the left of the frame almost throughout the whole film, it can be very distracting in some scenes, I'll see if I can do anything about it.

This is starting to become very time consuming, I recently went through the 8210 capture and noticed a few bad frames that slipped through the automatic IVTC process I made, something that actually should be expected as no automatic IVTC'ing is bulletproof so I'm currently going through and manually correcting the few interlaced frames and damn is it boring! But I want this to be perfect so I guess it'll be worth it in the end.

Jaitea also sent his PAL transfer my way earlier this week, so I should have it any day now which is great. :)

Also, I noticed that I resized my earlier two samples I posted a few pages back incorrectly, the last one have the correct aspect ratio. (In case you wondered why it looked a little funny)

There's a similar thing going on with the print Puggo used for Stjärnornas Krig in several places. It could be a light source leaking into the telecine?

And I think I've finally identified our mysterious white friend. ;)

EDIT:Fixed broken image link. I have no idea what happened with the first one!

SilverWook said:

There's a similar thing going on with the print Puggo used for Stjärnornas Krig in several places. It could be a light source leaking into the telecine?

 

There is? haven't noticed. Yeah, my thoughts was also that a light source could have caused those weird spots, just didn't think it was possible?

SilverWook said:

And I think I've finally identified our mysterious white friend. ;)

I knew it! he really is a evil looking dude. ;)

Maybe a potential new avatar...

I think we were lucky to even get THX on Laserdisc. I don't think they would have had a video master, so they probably made one from their archived print of the director's re-release. I agree with SilverWook's thought about it possibly being light coming in from the telecine. With it being in the corner I'd say maybe even leaking in from sprocket holes that were damaged. That used to happen on some of my 16mm student projects.

And the white fiend is...?

Aside from Criterion, I don't think too many of the studios were going the extra mile in proper video transfers of old dodgy print elements, even in the early 90's.

I just watched an animated film on a Hong Kong Laserdisc, (circa 1990) and I was surprised at the tracking errors and dropouts of the video master used. The soundtrack was in mono in spite of a Ultra Stereo credit at the end.

The white fiend is Coily the Spring Sprite, star of the short film "A Case of Spring Fever", made infamous by Mystery Science Theater 3000. :)

http://www.youtube.com/watch?v=v58argQ3l-M

captainsolo said:

I think we were lucky to even get THX on Laserdisc. I don't think they would have had a video master, so they probably made one from their archived print of the director's re-release. I agree with SilverWook's thought about it possibly being light coming in from the telecine. With it being in the corner I'd say maybe even leaking in from sprocket holes that were damaged. That used to happen on some of my 16mm student projects. 

Interesting. Overall, I think this transfer looking pretty good, especially compared to other films of this age released on LD in the early nineties, some scenes are worse than others but sometimes it look surprisingly good.

I have actually no problem with the dirt on this transfer except those two distracting things in the left of frame, they could've easily made it less apparent if they had just centered the framing like they did on the PAL release, the NTSC transfer is mostly cropped on the right side causing problem with the films stylish cinematography in a couple of scenes. That along with the oversharpened image maybe makes the PAL widescreen LD release superior even though it losing more picture-info overall.

I absolutely don't want to use any degraining on it but do anyone got any thoughts and suggestions on the haloing/edge enhancement in this transfer? As I understand it, if I try to reduce it with filtering I'll just causing damage to the rest of the picture, resulting in a softer image, you cannot get rid of it completely, the damage cannot be undone, or am I wrong?

The more you do to a LD cap, honestly the worse off the image will be. I did some similar things on some basic LD conversions and the end result looked horrible, especially when scaled to 16x9. I don't really think there's much you can do with filters applied to the source without compromising the image.

From what I've seen of the angrysun LD cap, this transfer is pretty good as far as dirt goes. I was able to watch through it without any real gripes. That white line on the left side at first looks a hair in the gate that was captured in the source transfer. On DVD it will clearly look like a hair, but on LD it gets very digitized and pixelated. Why it's white I don't know, but in the scene where it first appears you can see it fly up into the frame and get stuck there. Later on it shifts slightly in the frame. But then again, it doesn't move all that much so it could be print damage. It does stay for about 11-12 minutes so that could suggest that a reel was somehow damaged.

As for the light source, it doesn't even have to be damage. It can just be the actual filmstrip weaving around in the gate. At school the Steenbeck I had to use wasn't in the best of shape and this would happen to me all the time whenever I ran anything back. It needed some kind of repair or cleaning so the image on the viewer would kinda bounce around to every side while running. Then the corners of the frame would occasionally have flashes of brightness or light-this can mean heads need to be cleaned or wheels replaced or a bunch of other things.

This I'm fine with. The white mark now looks like print damage but I really didn't notice it before. It would be nice to see it fixed, but I don't think there's really a feasible way to do so.

I think that the white thingy is quite clearly a hair that got stuck in the gate but not during the telecine but when the print used for the telecine was being struck - that would explain why the hair is white and why it is only there for what seems to be one reel.