LexX said:
Are all the SW prints looking the same as PG or would there be others that have more color like TESB print?
Most of the ones I've seen are similar to PG. I think one came up on eBay with better color, and it brought a high price. PG color is pretty good except for reel 2.
Start browsing the preservation threads. There's a lot going on out there, and a lot to wade through. But it's exciting reading. And although most of the efforts have been on pre-SE preservations, I'm pretty sure there are 1997 preservation projects as well.
There are quite a few 16mm prints of the OT out there. They pop up on eBay and the 16mm forum periodically.
If the PSB up-conversion goes well, I may go back and do an upconversion of PG. It would be a fair amount of work, as it might entail having to redo a bunch of the edit/assembly. But at least I kept the original raw scans.
Plan to do three more scans of reel three of ESB today. I'm having a problem with the Workprinter... takeup arm is sluggish and I've been trying to repair that. It still works, just means I have to stay there every second so it doesn't spill film.
I've seen warpage in 16mm and 8mm films that I've worked with. It's a pain in the neck without an easy solution.
First, shrinkage actually means shrinkage... meaning that simply pressing the film flat won't solve the problems of the sprocket holes no longer aligning with the sprockets. This is fairly rare but I saw it once in an older Chinese film.
Second, warping causes the distance between the surface of the film and the camera to change as the film plays, meaning that it is difficult to keep it in focus. This is more common.
If the shrinkage isn't too bad, someone could probably jerry-rig some glass plate method of keeping the film flat to counteract the warping. But it would risk scratching the film in the process. A normal gate tries to hold down the film at the edges... the tradeoff of a more closed in gate would be to crop more of the edges. My workprinter has an enlarged gate which captures more of the frame but also makes it less successful with films with slight warpage. Some of the home movies I've worked with have had this problem, and I've sometimes gotten around it by just trying various projectors until I found the one that was the best match for the filmstock.
none said:
Color-me-stupid. Have equipment (pics or spec) been sent to the knowledgeable? Puggo, etc.
I have no capabilities with 35mm. Well, I do enough to know that making a homemade 35mm telecine is a significant undertaking.
Indeed, this is a fantastic time to be a SW fan, and a fantastic time to be a member of this forum. We are on the verge of having at least THREE high quality reconstructions of the OT: Harmy's, -1's, and DJs (and to a lesser-extent PSB which is looking mighty-fine too). Watching them unfold is really thrilling to be a part of. And when they're out, Lucas pretty much becomes an irrelevant fan editor. Heck, it might be enough to push him into releasing the OT just to try and compete!
mikeveli20 said:
So Harmy's upcoming Despecialized versions will be the same as the upcoming Blu-Ray GOUT except for the fact that that Harmy's will be 1080p and the GOUT will be 720p? Just want to make sure I have the details correct.
One comes from a hi-def source, and one comes from a poor transfer of a laserdisc master. Yes, DJ is doing amazing things with the GOUT, but it will be quite easy to tell the two apart, especially on a moderately large screen.
The appeal of GOUT is that it is entirely OT material. The appeal of the despecialized versions is that the transfers are of the highest quality. Both are fantastic and not in any way duplicate efforts.
thorr said:
Puggo - Jar Jar's Yoda said:Ok, that's a great method. So the horizontal stretching factor appears to be about 1.98. That matches the basic definition of 16mm scope being 2x.
Actually it is more like 2.68:1. I was going to say that the reason for the cropping the way it was was due to the way the anamorphic lenses were defined for 16mm, and that may still be true. If the lenses are about 2.68:1, then they would need to crop the top and bottom to match the lens while maintaining the correct aspect ratio, or add black bars to the frame.
Hey, my theory is correct and so is yours. I found this: http://www.film-center.com/scope1.html
So the original frame is 1.33:1 and then it is 2x'd to 2.66:1 which is what we are seeing.
Right. By 2x, I meant that I need to take the raw image of the frame and expand it by 2:1. Since the original image wasn't square to begin with, the end result ends up being more like 2.68:1.
mikeveli20 said:
So in that case, why even do the Despecialized editions? With the great things that dark jedi is doing with the GOUT, his versions and yours will end up being practically the same.
Harmy's are true full-blown HD resolution -- they are completely different experiences.
Actually it is cropped less than I was fearing.
In working on my 16mm restoration of ESB, I have been surprised by the amount of blue in the Hoth scenes. Not nearly as much as in the 2004 disks, but more than I was expecting. Also, some Hoth scenes on the 16mm are bluer than others.
thorr said:
Yes, I was keeping the height constant on the PSB and only stretching the horizontal until the width and height matched the original aspect ratio. I also had to grow and shrink the original frame by dragging the corner so both dimensions would grow and shrink and keep the original aspect ratio. Eventually I got the original frame to the right height and then stretched the PSB until the width matched. It was kind of a pain. :-) Then when I overlaid the two, I pasted the new onto the old and then did undo, redo, undo, redo so I could flip back and forth to make sure it matched and that is when I noticed the PSB frame was very slightly crooked. The edges are straight so it is the frame inside that is crooked (definitely not a big deal).
Ok, that's a great method. So the horizontal stretching factor appears to be about 1.98. That matches the basic definition of 16mm scope being 2x.
It was more important to me to make the edges straight, so that when it comes time to crop, there won't be much image loss. By the way, it is not at all easy to get the frame straight with the workprinter.
thorr said:
Hi Puggo,
- I gave it a shot and I took the frame of the two walkers and cropped it to 624x462 to get rid of the borders.
- Then I stretched it to 1236x462 and it looks pretty close to correct. Your scan is very slightly crooked so it doesn't match perfectly, but I can't get it any closer.
So at 462, the height of the material actually matches exactly? I was assuming that the whole thing would need to grow or shrink somewhat to make it match.
Thanks a lot for your analysis! This sort of thing helps me out a lot. I'll post a short clip and if you (or anyone else) wants to try running a BD test using your formula above, that would give us an idea of how well this might work.
A comparison group like that, but including more human faces, would make it easier to decide on a favorite.
Asaki said:
I might be way off the mark here, or you might already have considered this, but is there a way you could do like they do with HDR photography?Get one capture very bright, one very dark, one in the middle and layer them together?I know you said you're already capturing at different settings, I didn't know if that's what you meant or not. But then again, trying to get them all to line up perfectly would probably take too long to do =/
It would be fun to try, but I fear there would be too many slight variations in the centering of each frame, such that it would just make it look less sharp.
I actually have tried this technique with old video (I do 1/2" EIAJ as well), and while the results were promising, I ultimately was dissatisfied with the loss of sharpness.
So, anyone care to take the frames I posted, and do some investigation into the appropriate percentage of horizontal stretch to make the material match the ratio of the original movie? I haven't even found a definitive definition of the horizontal percentage stretch for 16mm scope projection. And I wonder if it wouldn't just be more accurate to eyeball it against some GOUT frames. I can do it of course, but some of you might do a more exact job than me.
DJ, your avatar is giving me nightmares...
Jar Jar force ghost?
Darth, What's His Name? said:
You got a lot of nerve. What happened to deleting all your posts and starting your own forum? Why not just stay there? In any case at least your future posts will provide us with lots of comedy relief.
Huh?
thorr said:
The current PG is a bit squished vertically from where it should be, so there is another reason to redo it. ;-)
Are you sure about that? Several of us went back and forth on that many times to be sure. The aspect ratio won't be exactly the same as the original film, because with 16mm there is significant cropping before the scope stretch, so you can't tell if it's right just by measuring the sides and calculating the ratio... you have to compare elements side by side with the GOUT.
Follow-up... it looks like the curved corners are part of the print, not the gate as I had presumed before. That means they stay. :)
Just finished two more scans of reel 2. Now back for some more scans of reel 3
Let me explain the dilemma with respect to the corners:
The scans you have all seen (with the rounded corners) are with the camera placed at a slightly further distance than was used for the PG. The pros of doing this are: (1) the side and top/bottom borders are more straight, and (2) the focus is slightly sharper throughout the image. The cons are: (1) the cropped rounded corners.
It isn't possible for me to get a wider gate... the gate I am using was specially ordered, hand enlarged by Roger Evans for capturing scope films with the workprinter. When I mentioned the rounded corners to Roger, his response was simply, "move the camera closer". Thus the tradeoffs mentioned above.
The SW print used in PG was such that I was able to capture the entire frame without any cropping, and only a very minimal parallax that I worked around with the gradually curved border on the top/bottom. I experimented with removing the parallax in software (there is an avisynth filter for it - I think it's called "barrel filter"), but the results were awful and I never bothered posting examples.
I haven't decided if I will use the same approach with PSB, or if I will accept the round corner croppings that you have already seen. Maybe something halfway in between. Once I decide, I'll make a mask that preserves as close to every bit of the captured image as possible.
dark_jedi said:
I would say this project is in VERY capable hands, it is not like Puggo has not done this before, I say you do what you think is right Puggo, I and I am sure a lot of others here will be VERY happy with whatever you end up doing, that said, good luck, can't wait to see it.
Actually, I greatly appreciate all the suggestions. I would have screwed up the PG royally if not for the excellent suggestions I received. And already I'm getting suggestions here that are changing the way I am planning to encode it (when I get to that point, which is likely a few months away).
thorr said:
Can you pause the capture and adjust the camera scene by scene? Another idea would be to do multiple captures with different settings and take the best version of each scene before adjusting in post.
Yes that's what I do. PG was pieced together out of 3 or 4 passes on each reel. I'm actually not done making capture passes through ESB, and I'm not convinced that I even have the right camera placement yet for this print.
Video Collector said:
May I just say that I second the request for a version that keeps the rounded corners visible. I believe it is an insignificant price to pay to preserve the maximum image area. The extreme outer edges of the frame are worth preserving too :-)
If I do mask the corners, it would be with a rounded mask so as not to lose any information off the sides, nor off the top/bottom. Don't worry, I certainly won't just chop it down to make square corners! The reason why I might ultimately chose to mask the corners with rounded borders, is that sometimes on the print those out-of-boundary areas can have all kinds of flickering distractions. Individual frames might look fine, but when in motion it can look worse. I'll see what it looks like when I get to that point. Thanks again for the suggestions!
Nerfherder said:
I think white balance is very important. Parts of the Puggo Grande seem very 'blown out' (almost like too much light was being projected through the film?)
I think you may be confusing "white balance" with "brightness/contrast". White balance refers to shifting the color balance so that the things that are supposed to be white are actually white. Actually, both things (white balance and brightness/contrast) are difficult to get accurately in a film capture. Film has a much larger spread than video cameras of being able to reproduce detail in bright areas and dark areas, and so one is always balancing whether to favor the details in the dark areas or the bright areas. The blowout of some of the whites in PG had nothing to do with white balance, and was more due to my trying to get people's faces to look as good as I could.
I'd be tempted to try out a raw capture without any processing at all and see how that looks, otherwise one is altering the print. It's nice to preserve exactly how one would see this if projected, even though the temptation is there to keep tweaking the settings.
See, that's the problem. A telecine capture doesn't ever preserve exactly how one would see the film if projected, because of a myriad of factors. The range of colors you get on the video depends on how the film is illuminated (what type of light), how the lens settings respond to the lighting used, the white balance setting on the camera, etc. Cameras don't work the same as our eyes do. And if you find a setting that works on one scene, it generally doesn't work as well for other scenes... sometimes it's way off. So, if you want it to look as close as possible to an actual viewing of the film, you HAVE to tweak it scene by scene in post.