team negative1 - star wars 1977 - 35mm theatrical version (Released) — Page 142

Yeah I think you’re right about the emoticon.

Arnied said:

Williarob said:

Therefore I’m re-encoding the main movie again to incorporate the Reel 3 cue mark fixes.

I love the quality of the film and especially the grain structure.

However, since you are going to re-encode again, is there something that can be done about the compression artifacts that are present in some parts of the film (mainly cantina sequence and some other dark parts)?

See for example:
http://postimg.org/image/7s64vt2sb/full/
http://postimg.org/image/qbc789hzv/full/

At 40 MB/s I doubt those are compression artifacts. If anything, they are imperfections brought out by the color correction process - these shots were brightened considerably from the original scan, and some of the shadow/color information just isn’t there. I suspect Poita’s scan will be considerably better, but until it arrives, this is as good as it gets I’m afraid.

yotsuya said:

I can’t speak for Darth Lucas about the recreation of the 81 crawl, but my purpose has been to recreate what was seen on LD and home video. While the handful of prints of the 81 crawl that have been examined don’t show many stars the LD archived versions do so many many stars at a brightness relative to the following scene. Also, in the version I have been working on (which is not complete due to the difficulty of fixing up the flyover portion of the TN1 35 mm scan included with the SSE 1.0 - for me no other source is acceptable) I am trying to use only components that are true to the 81 crawl. The title fade out is identical to TESB, the crawl is identical to the SE, the stars, moons, and planet have been recovered for the most part by Poita, that just leaves an accurate 1981 flyover. It is different in 77, 81, and the 97 SE so for me only the 81 flyover will work. Darth Lucas has pulled his starfield from TESB - which used the same painted starfield, but not the same shot of it. To properly match the 35mm scan of the 77 crawl, some added grain and a tiny bit of gate weave really make the difference. I’ve had too much on plate to work on the flyover or the improved starfield/moons/planet. To be honest, from the quality of the scan, I am going to have to manually edit ever frame of the flyover (roughly 400) to align each from to a known stable version. Some frames are warped and stretched noticeably.

I admire your dedication and look forward to seeing the finished product. In the version I created, using the '81 crawl with the '77 starfield, I used the “find Edges” effect in After Effects and then several instances of the Paint Bucket Effect to fill the old starfield with blue paint. Then I turned off the find edges effect and keyed out the blue to reveal the new starfield underneath. This worked very well for the crawl, and may also work for the flyover but I never got around to trying it. Give it a go, it may save you a lot of time.

yotsuya said:

I can’t speak for Darth Lucas about the recreation of the 81 crawl, but my purpose has been to recreate what was seen on LD and home video.

I think that sums it up well: you’re recreating the home video experience, whereas the SSE aims to recreate the cinema experience. I’m not saying that Rob should use one or the other for the BD ISO, just that I’d prefer the scan not the recreation. Perhaps he could simply make an alternate MT2S file that we could use as a drop-in replacement? That way everyone’s happy. 😃

Also Rob (& others who worked on v1.5), great work on this!! I haven’t had a chance to look it through yet. I love the name “Silver Screen Edition” and I think you should keep it.

As for branching that some people have said isn’t seamless - this could be because the playlist settings are not correct. When I open up the playlists in BDEdit they show a 10 minute in-time instead of 10 seconds (11.650 seconds is usually what I see on commercial discs) - Rob you might want to fix that, even if it means opening up the playlists in BDEdit and changing them manually after the disc is built. Also, I noticed you didn’t include the Teaser scan - it would be nice to include that on the next version if possible. Also the menus should be set up in the same way, and they aren’t in the SW 1.5 ISO (i.e. 11.65 seconds in-time and first item has a value of “1” under “c” column while other items are “5”), again you can change all those values using BDEdit after the disc is built if needed and it should work fine (but make sure you test it and keep a backup of the playlists). 😃

Here’s a graphic showing what I mean.

towne32 said:

I know you’ve done quite a bit of analysis of some of these flags in commercial discs, Valeyard, but it turns out to be quite simple to set up seamless branching with the right software.

You can do the same yourself. I looked at Blade Runner, Alien 3, T2 Skynet, and they all look identical. And you can compare to even more discs with the menus - they always start with “1” for the c value of the first entry and then “5” for all other entries, no exceptions, & “In Time” is always 11.650 seconds (branching is done in the same way to seamlessly looping menus). Changing those settings in BDEdit shouldn’t be difficult or cause any problems for the disc.

RU.08 said:

towne32 said:

I know you’ve done quite a bit of analysis of some of these flags in commercial discs, Valeyard, but it turns out to be quite simple to set up seamless branching with the right software.

You can do the same yourself. I looked at Blade Runner, Alien 3, T2 Skynet, and they all look identical. And you can compare to even more discs with the menus - they always start with “1” for the c value of the first entry and then “5” for all other entries, no exceptions, & “In Time” is always 11.650 seconds (branching is done in the same way to seamlessly looping menus). Changing those settings in BDEdit shouldn’t be difficult or cause any problems for the disc.

The seamless flag is set correctly and even in BDEdit the c value for the first entry (the crawl) is set to 1 (non seamless) and the second entry (main movie) is set to 5 (Seamless), just like in the retail discs.

I don’t know why the in time is at 10 minutes (though I do see that in BDEDit), there is no indication of that anywhere in the software, all start times are set to 00:00:00:000 in the properties window and these settings are read only. I can try changing the in time in BDEdit, but I really don’t see what effect that has on anything, it’s not like the movie doesn’t start until the 10 minute mark.

Also, you said the teaser scan wasn’t included, but it is, along with the other two trailers that were included with 1.0, or am I misunderstanding?

If you are using the software I think you are using then you should do an auto layout as one of the last processes to interleave the seamless branches so that there will not be a judder or pause. You’ve probably done this but if you haven’t it’s pretty essential when doing a dual layer disc and very essential when doing seamless branching to make sure all the stuff gets put physically where it should be on the disc.

The disc layout had to be done manually (“Automatic Layout failed to find a suitable location to place the layer break”), so I set the layer break 1 hour and 4 minutes into the movie, which is a long way from the branch point just 2 minutes in. It is 100% seamless in PowerDVD and also on my LG player, which is good enough for me.

Meanwhile I still haven’t resolved the subtitle sync problem. If you are just joining us, this is the story so far. Using Jubler, each .srt file (taken directly from project Threepio) was split in to two parts (Before the flyover and after the flyover). Also in Jubler, I retimed the second part, so that it counts up from 00:00:00.000 instead of 00:02:26.xxx and verified that they stay in sync from start to finish. That’s all good. Now, since I can’t just import .srt files, I then converted them using Lemony to the correct format, and checking the xml file it created (along with thousands of .png subtitle image files) I have verified that the timing is still correct, in that the in and out timecodes match their .srt counterparts. So far, so good.

Now when I import the BDN files into the authoring software, it asks me if they are drop or non-drop, and regardless of which option I pick, they still drift out of sync.

As I understand it, the original .srt files are timed to match a true 23.976 timecode, but the film is almost certainly using a 24 fps timecode, which is why they slowly drift out of sync. (Excellent explanation of all this here: https://forums.creativecow.net/thread/279/1230).

It sounded to me like the solution should be to convert the timing to match, so, using Lemony, I tried adding a frame rate conversion (23.976 to 24) and rebuilt everything. (It takes about 5 minutes to convert the subtitles, another 2 or 3 to import them, and then about an hour to rebuild the disc before I can test it, so it’s a time consuming process!)

It still goes out of sync (though at a different rate than before) and this is regardless of whether I select drop or non-drop on import or mark the film as drop or non-drop (I think I’ve tried all combinations). At this point I am struggling to wrap my tiny brain around the problem. Any ideas?

I’ll let you know in an hour or so…

JEDIT: The SUP files are not compatible, and would still need splitting and re-timing even if they were.
JEDIT2: Actually, setting it to 30 FPS probably won’t work, because when it gets converted to BDN, it uses a frame number instead of milliseconds, so it a title has a frame number > 23 it might do something weird…