Harmy's RETURN OF THE JEDI Despecialized Edition HD - V3.1 — Page 107

Clutchins, you’re not supposed to display the link in your quote. Let’s not let this spread through carelessness.

Tom220 said:

Hey Harmy, first of all thanks for all the work you have done to preserve these films. I can’t tell you how grateful I am.

I watched most of Return of the Jedi last night and I couldn’t help but noticing an issue with the blacks. They look severely crushed in 2.5. I thought this might be due to the source material with the BluRay, but then I looked at version 1.0 and the blacks didn’t look as crushed there. In some scenes it’s more obvious than others, but this is a glaring issue to me and I surprised no one else has brought it up. I also felt as though the brightness level is set a little too low. Again, comparing it to 1.0 where the brightness level felt right. Fleshtones and overall colors are a bit too red for me personally, especially on the Endor scenes. Han looks like he’s spent three years on a tanning bed, not frozen in carbonite. I saw a print of Jedi about 4 years ago and I don’t remember it looking like this. Just wanted to share my thoughts.

1.0 looked a bit overly lightened, kind of bleached out.

But no, you’re not the first person to make this complaint. I think you are the second. I know you say you’ve seen a print, but part of the issue is that everyone who has seen multiple 35mm LPP scans can tell you that some parts really are quite dark. He includes so much of the 35mm footage, that I suspect he pushes it to as far as is reasonable, and then brings the BD footage to a matching level. Either way, aside from a couple Emperor shots, it should be at least as bright as the BD. I do wonder if the shot of the crossed sabers in front of his face might have been better with increased brightness, regardless if it makes his ‘slugs’ visible. But that’s my only contrast complaint.

As far as the red skin tones go, it is indeed quite variable. Those extremely bright shots are thanks to the 2004 mastering. As someone who has messed with the color of this film, the skin tones do turn bright red when you bring down the blues (generally, the film needs that done, though to a lesser degree than ANH. I’ve done it manually and with Neverar’s LUT, and it results in the way Harmy’s looks). The severity of the redness depends on the monitor and calibration, but I agree that a few shots looked more natural in 1.0.

Mavimao said:

FYI, there are no technicolor prints of Jedi (just the original Star Wars is so lucky). A kodak low fade print was scanned and referenced.

I said LPP, I’m not sure if someone else was confused or not about this point. As great as Technicolor is, when done properly, I would much rather that a SW77 LPP existed than a Jedi IB Tech. 😃

Edit: Nevermind, someone posted before me and I didn’t see it as it was on the previous page. 😄

ilovejarjar said:

Sartre said:

Harmy said:

Did you try demuxing first?

Yes. When I demux the h264 stream before remuxing, the movie does start when it’s double-clicked but it still takes a while to catch up when you click in the seek bar. I notice when muxing im mkvmergeGUI that it takes a long time for the progress bar to show anything, like it’s reading through the entire file, then it shoots up to a hundred percent and takes another minute as it seems to write the file. I do lots of video converting/muxing and this is definitely unusual behaviour. I analyzed it in avinaptic and it showed a lot of buffer overflows:

Selected profile DivX 720HD
Resolution Ok
Framerate 23.976024 <> 30
Buffer underflow 00:12:38 (frame 18173)
Buffer underflow 00:12:39 (frame 18204)
Buffer underflow 00:12:41 (frame 18234)
Buffer underflow 00:12:49 (frame 18449)
Buffer underflow 00:13:02 (frame 18759)
Buffer underflow 00:13:04 (frame 18808)
Buffer underflow 00:13:06 (frame 18843)
Buffer underflow 00:13:45 (frame 19769)
Buffer underflow 00:29:09 (frame 41925)
Buffer underflow 00:29:11 (frame 41973)
Buffer underflow 00:29:12 (frame 42014)
Buffer underflow 00:29:28 (frame 42390)
Buffer underflow 00:31:14 (frame 44930)
Buffer underflow 00:32:03 (frame 46097)
Buffer underflow 00:32:07 (frame 46198)
Buffer underflow 00:32:09 (frame 46259)
Buffer underflow 00:32:11 (frame 46287)
Buffer underflow 00:32:28 (frame 46709)
Buffer underflow 00:32:39 (frame 46970)
Buffer underflow 00:32:42 (frame 47048)
Error Too many violations

It’s trying to play it with the DivX (MPEG4-ASP) directshow codec and not H.264 (MPEG4-AVC). MPC-HC has its own codecs. You don’t need CCCP, ffdshow or anything like that to play it back, use the internal codecs:

http://www.ezoden.com/htpc/11/mpc-hc-internal-filters

If that doesn’t work try VLC which only uses its own codecs.

No it isn’t. The divx reference is a quirk with avinaptic. I’ve tried internal and external avc filters in MPC as well as VLC and every other player but no go. VLC doesn’t show a picture at all.

Sartre said:

No it isn’t. The divx reference is a quirk with avinaptic. I’ve tried internal and external avc filters in > MPC as well as VLC and every other player but no go. VLC doesn’t show a picture at all.

You’ve probably got a corrupt download then. If you’re looking for an mkv why not just bite the bullet and get the larger version?

This is great and I’m really looking forward to showing the OT to my son without all of the “modern” touches. Thank you, Harmy (and other contributors), for giving so much of your time and expertise to this project.

Performed my first MKV muxe to BD ISO tonight. Simple and fast and worked like a charm. Kids are already engrossed in the newly burned Blu-Ray as I do the dinner dishes. Awesome!

CatBus said:

Often it’s not a matter of hardware. Most software just decodes the lossy core of the DTSMA track, but if you have something special that may be trying to decode the lossless file, it could be making a mess of it. Which is why I asked you to see if an AC3 track synced OK. The 93 English AC3 track is within a few milliseconds of in sync with the 5.1 DTS-MA track. If the decoder isn’t the problem, they should have the exact same sync issues.

Just tested on a laptop with headphones in VLC. Both the 5.1 and 1993 2.0 track seem equally out of sync (audio lagging behind video). The scenes I’m looking at are the beginning of the crawl, when Leia unfreezes Han, and when Vader talks with Luke in the Endor hallway after he turns himself in.

For sanity, I verified the MKV checksum. Is anyone seeing this sync issue?

Harmy said:

He’s definitely wrong in that v2.5 is darker than v1.0 - there’s a couple of scenes in v2.5, which are a bit darker, because from what I can tell from the 35mm LPP print and Puggo’s 16mm print, they were originally darker but generally v2.5 is definitely brighter and has less crushed blacks.
Actually, here’s a comparison gallery showing v1.0 vs v2.5:

http://picsurge.com/rrGUtW

For best comparison, I recommend downloading the whole gallery and viewing it in full screen mode in your image browser.

Harmy - I definitely see what you are saying with regards to many parts being lighter, but in some frames, even if you lightened a portion of the frame, you really lose the shading detail in the blacks (if I am correct in that is what is defined as black crush).

For some comparison frames:
Stills: 57-58 - loss of detail to Luke
Stills: 149-150 - loss of detail to emperor clothing, architecture, and Vader’s helmet.
Stills: 155-156 - smaller ship is absorbed into space
Stills: 181-182 - loss of ship detail behind Luke.
Stills: 183-190 - crushed detail in vegetation shaded areas
Stills: 223-224 - loss of architectural details, emperor’s robe detail
Stills: 241-242/245-246/243-244 - loss of Vader lower mask/chain patterning detail

Fantastic work. This is now my preferred version of ROTJ. Can’t wait for the ESB and SW versions with 35mm footage.