Harmy

I'm sorry, I didn't mean to be a dick - what you've been able to do there is pretty amazing - in my original comment, I just wanted to say that the difference can't be properly appreciated at this resolution.

And I actually agree that preserving print quality is too low of a goal for a good restoration but since you have multiple prints, couldn't you just compare them and see, which flaws (in this particular case misalignment) they have in common and preserve those - In this particular case, the misalignment of separate elements could only have been introduced during the compositing and therefore was definitely a part of the original negative of the movie and any misalignment introduced at any other stage of the process would have to be consistent over all the elements, so if you looked at a Kodak or Eastman print and the over-all misalignment wasn't there, you could safely assume, that it wasn't on the o-neg and was introduced in the I.B. printing process and if it was there on three or more different prints, it should be quite safe to assume that it was on the o-neg.

This kind of scrutiny wouldn't of course be normal for a standard restoration but you set the bar so high with the level of scrutiny that you're doing for your project, that doing this doesn't seem that far fetched and, since as a result you could do a single alignment of the entire frame and arrive at something actually closer to the o-neg, it actually seems easier than rotoscoping and re-aligning all the elements separately.

That is undeniably true but after a simple color adjustment the raw footage does look much better than before. Plus if the channels are off by a different amount on different elements, the misalignment must have been introduced in the original compositing and not in the making of the print and thus realigning the elements separately (not to mention harvesting detail from neighboring frames, which was never even on the negative) is an act of revisionism. All of that is of course ok for Mike's version, since he isn't going for 100% original and I completely agree that it is amazing what can be pulled from film but I don't think it would fit a proper restoration.

I guess at 4K the difference is probably quite big. At this resolution, the main difference is the color, mainly the black level, which is of course completely off in the untouched scan video.

My new setup (click to view the whole gallery):

Oh, yeah, I'd love to see those too. And I think, if he doesn't count on making a profit on those films, it would be pretty easy to just put them out on-line for free or for a small fee.

Yeah, I do this all the time for single shots and it's a big ol' pain in the ass, so doing it for a whole movie would be more trouble than it's worth.

Yeah, I definitely like TN1T better than DOUT but I'm sure TN1 will come up with some cool original name for their project, like they did with their Renagade Grindhouse version.

Yeah, it was definitely mastered digitally. How else could the master have all that DVNR (Digital Video Noise Removal)?

Also, the '93 LD master (source of the GOUT) was digital too.

Here are a few screenshots:
http://s25.postimg.org/itfnsztal/Pre_comp_1_0_00_56_01.png

http://s25.postimg.org/ay0j0fua5/Pre_comp_1_0_01_38_00.png

http://s25.postimg.org/og7fcq6fh/Pre_comp_1_0_01_33_03.png

I find this idea of trying to simulate a CRT display equal parts ridiculous and fascinating, so I decided to take a stab at it myself :-D

https://drive.google.com/file/d/0B0dtCY1O9WXzVmNKRU1TbWprWEk/view?usp=sharing

I think it's just the higher contrast - you can see the corners being a bit darker in the exact same places in the GOUT too, only it's just brighter and less contrasty over-all.

That might be the best way to do it actually.

OK, so this is something I'm very interested in - what exactly do I have to do in order to keep my projects in perfect sync with hairy hen's audio preservations?

Interesting - I have the same problem with my OT Bonus disc - it plays fine in my BD player but my computer's BD drive refuses to read it and I have the same problem with a couple more discs (among them the bonus disc in the Blade Runner 30th Anniversary set). I was planning on getting a different BD drive eventually (the one I have is external) hoping that it would solve the problem with those discs but it would seem that it may not.

Well many BDs come from the same master as their DVD counterparts but that doesn't mean they are the same thing - the resolution is the main difference between the two formats after all.

Like it or not, the policy is that if you want to download an HD restoration, you have to own the BDs and it makes sense because since the added resolution is the main value added for the BDs, acquiring the film in higher than DVD resolution some other way without also buying the official HD release is piracy. 

You should own the official BD in order to download the Despecialized or any restoration or fanedit based off of the BDs!!!

Thanks, Mike. And it's not so much about Despecialized being the #1 restoration available, which I never even claimed it was but about the fact that I tried to stick to those scans pretty religiously a waved them around as an argument whenever someone came up with some criticism of the colors in Despecialized and now it seems that I may well have been completely off base in some of those arguments.

So what you're saying is, that the colors in despecialized are essentially completely wrong? Because I remember I thought this may be the case and it may be an uncorrected scan, so I asked whether the colors in those scans looked close to the print itself when projected and you said they did.

How come that frame looks nothing like this shot does in the technicolor references I got from you?
There's way more green in those and the colors there are much more saturated and warm.

Looks like a light leak from the crawl element, so 1980 artifact would be my guess.

In despecialized v2.5, I still didn't remove all of the changes (I left in the new wipes and some of the less obvious re-composites) and the run-time of the shots I did despecialize still mounts to almost 20 minutes not counting the credits (which would add a few more minutes of the runtime still), so that's much more than 5%.

That is because the background in v2.0 is a still image from the SE. Also, the compression in the clip hides some of the grain in the cleaned up v2.5 version, of which there's actually plenty.

Plus I really don't see where any parts of the probe have "become thinner" in the filtered version.

Whereas in the v2.0, they certainly were thinner because of the way the probe and the smoke were composited from GOUT onto the SE HD background:

Thanks for the GIFs, g-force :-)

It seems from the references I've seen, that the Endor colors were actually a bit colder than they are in v1.0, so I will adjust them accordingly.

Still 720p for now but I'm doing the 35mm cleanup at 1080p for a bit of future-proofing.