The frame I am most curious about is the one where Luke is recovering inside the bacta tank with Han, Leia, Chewie and 3PO watching. That scene looks so different from one edition to another that I don’t know which one is the correct one.
This shot?
(Note this isn’t from the print, I just want to clarify which shot you mean.)
Is that if you try to donate without changing currency (e.g. Donating in USD if you are in the US)?
There has been a lot of work done, particularly in the retro gaming community to get an LCD TV to look like a CRT TV.
Check out some of the retro gaming forums on CRT emulation, there are some great AVISynth filters that really do recreate the look.
Okay, I’ll see if I can setup a US account, a $5 fee is crazy, it kind of negates the whole purpose of paypal, greedy buggers.
In the meantime, I can use a friends US account, PM me if you would like to donate that way. Thanks so much!
Now that we have 4k77, with reel-level correction, isn’t this all moot?
Not really. The 4K77 is great, but they didn’t sit and watch a projected print and then grade to match that print. The grade on 4K77 is somewhat arbitrary, but probably closer than anything else currently out there to how the film was.
As mentioned, this project is something different altogether. It is a recreation of a UK TV broadcast colouring.
Yeah I hear what you saying about memory, and I know the print is correct or accurate but there is something I can’t quite put my finger on.
And my thoughts lie with perhaps if it is indeed a broadcast print then how would that handle with a CRT and how would that also change the information.
The Blu ray when it comes to R2-D2 panels they won’t go the same color as the technicolor print in fact there is one shot that even if I desaturate the whole image his front part round his eye is blue and everything else is black and white.
So yeah the Blu-ray has some serious issues…
I am just pondering these variations and It might only be small, but it is signifficant enough that it’s coming over like there is something not 100% correct but say 98% correct and that’s not to say I am calling it out as wrong but it does seem that there is evidence to support that is displayed at least a certain way and that perhaps should be taken into account.
I mean I’m not bonkers but I think action is very important part of the film and It should not be overlooked even though it is a small thing ultimately.
Basically they were yellow flashes until the DVD came out in 2004… And I still had a CRT so I am wondering if it has to do with this or that as you say it’s been prepared by design for CRT or the CRT simply misinterprets highlights… But since the new 2004 it has lost the luminence that it did have in my opinion.
Either way you can’t erase 20 years of something being a certain way even if it’s wrong…
The Blu-ray is an absolute mess when it comes to colour, and they digitally altered specific parts, like some of R2’s panels by masking them out and changing the colour separately etc. so it is useless as far as a cohesive colour reference goes. Film scans, laserdisc and VHS will at least be consistent within their own medium, whereas the blu-ray has been tweaked within frames.
As for CRT, it is highly unlikely that your particular television(s) were ever calibrated, so who knows what colour bias they had, generally though a calibrated CRT TV will give accurate colour.
The explosions likely had a bit more yellow bias, telecine tends to lean towards the warmer end, and an operator again would likely decide that an explosion should be orangey-red and colour correct towards that end.
One thing CRT does have is an interesting luminance curve, and absolute blacks and a diffusion due to the spot size being larger than the smallest pixel element. This gives a diffuse glow to everything, which coupled with the blacks and scanlines, can give an impression of more colour depth, and a less flat image.
Watch a laserdisc on a CRT TV vs on a LCD screen and you will see what I mean. It is a different look, and the loss of luminence you are talking about might be to do with that.
Anyway, you are going down the path of creating a version as you remember it, whether that’s how it was or not is pretty irrelevant for that kind of project, colour it the way it looks best to you.
I don’t know yet as I haven’t watched this print projected. LPP from 1983 tends to be grainier and lacking shadow detail, so hopefully this Eastman print will be much better, but I will know more once it has been cleaned and scanned, then I will risk projecting it.
The LPP was made a few years later, so may not be the same colour-wise as the 1980 release, but we will soon know for sure.
It certainly wouldn’t have been a ‘raw’ take, but they may have gone back to the neg and made a new print from it for the purposes of a TV version to do the telecine from, and then the grade would have been different, and then the telecine process, as mentioned, would have had an on-the-fly grade done to it as well.
I’ll get the broadcast tapes captured and see how they look. It will be 4:3 of course!
Looking at that VHS cap, the magenta is still definitely there in the flashes and explosions, so it looks like the same film master to me, but it is clear that that version has had the magenta graded way down, looking at the extremely yellow skintones. So the flashes are still there, but the whole thing now has a very yellow push.
You can also see the ‘neutral balance’ that is typical of a telecine transfer, in the grey moon and very white ship, telecine operators tend to go for neutral colouring, even when the original might have had an intentional colour shift (a scene being blue-ish to convey cold for instance, they will often re-balance to be a neutral grey).
Also, memory for certain colours is terrible, even among trained colourists. In colour science there are a range of what is known as memory colours, it applies to things like sky, grass, fire and explosions.
We have a tendency as a species to remember these items as particular colour palletes, even if what we saw were quite different. e.g. We remember a STOP sign from a movie being blood red, even if it was actually quite orange, or even almost grey. We will clearly remember a sunset as being orangey-red, even if it wasn’t at the time (it may have been, but even if it wasn’t, we will remember it as being reddish orange).
If anyone wants to read up on it, and why we are so incredibly bad at remembering colours for certain items, I’ll pop a reference paper below.
It doesn’t mean anyone shouldn’t re-colour a film to how they remember it, but it is interesting that how they remember certain parts of it, skintones, sky, grass, coffee, taxis, water etc. will alomost certainly be way off from how it actually was in the film.
Our memory, and expectations make us see and remember colours that aren’t there, similar to the way that we ‘know’ that squares A& B are not the same colour, so our brain ensure we see them as different colours, even though sqaures A & B are absolutely identical (as shown in the image at the end of the post.
A less technical, but still interesting read: https://prolost.com/blog/2010/2/15/memory-colors.html
More technical reading.
https://www.medicaldaily.com/memory-color-shades-why-human-brain-struggles-remember-color-336396
http://psycnet.apa.org/record/2015-22398-001
Bodrogi, Peter & Tarczali, Tünde. (2001). Colour memory for various sky, skin, and plant colours: Effect of the image context. Color Research & Application. 26. 278 - 289. 10.1002/col.1034. In memory-matching techniques, the remembered colour might differ from the original colour even if the viewing situation is the same. Our aim was to point out whether these so-called memory shifts are significant in the everyday situations of viewing photos depicting sky, skin, or plant, or viewing standalone uniform colour patches of sky, skin, or plant colours. In many cases, significant memory shifts have been found. Considering only one type of object (sky or skin or plant), memory shifts turned out to be systematic in the sense that they were directed toward specific intervals of hue, chroma, and lightness. This tendency was more explicit for photos than for standalone colour patches. A method to quantify prototypical colours and their tolerance bounds was suggested. © 2001 John Wiley & Sons, Inc. Col Res Appl, 26, 278–289, 2001
Block out the little ‘bridge’ joining squares A & B with your finger, and your brain will go back to seeing them as different colours again. Our expectation of colour really affects our perception and memory of it.
Erm, not really.
It involves boosting the cyan mostly, but in very specific ways. Film doesn’t fade evenly, the highlights end up with the most fade, so it is a matter of looking at the fade across each layer of film (CMY) and working out how much it has faded at each level and in each channel.
Having a good colour reference is essential for this, then you can do some fun computations based of probability density functions to get the two sources to match, and then use the results from that to correct the rest of the reel.
It is based on this technique http://www.mudgee.net/ot/colourgrading.pdf
The basics are here: https://originaltrilogy.com/topic/Colour-matching-for-fan-edits/id/7257/page/1
Dre’s tool originally started off using this technique and then he went a lot further, but the images I’ve used on the front page of my other thread, where the very faded ‘red’ German print is adjusted to the less faded print are using the code from the links above, and then some more correcting in Nucoda.
The images on the front page of this thread are from the other print I had access to which was only very slightly faded, but unfortunately had a lot of mold on a lot of frames. That print will be used for any frames that are damaged/missing on the new print.
Hope that helps!
No my screen is fine… Let me explain. Honestly I find it interesting no hijacking.
This is the version that was on TV I do remember it vividly. I recorded this on Betamax and I just used to watch it over and over again and If I could not watch it on Betamax then I used to have an old maxwell 90 audio tape with the audio of the film on It I used to take about with a Battery powered tape deck.
So I guess what I am trying to say is this is the version I remember vividly but it was prepared a little different but it’s for want of a better word the same thing.
But the levels on the clips look like they are different. and not the right contrast, contrast much higher.
What I have been trying to do is re-create the look I remember off the TV, that version I remember. But I must admit The whole twin suns is an oddity and I am more and more sure there are 2 versions of the scene now.
Now I seen the next part where they are running this part looks fine… Seems like it has brightness issues, if you compare it to the bit where R2-D2 cross the corridor this part is perfect.
On dealing with that part in the Blu-ray was having to drop the high levels to near zero on one shot, it’s over accentuated the blu-ray or 2004. So it’s more of an inherrant problem for that part than a problem with your footage. Not your fault but it’s salvageable from what you have not in the Blu-ray though.
There is a version where the sparks around the Door are yellow and the flashes are yellow but it’s more de-saturated and this was what was on UK TV back in the day.
Ah, so a recording of the UK television broadcast that you recorded onto Betamax?
I have the U-Matic broadcast tapes here that the television stations used. I must get around to capturing them one day, it would be interesting to see what the colour is like on them.
Thanks, first thing that struck me is how un-naturally over bright this is, I mean seriously I need to put sunglasses on.
The color is very good but there is no depth it looks flat as a pancake.
It looks like a series of really bright images. It does not have any step down like a levels curve it’s just all very bright as can be.
So if this was projected as I said it would drop in the mids then further in the shadow. But I seriously remember the yellow flashes not dreaming it up or anything it’s on the VHS 😃
Would you say it’s missing yellow?
The medium this has been prepared for is an old 35mm projector not a tv.
Should have called this thread hello yellow perhaps?
If I de-saturate that drop the brightness a bit up the contrast (Big jump) and add a bit yellow that is basically the VHS…
It might be worth a check you monitor if that clip looks un-naturally bright.
The levels on the scopes are correct, but that clip is at broadcast levels, so blacks start (correctly) at 128, not at zero, and it stays way under peak brightness levels - so it depends on your monitor calibration, and what software you are using to play it back.
Here is a test-pattern clip set to similar levels, check how the test patterns look. The second bar on the first pattern should be almost black for video level playback.
https://wetransfer.com/downloads/f22767e2f8d92ad1ac1a6de77f48f7df20180602010248/292ea0ca385130b7d0e0b724bf51707320180602010248/a95e73
Modern movies are way more contrasty than older movies, the current trend is to crush the blacks and create a very ‘punchy’ look. It often has way less actual detail and nuance than the older films, but people are now very used to a ‘punchy’ image.
Anyway, I’ll stop hijacking your thread, I just thought I’d drop in to show what the original film pallete was back in the day, and remind anyone doing any colour work to calibrate their screens and remember the difference between video levels and full range computer levels.
EDIT on re-reading your post, I see you are probably talking about that photo of the sunset looking overly bright, yes it is!
Yes, on the original prints you can see that the frames are intentionally graded that way by looking at the frame edges.
I haven’t watched the Blu-Ray, so can’t comment.
Below is footage from a UK release print.
https://we.tl/PDf9j9SQ4r (42MB .mp4 file)
By all means, colour the film to the way you most enjoy it 😃
So you are telling me that loads of pink and magenta flashes and stuff are intentional?
Yes, there absolutely are pink/magenta flashes during laser blasts etc. on all surviving prints, and it was that way when I screened it many, many times back in 1977 and again in 1980 and again in 1982.
That is how it always was in the cinema. Most people wouldn’t notice it as it is one or two frames, i.e. often just 1/24th of a second, but it is how it was.
Anyone is free to re-colour the films to a version they prefer of course, but if you are interested in how it was in cinemas at the time, the blasts did have magenta flashes in many of the ‘explosions’.
Ah ok interesting… So although a print may have all the detail it might not have the correct exposure?
Hence why a laserdisc has more detail than a print because of being over exposed? or blown in bright part…
So like what I am questioning here about although prints may have more detail over all… They may miss out on certain details depending on how it is handled during exposure?
No, it is not that the print has the ‘wrong’ exposre, the print has the exposure desired by the Director.
The negative will have a lot more shadow and highlight detail than the print.
The print colour and exposure is intentionally different to the negative, this is what grading is - the job of the colourist and the director to decide how the colour and exposure etc. should look on the final print.
So a decision might be made on a particular shot to crush the blacks a bit to set a particular mood, or indeed to hide unwanted detail in the shadows. The director might decide to blow out the highlights to make a scene look more bleak, or an explosion to look more intense etc.
When doing a telecine for home video, often the Director isn’t there for the process, but even if they are, the telecine operator has relatively little control over the exposure and colour choices, as it was adjusted as the film was running.
Also, for home video, decisions were often made, because of the limited colour space of NTSC to wind the highlights back, so a telecine transfer is always going to be a different beast to the release prints.
It isn’t that the film printwasn’t correctly exposed, it was a conscious choice to achieve a particular look, and a different look was often chosen for home video, where the viewer used to be watching on a 12" to 27" 4:3 CRT screen, possibly in black and white, and probably with the living room lights turned on. This makes for a very different decision making process for how the images should look, compared to a dark cinema with a very large screen.
None of the laserdiscs are anything much like the movie was in 1977.
When the D2 tapes were created, the film elements used were run through a telecine, with an operator/colourist whose main job was to get the colours within NTSC legal, and would have balanced the shots by eye on the fly in real time.
So the colour on any of the home versions was whatever decisions the telecine operator made at the time, many years after the original release.
The Technicolor IB prints will not have measurable fade, and certainly won’t have lost any real detail in the highlights vs what was there in 1977. The IB prints were struck the same year as Star Wars was released, so the original elements would not have faded either.
However, if a new print was made for the purpose of the telecine at the time, that print would have had different exposure etc. and may have picked up some detail from the negative that was never in the release prints, such as in the explosions.
It gives some insight onto what may have been on the neg, but not into what was on the prints screened in 1977.
Funnily enough, a DeLorean was genuinely involved.
Beautiful print. I hope you store it during and after the preservation in the same cold conditions. 😃
Yes, it is currently in cold storage at the Academy in LA until I can get it cleaned.
It spent its life from 1985 til 1993 in archival storage, it would have been kept in ideal conditions, at about 7C, it then went to a collector for a couple of years, and then went into cold storage until now.
It is a very lucky (and stupidly expensive) find.
The teaser for ‘Revenge’ is this one:
https://movieweb.com/star-wars-revenge-of-jedi-teaser-trailer-rare/
Ah, it would make sense to download the 1080P version then 😃
Anyone figure out a definitive way to fix the jerky/choppy playback? I’m trying to play the file on my Apple TV 4K via Plex. I’ve tried remuxing it with both MKVToolNix and MP4Tools. Neither worked…
I’m streaming from my iMac to my 4th generation Apple TV via Plex without issue… I can take a look at my Plex settings when I get home if you’d like to compare.
Out of curiosity, it wouldn’t be an issue with your router, would it? I used to experience notable buffering times when streaming large movie files or streaming trailers via the Apple Trailers app before upgrading my router.
Does your TV do 24P? It works on my Apple TV @ 24Hz smoothly.
Harmy’s post earlier about the comparison between the 4k and the 1080 being only more detailed in the dirt and scratches had me digging around for resolution information. I know for example that the theoretical information contained in a 70mm film frame is around 6K. So, the question is, how much resolvable information is actually in a film print, like this 35mm IB tech print? What I found is a suggestion that the negative is very high resolution and then each subsequent generation from the negative (interpositive, print) loses resolution and therefore information. This would be like taping a tape of a tape - analog loses signal over generations. So, 4k might actually be overkill for this print in terms of resolution, but a UHD version of the 4K could really show off the color space where 1080 just can’t. Anyway, just thought I would post the relevant link, etc. https://www.quora.com/Does-4K-resolution-approach-the-resolution-of-a-35mm-film-print
By the way, I love this print. Projected it last night…
The answer is… it depends.
I have had 35mm prints where you need to scan in 10K to capture all of the resolution on the print. I have had others where 2.5K is overkill, the resolution is just not there on the print.
Yes, negatives hold way more colour and resolution and dynamic range than a release print, which has often done through multiple generation losses along the way (Some prints come straight from the negative, and with digital you can go straight to a print directly) , so it really does depend a lot on what you have.
With the main IB print that was used for the 4K77 restoration, it was scanned on a 4K Scanity, but the horizontal resolution on the scan was less than 4K as it included the soundtrack and sprocket areas. Also the Scanity uses a line sensor, so mechanical tolerance issues in its drive mechanism lead to lower resolution captures than on full frame sensor scanners. The Scanity is designed for negs, not prints, so it also loses out a bit in shadow detail and colour capture vs more modern scanners.
So in this case, there probably isn’t much difference between 2.5K and 4K for the scan that is being used, but having the scan in 4K allows better granularity when stabilising and doing repairs, but the end result will basically look identical if you rendered it out to 1080P, and then took that 1080P result and upscaled it to UHD.
Yes, a wider colour gamut than Rec709 gives some advantages if you use the wider colour space that the newer UHD formats offer.
It will be a combination of the two prints, with this new print being the main source.
There are bound to be some missing or damaged frames that are better on one print than the other.
Yes Laserschwert, the flare on the left hand side is due to some smudging on the lens, the lens was dirty, the scanner had just run a print that had been treated with filmguard, and some ended up on the lens. It meant that print had to be taken off and cleaned, so it gave me a ten minute window on the scanner, we just chucked it on quickly to see how it looked.
Arnied, cleaning and prep is USD145 per reel, so $870 (and freight).
CatBus, the print was put into cold storage nearly 25 years ago, so it has the slightest amount of fade, but it is really slight. It means we finally have a truly reliable colour reference for Empire as it was in the cinema. It is very exciting.
Remember that the cinema presentation is more cropped than what is on the full frame of the film.
Here are some larger files for the pixel peepers.
The scanner wasn’t calibrated or refocused, just sneaked in a few frames in-between other jobs.
Same three frames, just higher res and not scoped.
It is just so good to have an actual colour reference to work on. I’ll be able to project any restoration side by side with the print to check the colour.