logo Sign In

Info: The Matrix - with original theatrical color timing? — Page 5

Author
Time

I like the results shown in the test clip. It seems to me that the green beard really stood out but that could just be because you mentioned it. Can't be sure I'd have noticed otherwise.

George creates Star Wars.
Star Wars creates fans.
George destroys Star Wars.
Fans destroy George.
Fans create Star Wars.

Author
Time

I am actually impressed by those results. The green beard was also evident even in the original colour timing of the first DVD release as well, so it doesn't really both me personally.

Author
Time

_,,,^..^,,,_ said:

Using one of these two methods (ColourLike or BDluma+DVDchroma) it's possible to retain the BD resolution with DVD color timing; now, if the DVD color timing is right, this methods could be used to do a proper restoration.

The problem with both these techniques is that the sources are so different in both luminance and color, their selective recombination still suffers from the Blu-ray's foibles. Getting luminance is not the same thing as getting resolution.

Fortunately, there is an alternative -- color correction. As long as this was not some George Lucas re-imagined "director's cut" hi-def, a single correction might be usable over the entire movie. Worse, some scene-by-scene exceptions. Anyway, here's a very quick proof-of-concept going by the numbers (of the histogram graphs):

This is what needs to be changed -- everything except the resolution. :)

This DVD is the target -- an end result, only in higher resolution. Notice the RGB graphs. These are what is to be duplicated on the Blu-ray. By eye, it is easy to use histogram functions to squeeze down and slide around each Blu-ray graph to roughly match it's DVD counterpart:

HISTOGRAM
RED       Low 16  Gamma 1.2  Midtones +25
GREEN   Low 16  Gamma 1.2  Midtones +25
BLUE      Low 12  Gamma 1.3  Midtones +20

The original Blu-ray is in grey and the red overlay is the result of spectrum squeeze-down ("Midtone: compress") and slide-around ("Low" and "Gamma"). Done (except for fine tuning the numbers -- always use an eye-dropper for dispassionate precision)!

Of course, if you wanted to repair any Blu-ray crush or blow-out, those areas would be patched in from the non-crushed, non-blown-out DVD itself.

Author
Time

Looks great spaceranger.

I need to get better at colour correction.

Author
Time

Spaced Ranger, could you elaborate an avisynth script for your method?

I'm testing further, and here you go the last results:

Tomorrow will be "The Matrix recoloration day...

Sadly my projects are lost due to an HDD crash… 😦 | [Fundamental Collection] thread | blog.spoRv.com | fan preservation forum: fanres.com

Author
Time

Truly amazing work here!

Have you improved on Neo's green beard or is it still unavoidable for some shots?

Author
Time

Andrea: Which method is `ColourMatch'? It looks the best IMO.

A picture is worth a thousand words. Post 102 is worth more.

I’m late to the party, but I think this is the best song. Enjoy!

—Teams Jetrell Fo 1, Jetrell Fo 2, and Jetrell Fo 3

Author
Time

AntcuFaalb said:

Andrea: Which method is `ColourMatch'? It looks the best IMO.

Agreed. Looks good.

Author
Time

Agree with ColourMatch looking the best. Just a little too much red though in every shot. Is there any way to globally bring the red down slightly?

Also, out of interest, all the shots you've posted are scenes set within 'the matrix' which are the ones affected by the BR's green tint. Do you have any examples of what this process would do to the 'real world' scenes since the BR timing for those is pretty much the same as the DVD to start with?

George creates Star Wars.
Star Wars creates fans.
George destroys Star Wars.
Fans destroy George.
Fans create Star Wars.

Author
Time

This is looking really good guys. I always kept my DVD and refused to buy into the screwy BD version. If we can manage to get the green out, then address the contrast, all that remains is the overall DVD color which is more of a dull red/yellow and may not be accurate. It's an older master for LD/early DVD and many of these had altered or differing color timings. (Blade Runner, The Wild Bunch etc.)

VADER!? WHERE THE HELL IS MY MOCHA LATTE? -Palpy on a very bad day.
“George didn’t think there was any future in dead Han toys.”-Harrison Ford
YT channel:
https://www.youtube.com/c/DamnFoolIdealisticCrusader

Author
Time

ColourMatch is simply a "codename" for my messing tests on ColourLike mixed with RGBMatch... tomorrow I'll do further tests, and, as it seems results are quite good, I think I could start a restoration program - what do you all think? I'll test it even on "The Thing" to see how it looks...

Here you are the "green bearded Neo" and two "out of the Matrix" live scenes - this time I added a ColourMatch version with 0.75 saturation, and it seems the best compromise:

(this is going to be the tallest forum thread ever made...)

Sadly my projects are lost due to an HDD crash… 😦 | [Fundamental Collection] thread | blog.spoRv.com | fan preservation forum: fanres.com

Author
Time

Great Work.

All I can say is its the best ive seen so far in trying to re-create the original experience.

Coupled with the Theatre DTS this restoration would be a dream come true.

Are you planning on doing a release, or supplying an Avisynth Script of some kind?

Thanks for attempting this.

Author
Time

This is looking really great! I saw this in theaters twice with my dad when it came out and the Blu-Ray colors really annoy me. My dad even complained about Neo's green beard the last time I lent him the BD.

Author
Time

_,,,^..^,,,_ said:

Spaced Ranger, could you elaborate an avisynth script for your method?

I never found an Avisynth plug-in that accepts a paint program's histogram functions numbers. However, in the Song Of The South thread, I noted that VirtualDub with a plug-in (Gradation Curves v.1.46b by BBugBunny, A.K.A. Alexander Nagiller) allows for a manual reconstruction of the histogram graph:

http://originaltrilogy.com/forum/topic.cfm/Song-Of-The-South/post/641915/#TopicPost641915

I know it's cludgy but that's the best I could find.

Author
Time

Spaced Ranger said:

I never found an Avisynth plug-in that accepts a paint program's histogram functions numbers. However, in the Song Of The South thread, I noted that VirtualDub with a plug-in (Gradation Curves v.1.46b by BBugBunny, A.K.A. Alexander Nagiller) allows for a manual reconstruction of the histogram graph:

http://originaltrilogy.com/forum/topic.cfm/Song-Of-The-South/post/641915/#TopicPost641915

I know it's cludgy but that's the best I could find.

IIRC, I found that script where I was following some of your posts (can't remember in which thread, maybe THX1138?), and I used it with my SW:ANH '97... it was really hard to achieve a good result, took a long time and a lot of tests, and I was not really satisfied with it... hope the new tests with this ColourMatch method will be even better than the current ones.

Nien Nunb said:

My dad even complained about Neo's green beard the last time I lent him the BD.

... at one moment, I thought to be crazy, but now that someone else noted it immediately, there are two options:

  1. me and your dad are good observers
  2. me and your dad are mad about wrong details

at least now I'm not alone anymore... (^^,)

***   ***   ***   ***   ***   ***   ***

About a preservation project: it seems there is some interest showed in this thread, and I definitely like the idea to do it, along with theatrical DTS soundtrack. Leave me one day to test further, post new screenshots, and if they will be better than these ones, BAM! "The Matrix [spoRv]" could be reality! Should I dare to do it?

Sadly my projects are lost due to an HDD crash… 😦 | [Fundamental Collection] thread | blog.spoRv.com | fan preservation forum: fanres.com

Author
Time

_,,,^..^,,,_ said:

Spaced Ranger said:

I never found an Avisynth plug-in that accepts a paint program's histogram functions numbers. However, in the Song Of The South thread, I noted that VirtualDub with a plug-in (Gradation Curves v.1.46b by BBugBunny, A.K.A. Alexander Nagiller) allows for a manual reconstruction of the histogram graph:

http://originaltrilogy.com/forum/topic.cfm/Song-Of-The-South/post/641915/#TopicPost641915

I know it's cludgy but that's the best I could find.

IIRC, I found that script where I was following some of your posts (can't remember in which thread, maybe THX1138?), and I used it with my SW:ANH '97... it was really hard to achieve a good result, took a long time and a lot of tests, and I was not really satisfied with it... hope the new tests with this ColourMatch method will be even better than the current ones.

Nien Nunb said:

My dad even complained about Neo's green beard the last time I lent him the BD.

... at one moment, I thought to be crazy, but now that someone else noted it immediately, there are two options:

  1. me and your dad are good observers
  2. me and your dad are mad about wrong details

at least now I'm not alone anymore... (^^,)

***   ***   ***   ***   ***   ***   ***

About a preservation project: it seems there is some interest showed in this thread, and I definitely like the idea to do it, along with theatrical DTS soundtrack. Leave me one day to test further, post new screenshots, and if they will be better than these ones, BAM! "The Matrix [spoRv]" could be reality! Should I dare to do it?

[spoRv]?

 

Author
Time

Andrea,

Since the DTS audio is already completed for this movie we could get it to you once you're ready to proceed.

:)

 

Author
Time

stretch009 said:

[spoRv]?

I like acronims... for the moment, it stands for Special Preservation Of Remixed Versions - as it will take BD for luma, DVD for chroma reference, and DTS theatrical soundtrack (Jetrell, you got a P.M.)

I'm waiting to receive the HD-DVD video for "The Thing [spoRv]" where it will be used for luma, LD for chroma (maybe merged with the PAL DVD), and different DVDs and LDs for soundtracks...

I know it's not the best one, and I'm open to new acronims; any idea? It must mean something about preservation and/or restoration, possibly from different sources... I've got an idea into my mind about SPORV, but I'll reveal more about it in the next days...

Sadly my projects are lost due to an HDD crash… 😦 | [Fundamental Collection] thread | blog.spoRv.com | fan preservation forum: fanres.com

Author
Time

_,,,^..^,,,_ said:

stretch009 said:

[spoRv]?

DTS theatrical soundtrack (Jetrell, you got a P.M.)

You have a response ......  ;)

Author
Time

"The Matrix recolor day" is almost finished, and I've made a lot of tests... we all know that the BD colors are wrong, and DVD colors are better... but are they "right"? After I spent all the night trying to reach the closest colors to DVD, I realize that in many scenes they are wrong too (IMHO), in particular, the "off the matrix" scenes were way too dark and blue/violet, while often "in the matrix" faces were too "sunburnt", and sometimes still yet a bit greenish...

Then, I thought to give a chance to an old script I made during the color correction of my SW:ANH (SET ruLes); at that time, it worked not so well, but now I have fixed it; what this script make is trying to match two clips and extract the right colors. I think the results you can see below are difficult to replicate using another tecnique with only one color correction setting for the whole movie... obviously, this is far from perfection, but a good compromise, retaining the most "character" of the DVD colors, while correcting the worst scenes. It worked in 99% of the scenes, and failed only when a super bright scene is present (like in the scene where Morpheus explains to Neo what is the Matrix)... if this method - called ChromaRite, just to remeber it when I will use it again - is, well, "right" enough, I will start "The Matrix" preservation project, using it for the whole movie, apart the 1% when it doesn't work - there I'll use the DVD uspcaled or another color correction script.

At the contrary, if you don't like it, I could always use the ColourMatch method; I refined it in the morning, and now in more than 90% of the scenes, colors are almost exactly the same of the DVD, while the other 10% are close enough. I have not posted ColourMatch screenshots here, as I personally think ChromaRite method is better, but it's up to you. Choose your pill... (^^,)

I've made a sample clip - inside, there are some seconds from the DVD, upscaled, where ChromaRite failed - the eventual project video quality will be the same.

Sample Clip (40sec. 14Mbps, 69.4MB): http://www.sendspace.com/file/l8uvt5

Waiting for your comments as usual.

Sadly my projects are lost due to an HDD crash… 😦 | [Fundamental Collection] thread | blog.spoRv.com | fan preservation forum: fanres.com

Author
Time
 (Edited)

Interesting shots, for me it's hard to say which is best or correct, my only sources for the original color timing are my pal vhs tape and the hdtv rip i got from rutracker. Would the hdtv rip be a better reference than the dvd in this case or not?

Author
Time

This is just my view and this is, of course, your project but I'd much prefer the  regrading be based on reference material than on an opinion of what looks right or wrong. Perhaps the DVD may not match the original theatrical colors exactly but without a reference it all becomes subjective and at worse risks becoming just another 'revisionist' version of the kind that put many of us off the BR in the first place.

In short, I'd take the ColourMatch pill without hesitation. :)

George creates Star Wars.
Star Wars creates fans.
George destroys Star Wars.
Fans destroy George.
Fans create Star Wars.

Author
Time

Unfortunately, I think I'll take the Colourmatch pill as well. It seems to be producing results more accurate to the DVD.

Kingherb said:

Would the hdtv rip be a better reference than the dvd in this case or not?

No, it's the same master as the DVD, so there shouldn't be any difference in color.

The Academy Awards Screener tape I have strongly resembles the DVD's pallette and I'm pretty sure it's a different telecine. It's pretty low quality though, I have to say.

Author
Time

A question that doesn't seem to be answered: what is the sample size being used on colourlike?

The program works by taking a 'read' of a sample clip, averaging what the frames are like, arriving at a single answer and applying it to the full range of the replacement video.

Are the screenshots a sample of 1 frame applied to 1 frame?  The entire scene?  Large chunks?  The entire DVD (which shouldn't work because the color timing is so different between scenes)?

Just single shots like this don't really prove anything.

Dr. M

Author
Time

I can't help but think you must have done something wrong when you tried ColourLike. (Maybe you didn't run it in RGB32?) I tried a couple of the latest shots and it matches them very nicely: