Color Correcting the Dragon Box - 3 Part Spectacular

[lansing]

I don't want to go into discussion about if there're macroblocking in the Dragon box, I don't care, I just want the contributions here to stay on topic.

You are really rude with him, he is the only one that make this topic alive since months, all people with genius idea mostly left their project out of anything or just don't talk about them anymore, i am okay that his methods is not yet ready and it's mostly why i didn't answered for my part except for some kind of encouragement. it's really nice that unlike 99% of all people that show their results only (like a "yes i am strong, i can do that but you can't" and never helped others) he kindly shared what he are doing and how he can do it. the only point where i can agree with you is that he shouldn't have made his tutorial until he reach his final goal as we can loose ourself in all the steps (and it's not a way i usually follow, i am a noob using his way i don't really understand lol) but there's others way to said him

as long as he work about color corrections it's in the topic, if some step don't fit you then just don't follow them that's all. i remember that you said nothing with people that was showing ugly waifux2 things that are farrrrrrrrrr worst than any issue i have seen here

Try and read his guide, 90% of that has nothing to do with color correction, that is the main problem. I am perfectly fine with people going off topic once every while, but not this. He has been going off on that irrelevant guide for close to a month!

What final result did you managed to get? show yourself where he is wrong with example, screenshot if you really think he is wrong, i think it's the best way to give evidence and make things go forwards instead of just a useless and stupid opinion fight and why not helping him too?

What he should be working on is his color matching method, not dithering or deblocking or whatever irrelevant stuff he was doing.

So far this is what he was up to from a couple of page back:

This is kai screenshot:

Spoiler:

This is his result matching to the kai shot:

Spoiler:

Obviously there is still a long way to go on the accuracy.

[crabshank1]

For the (n+1)th time: the macroblocks were coloured and fucking up my white balance (my CC), so I had to lose them even if I didn't hate them. But seeing as you said white balance isn't needed, I finally realised earlier why you keep saying this.

you talk about coloured pixels in the white colors? i used to see some when i tried to cc, the only way i managed to find to removed them was DNR, i was thinking that it was caused of the grain lol

In the dark greyscale colours mostly, like in the hair.

[crabshank1]

What he should be working on is his color matching method, not dithering or deblocking or whatever irrelevant stuff he was doing.

So far this is what he was up to from a couple of page back:

This is kai screenshot:

Spoiler:

This is his result matching to the kai shot:

Spoiler:

Obviously there is still a long way to go on the accuracy.

I was not matching to Kai colours, but Kai HUES (HSV)!

Also, I find it strange that you haven't mentioned my work on Dragon Ball E109.

[lansing]

What he should be working on is his color matching method, not dithering or deblocking or whatever irrelevant stuff he was doing.

So far this is what he was up to from a couple of page back:

This is kai screenshot:

Spoiler:

This is his result matching to the kai shot:

Spoiler:

Obviously there is still a long way to go on the accuracy.

I was not matching to Kai colours, but Kai HUES (HSV)!

Also, I find it strange that you haven't mentioned my work on Dragon Ball E109.

Any kind of color correction will look better than the original for such bad source, it didn't prove anything. What I'm looking for is a reference and the matching result of the same frame so we can compare the color matching accuracy. What is the reference for your ep.109 result?

[crabshank1]

Any kind of color correction will look better than the original for such bad source, it didn't prove anything. What I'm looking for is a reference and the matching result of the same frame so we can compare the color matching accuracy. What is the reference for your ep.109 result?

The reference hues were from Hikari TV, pictures on the internet and educated guesswork. That's what you have to do when there's no good reference.

This is an old problem, no one knows the true colours, but I think matching hues is a good and feasible solution.

Why don't you post some of your work, so we can compare results?

[crabshank1]

What he should be working on is his color matching method, not dithering or deblocking or whatever irrelevant stuff he was doing.

So far this is what he was up to from a couple of page back:

This is kai screenshot:

Spoiler:

This is his result matching to the kai shot:

Spoiler:

Obviously there is still a long way to go on the accuracy.

I was not matching to Kai colours, but Kai HUES (HSV)!

Also, I find it strange that you haven't mentioned my work on Dragon Ball E109.

Any kind of color correction will look better than the original for such bad source, it didn't prove anything. What I'm looking for is a reference and the matching result of the same frame so we can compare the color matching accuracy. What is the reference for your ep.109 result?

For exact colour matching I have this:

https://github.com/crabshank/HTML5-medi ... r-webpages
https://chrome.google.com/webstore/deta ... edjhbjceli

I uploaded this colour bars image:


to Youtube and got (probably Rec709):


then with a colour matrix I got:


But I don't think the colour matrix method is suitable for this project.

[lansing]

This is an old problem, no one knows the true colours, but I think matching hues is a good and feasible solution.

Well that is not what your matching result is telling me on that match to the kai shot. Your method needs a lot more works.

Why don't you post some of your work, so we can compare results?

Start reading from about 15 pages back on this thread and you'll see what I was doing. As of right now the result of your color matching work is no better than the selective color Ajay was doing in After Effect 3 years ago.

[HakkaiBills93]

This is an old problem, no one knows the true colours, but I think matching hues is a good and feasible solution.

Well that is not what your matching result is telling me on that match to the kai shot. Your method needs a lot more works.

Why don't you post some of your work, so we can compare results?

Start reading from about 15 pages back on this thread and you'll see what I was doing. As of right now the result of your color matching work is no better than the selective color Ajay was doing in After Effect 3 years ago.

color match tools is not perfect either, when you create a lut it don't take little area colors and even when you managed to get an accurate result in color match tools and export it as a lut, the result isn't the same in the software in which you load it
for example i took this cel

you can perfectly reach the rights colors only by using the rights settings, it's not cause some ajay attempt were wrong that you can say that it is not possible to reach it with after effect (it's not cause you can't do something that it is impossible) here some example



here the frame from dbox

about the namek sky little more yellowish in the cel and not in the cc i did it's just that i forgot to settle it, selective colours allow you to alter colors easilly, there's plenty of way to color corrected and color matching using season sets for example will give you crushed white and black (and i already show you that cause no in no way the white was supposed to be crushed like in the season sets and it's mostly cause they overcontrasted the release that it look like white was balanced but it's wrong)

some db episodes like 109 in the dbox don't seems to have tint in the white only cause it was crushed the same way



here the dbox with color match tools



here what i reach with after effect

[lansing]

Start reading from about 15 pages back on this thread and you'll see what I was doing. As of right now the result of your color matching work is no better than the selective color Ajay was doing in After Effect 3 years ago.

color match tools is not perfect either, when you create a lut it don't take little area colors and even when you managed to get an accurate result in color match tools and export it as a lut, the result isn't the same in the software in which you load it

you can perfectly reach the rights colors only by using the rights settings, it's not cause some ajay attempt were wrong that you can say that it is not possible to reach it with after effect (it's not cause you can't do something that it is impossible) here some example

You probably doing something wrong if you're getting that kind of result with dr dre's color matching tool.

This is what I got using dr. dre's color matching tool and 3d lut creator, both giving excellent results.


dr. dre's program:

Spoiler:

3d lut creator:

Spoiler:

selective colours allow you to alter colors easilly

No, manually adjusting every color in selective color in After Effect is extremely time consuming, and it will be a nightmare to update a color if something goes wrong. Let's say you have done 50 episodes in the line and suddenly realized that you've been using the wrong color for goku's skin, then you're fucking screwed because you would have to go back to every single episode to fix them. Also this method is not reproduceable by others and undistributable meaning you can't share your work with anybody.

[crabshank1]

Just a heads up to anyone interested, I need to make a lot of updates to the guide already. However, I won't do it yet as I'm still working on it. If you are interested in the technicalities, I have updated my Github

Preview:
FFT3Dgpu(sigma=10) applied to my CC on the left, dBox on the right

[crabshank1]

Updated chain: Upscale (NO LONGER OPTIONAL!) -> hues -> grey dither (https://github.com/crabshank/HLSL-Resha ... ither.hlsl) -> white balance -> crush (OPTIONAL) -> downscale.

More to come soon.

P.S. I have updated the AvsPmod.zip file https://1drv.ms/u/s!AjHPdv1_75reiQwQwkH ... 9?e=Z5gZKn
Backup: http://s000.tinyupload.com/index.php?fi ... 6787188118

P.P.S. I recommend Just Color Picker http://annystudio.com/software/colorpicker/#download, the only eyedropper I know that lets you take a the average of a sample. 3x3, 5x5, or 7x7, it's slow but works in real time.

[PremiumSalt]

Just a heads up to anyone interested, I need to make a lot of updates to the guide already. However, I won't do it yet as I'm still working on it. If you are interested in the technicalities, I have updated my Github

Preview:
FFT3Dgpu(sigma=10) applied to my CC on the left, dBox on the right

Correct me if I'm wrong but isn't the left a wildly inaccurate depiction of Boo's color? I was under the impression that what's seen on the right is what he's supposed to look like.

[crabshank1]

Just a heads up to anyone interested, I need to make a lot of updates to the guide already. However, I won't do it yet as I'm still working on it. If you are interested in the technicalities, I have updated my Github

Preview:
FFT3Dgpu(sigma=10) applied to my CC on the left, dBox on the right

Correct me if I'm wrong but isn't the left a wildly inaccurate depiction of Boo's color? I was under the impression that what's seen on the right is what he's supposed to look like.

At first I used hues from the full colour manga (349° for him) but it looked weird, so I eyeballed the hues and I thought 311° looked good. Plus I think I've seen him with a similar hue before.

You got a recommendation?

P.S. If you think of a hue as containing all the colours in the spectrum + magenta and pink, under all intensities of neutral (white with correct white balance) lighting, it might help you understand why I match to them instead of specific RGB values.

[crabshank1]

UPDATE:

So I think I have cracked the problem of how to adjust the saturation channel. In my Github and the version of AvsPmod in my OneDrive, there is an image called "Almost uniform pattern.png". I match the saturation (0-100) in my upscaled dbox frame, to the value (0-100) in Almost uniform pattern.png, in my remapping page.

So the updated chain:

• Upscale
• Hues
• Saturation matched to value of noise image (OPTIONAL on most episodes, unless especially desaturated)
• Value matched to image created in GIMP by using Colors > Auto > Stretch Contrast on a DBox frame (OPTIONAL on most episodes, unless highlights blown out)
• Grey dither (https://github.com/crabshank/HLSL-Resha ... ither.hlsl)
• White balance
• Crush (OPTIONAL)
• Downscale.

I'm going to redo Dragon Ball E109 and post it to YouTube.

P.S. re:

P.S. If you think of a hue as containing all the colours in the spectrum + magenta and pink, under all intensities of neutral (white with correct white balance) lighting, it might help you understand why I match to them instead of specific RGB values.

the whole big rectangle is 42°.

[crabshank1]

Playing around with the saturation a bit more, and I thought to create a greyscale gradient and then put a gamma of 2.2 on it:


And remap the saturation of a test image to the value of the gradient:


The top is mine. Looks so good IMO.

So the updated chain:

• Upscale
• Hues
• Saturation matched to value of gamma image (https://github.com/crabshank/HLSL-Resha ... %202.2.png; also in OneDrive AvsPmod.zip file) (OPTIONAL on most episodes, or remapping page will not recommend any remapping, unless especially desaturated)
• Value matched to image created in GIMP by using Colors > Auto > Stretch Contrast on a DBox frame (OPTIONAL on most episodes, unless highlights blown out)
• Grey dither (https://github.com/crabshank/HLSL-Resha ... ither.hlsl)
• White balance
• Crush (OPTIONAL)
• Downscale

[HakkaiBills93]

You probably doing something wrong if you're getting that kind of result with dr dre's color matching tool.

Probably , i just load the source i want to correct (dbox screen)
load the cel
then use 100 colorspace and 0.9 in smooth (i tried also with 50c and 0.1 i got the same) probably somewhere else
anyway even if colors match with the shot and even if i can't say that colormatch tools is not something interesting, it have his own issue like the small area color issue
taking one shot from each part and find the same shot as references is a huge work , longer than what you can do with software adjustment cause except color cast that is differrent in each episodes part and sometimes even differrent within the same part (flashback or some damaged frames etc like ep with yamcha against invisible man where few frames are worst coloured than all others) , with after effect presets can be done for those part as long as you trimmed each part that need differrent color cast removal, between episodes you can have to alter a little some colors but compared to the color match tools that need a lut for each part trimmed, it's quickier


GT new upscale made by someone everyone know (i almost want to say ROUND 1 "fight")
https://www.youtube.com/watch?v=sc-tCnGlvE0

[lansing]

You probably doing something wrong if you're getting that kind of result with dr dre's color matching tool.

taking one shot from each part and find the same shot as references is a huge work , longer than what you can do with software adjustment cause except color cast that is differrent in each episodes part and sometimes even differrent within the same part (flashback or some damaged frames etc like ep with yamcha against invisible man where few frames are worst coloured than all others) , with after effect presets can be done for those part as long as you trimmed each part that need differrent color cast removal, between episodes you can have to alter a little some colors but compared to the color match tools that need a lut for each part trimmed, it's quickier

Dude, we are 63 pages into this topic and you're still talking about color correcting it by removing color cast? It had been blown out of the water by color matching like 2 years ago.

And I thought this debate was on selective color and color matching tool, why are you comparing it to an inferior method?

[HakkaiBills93]

You probably doing something wrong if you're getting that kind of result with dr dre's color matching tool.

taking one shot from each part and find the same shot as references is a huge work , longer than what you can do with software adjustment cause except color cast that is differrent in each episodes part and sometimes even differrent within the same part (flashback or some damaged frames etc like ep with yamcha against invisible man where few frames are worst coloured than all others) , with after effect presets can be done for those part as long as you trimmed each part that need differrent color cast removal, between episodes you can have to alter a little some colors but compared to the color match tools that need a lut for each part trimmed, it's quickier

Dude, we are 63 pages into this topic and you're still talking about color correcting it by removing color cast? It had been blown out of the water by color matching like 2 years ago.

And I thought this debate was on selective color and color matching tool, why are you comparing it to an inferior method?

i never said that removing the color cast is the only thing to do in color correction, selective color don't remove any color cast at all, and i know agument with you is useless as you can't accept being wrong.

inférior méthod only when you are unable to use others.

yes the color match tools require to have two sources that have the same level as contrast and luminosity to work perfectly and good luck if you are thinking that it can be reach with references available.

Even xerox code don't solve the issue as it will require to change brightness and luminosity for all color to match your source but all source aren't the same meaning that if you use it on one part, it won't work the same on another, if you match an overbright references you'll get overbright result that will be destructive for sure and that you'll never be able to reverse , like when you was trying to match season sets which have crushed black and overbrightness and loose some lines.

at least you have more control with after effect with lumetri colors, levels , selective colors and probably other plugins (and if you are unable to match a source using after effect it doesn't mean that it's impossible or inferior) than with a .lut files that can't work for every single episodes and all parts but will require as much lut than part to correct unlike after effect preset that you can create that will only need white balance adjustment for each part and all others colors will be okay with the remaining settings

i am curious to see one full episode done with a single lut (and i talk about dbox episodes really dark or really bright) to see if what you think of the "superior" way is really superior way, there is plenty of differrent way to cc, for me color match tools is perfect when you have two similar sources where only colors are differrent with one really reliable as references, dragon ball don't have this kind of release, none are perfect

[crabshank1]

yes the color match tools require to have two sources that have the same level as contrast and luminosity to work perfectly and good luck if you are thinking that it can be reach with references available.

Even xerox code don't solve the issue as it will require to change brightness and luminosity for all color to match your source but all source aren't the same meaning that if you use it on one part, it won't work the same on another, if you match an overbright references you'll get overbright result that will be destructive for sure and that you'll never be able to reverse , like when you was trying to match season sets which have crushed black and overbrightness and loose some lines.

at least you have more control with after effect with lumetri colors, levels , selective colors and probably other plugins (and if you are unable to match a source using after effect it doesn't mean that it's impossible or inferior) than with a .lut files that can't work for every single episodes and all parts but will require as much lut than part to correct unlike after effect preset that you can create that will only need white balance adjustment and all others colors will be okay with the remaining settings

i am curious to see one full episode done with a single lut to laugh about all white balance issue you'll have in the result

And this is exactly why I match hues, they're totally separate from brightness and white balancing. Thanks for making my point.

[lansing]

taking one shot from each part and find the same shot as references is a huge work , longer than what you can do with software adjustment cause except color cast that is differrent in each episodes part and sometimes even differrent within the same part (flashback or some damaged frames etc like ep with yamcha against invisible man where few frames are worst coloured than all others) , with after effect presets can be done for those part as long as you trimmed each part that need differrent color cast removal, between episodes you can have to alter a little some colors but compared to the color match tools that need a lut for each part trimmed, it's quickier

Dude, we are 63 pages into this topic and you're still talking about color correcting it by removing color cast? It had been blown out of the water by color matching like 2 years ago.

And I thought this debate was on selective color and color matching tool, why are you comparing it to an inferior method?

Even xerox code don't solve the issue as it will require to change brightness and luminosity for all color to match your source but all source aren't the same meaning that if you use it on one part, it won't work the same on another, if you match an overbright references you'll get overbright result that will be destructive for sure and that you'll never be able to reverse , like when you was trying to match season sets which have crushed black and overbrightness and loose some lines.

at least you have more control with after effect with lumetri colors, levels , selective colors and probably other plugins (and if you are unable to match a source using after effect it doesn't mean that it's impossible or inferior) than with a .lut files that can't work for every single episodes and all parts but will require as much lut than part to correct unlike after effect preset that you can create that will only need white balance adjustment and all others colors will be okay with the remaining settings

i am curious to see one full episode done with a single lut to laugh about all white balance issue you'll have in the result

What are you even talking about?? The use the color matching is going to be apply in a scene by scene basis, as I had posted about it in this very thread a year ago. What are you smoking?

yes the color match tools require to have two sources that have the same level as contrast and luminosity to work perfectly and good luck if you are thinking that it can be reach with references available.

Dude, I'm going for the highest quality color correction, not some amateur cc people have been doing that ended up going nowhere. When you have seen the best, there is no point to even go for the inferior methods. And in my latest finding here I had explained why collecting references is actually doable now.

i never said that removing the color cast is the only thing to do in color correction, selective color don't remove any color cast at all, and i know agument with you is useless as you can't accept being wrong.

Well, then there you go with that limitation on selective color when compares to color matching programs. The latter one doesn't care about color cast.

Spoiler: