Color Correcting the Dragon Box - 3 Part Spectacular

Discussion regarding the entirety of the franchise in a general (meta) sense, including such aspects as: production, trends, merchandise, fan culture, and more.

Moderators: General Help, Kanzenshuu Staff

User avatar
jjgp1112
Born 'n Bred Here
Posts: 7478
Joined: Mon Jul 23, 2007 10:15 pm
Location: Crooklyn

Re: Color Correcting the Dragon Box - 3 Part Spectacular

Post by jjgp1112 » Thu May 05, 2022 4:50 pm

vanner64 wrote: Thu May 05, 2022 4:08 pm
Grande Pesquisador wrote: Thu May 05, 2022 3:52 pm Is this the SoM group project? I've been looking for information about this for a while. I'd love to see more of what you've been up to and when it will be available.
Thanks appreciate it! My post saying that the original Dragon Ball show has been white balanced and we have a draft of color correction for all 153 episodes is really the most up-to-date news at this time. I'm not sure what the rules are for posting Discord invite links here, but you can find an invite link on any SoM releases. We often discuss smaller updates there.
I'm no mod but people have posted Discords on here with no problem
Yamcha: Do you remember the spell to release him - do you know all the words?
Bulma: Of course! I'm not gonna pull a Frieza and screw it up!
Master Roshi: Bulma, I think Frieza failed because he wore too many clothes!
Cold World (Fanfic)
"It ain't never too late to stop bein' a bitch." - Chad Lamont Butler

User avatar
lansing
Beyond Newbie
Posts: 282
Joined: Mon Jan 06, 2014 1:49 am

Re: Color Correcting the Dragon Box - 3 Part Spectacular

Post by lansing » Thu May 05, 2022 5:19 pm

jjgp1112 wrote: Thu May 05, 2022 3:12 pm WHere in his elaboration did he say he didn't believe this? He said he didn't think Toei was trying to find the best film stock to achieve a certain color.

But what we are we trying to say is that ALL film stock will alter colors, no matter what your artistic goal is.

Meaning, that Toei will keep this in mind. They're not using a particularly ambitious film stock, but they're using *a* film stock, meaning that they will tune the colors to the expected output. A standard damn procedure. All he did was point out the more ambitious uses of film stock as an easy way to demonstrate how artists are aware of how film stock enters into the equation - and now you're fixating on that because you keep trying to play "gotcha!" I mean holy crap, the very same sentence you're lasering on even establishes the nuance:
So now there is a tuning process added to the story? Where in the chain does this happened? I'm lost.
jjgp1112 wrote: Thu May 05, 2022 3:12 pm
Now because Dragon Ball was produced for TV on the cheap, it's very unlikely that anyone was purposefully selecting film stocks to modify the cel colors in an artistic way, however that does not negate the affect that film has on the colors
Our entire damn point is that colors on the original film print were the intended final product and that the cels were painted in such a way to facilitate this. Since we don't have the original film prints, a combination of the cel colors (figuring out the values and such) and promotional material is the best way to figure out how the show was intended to look.
That statement means that the animators did not intentionally go for any particular looks when choosing film stock, which means that they wouldn't know/don't care about the outcome of the film stock they use. So the extra properties the film stock added to the color would not be their intended color.
And your colors are still oversaturated.
What does this even mean? You have been told that those are the accurate cel color and you don't like it? What do you want me to say??

User avatar
vanner64
Newbie
Posts: 36
Joined: Wed Jan 26, 2022 3:43 pm

Re: Color Correcting the Dragon Box - 3 Part Spectacular

Post by vanner64 » Thu May 05, 2022 5:53 pm

lansing wrote: Thu May 05, 2022 5:19 pm What does this even mean? You have been told that those are the accurate cel color and you don't like it? What do you want me to say??
You say this, but no one knows what method you use (unless you shared it awhile ago but I cannot find it). So I ask again to better understand your process:

Which of the following methods would you say is the best/correct way for color correction (not for all media ever, just this specific show):

1. Strict adherence to the colors of cels found online.
2. Strict adherence to color charts.
3. A subjective combination of cels and color charts but nothing else.
4. A subjective combination of cels, color charts, promotional images, and original broadcast.
5. Some other method (please share what that would be if you say this one)
Last edited by vanner64 on Thu May 05, 2022 6:13 pm, edited 1 time in total.

User avatar
jjgp1112
Born 'n Bred Here
Posts: 7478
Joined: Mon Jul 23, 2007 10:15 pm
Location: Crooklyn

Re: Color Correcting the Dragon Box - 3 Part Spectacular

Post by jjgp1112 » Thu May 05, 2022 6:04 pm

lansing wrote: Thu May 05, 2022 5:19 pm

snip
I'm going to ask you to please pay closer attention to the words that you are reading and THINK

-ikaos says that all film alters the colors of the source
-ikaos then points out that some film makers even go as far as picking film stock to achieve a specific effect just as an example to reinforce that
-ikaos clarifies that while Toei probably didn't go as far as picking something to achieve a special effect to the colors, they were still aware that even the standard stock they used was going to alter their colors

You: "yOu jUsT cOnTrAdIcTeD yOurSeLf"
So now there is a tuning process added to the story? Where in the chain does this happened? I'm lost.
Please tell me you're being literal and pedantic on purpose. There is no "tuning process." I used tune as what I sure thought would be a simple way of just saying "They picked the colors that would lead to the results they wanted from the 16mm film prints."

(Actually you know what? There probably is a tuning process. Colorists put a lot of thought into what they do even on a low-budget production like Dragon Ball. So they most likely assessed how their junky film stock would alter the colors and then based their cel paint pigments on their findings)

Again, you seem to be stuck on ikaos' example and somehow think it's contradictory. It's not. Just because Toei wasn't meticulously going through different types of film stock to get some neat effect on the colors doesn't mean they're not aware of the fact that whatever they do use will ultimately change the colors.

Toei may not be ambitious, but they're not idiots either - they are a multi-billion dollar company pushing out a brand and want to keep a consistent look and presentation with that brand, which means making sure the primary color palette is correct across all materials. And so they're going to make sure that with whatever film stock they use, the colors that come out are the same as their promotional material...which means, everything I just wrote in the parentheses above. Trust me, I work in media and see first-hand how calculated all this stuff is.
Yamcha: Do you remember the spell to release him - do you know all the words?
Bulma: Of course! I'm not gonna pull a Frieza and screw it up!
Master Roshi: Bulma, I think Frieza failed because he wore too many clothes!
Cold World (Fanfic)
"It ain't never too late to stop bein' a bitch." - Chad Lamont Butler

User avatar
lansing
Beyond Newbie
Posts: 282
Joined: Mon Jan 06, 2014 1:49 am

Re: Color Correcting the Dragon Box - 3 Part Spectacular

Post by lansing » Thu May 05, 2022 7:53 pm

vanner64 wrote: Thu May 05, 2022 5:53 pm
lansing wrote: Thu May 05, 2022 5:19 pm What does this even mean? You have been told that those are the accurate cel color and you don't like it? What do you want me to say??
You say this, but no one knows what method you use (unless you shared it awhile ago but I cannot find it). So I ask again to better understand your process:

Which of the following methods would you say is the best/correct way for color correction (not for all media ever, just this specific show):

1. Strict adherence to the colors of cels found online.
2. Strict adherence to color charts.
3. A subjective combination of cels and color charts but nothing else.
4. A subjective combination of cels, color charts, promotional images, and original broadcast.
5. Some other method (please share what that would be if you say this one)
I'm using color matching program now. There is an accuracy rating in color matching called average deltaE, the lower the number the better. An average deltaE of 3 means really good. And in this example, I got a 4.24 among the two character models used, which means the matching is pretty accurate.

User avatar
vanner64
Newbie
Posts: 36
Joined: Wed Jan 26, 2022 3:43 pm

Re: Color Correcting the Dragon Box - 3 Part Spectacular

Post by vanner64 » Thu May 05, 2022 9:18 pm

lansing wrote: Thu May 05, 2022 7:53 pm
vanner64 wrote: Thu May 05, 2022 5:53 pm
lansing wrote: Thu May 05, 2022 5:19 pm What does this even mean? You have been told that those are the accurate cel color and you don't like it? What do you want me to say??
You say this, but no one knows what method you use (unless you shared it awhile ago but I cannot find it). So I ask again to better understand your process:

Which of the following methods would you say is the best/correct way for color correction (not for all media ever, just this specific show):

1. Strict adherence to the colors of cels found online.
2. Strict adherence to color charts.
3. A subjective combination of cels and color charts but nothing else.
4. A subjective combination of cels, color charts, promotional images, and original broadcast.
5. Some other method (please share what that would be if you say this one)
I'm using color matching program now. There is an accuracy rating in color matching called average deltaE, the lower the number the better. An average deltaE of 3 means really good. And in this example, I got a 4.24 among the two character models used, which means the matching is pretty accurate.
Oh interesting. What things are you trying to match to each other? I assume one is DBOX (maybe a DBOX upscale) and the other image that is your color reference is..?
Last edited by vanner64 on Thu May 05, 2022 9:25 pm, edited 1 time in total.

nward50987
Newbie
Posts: 18
Joined: Tue Feb 22, 2022 12:47 am

Re: Color Correcting the Dragon Box - 3 Part Spectacular

Post by nward50987 » Thu May 05, 2022 9:21 pm

Here is a CC using a new method. Let me know how it looks. https://imgsli.com/MTA2NzE0

User avatar
lansing
Beyond Newbie
Posts: 282
Joined: Mon Jan 06, 2014 1:49 am

Re: Color Correcting the Dragon Box - 3 Part Spectacular

Post by lansing » Fri May 06, 2022 2:16 am

vanner64 wrote: Thu May 05, 2022 9:18 pm
lansing wrote: Thu May 05, 2022 7:53 pm
vanner64 wrote: Thu May 05, 2022 5:53 pm

You say this, but no one knows what method you use (unless you shared it awhile ago but I cannot find it). So I ask again to better understand your process:

Which of the following methods would you say is the best/correct way for color correction (not for all media ever, just this specific show):

1. Strict adherence to the colors of cels found online.
2. Strict adherence to color charts.
3. A subjective combination of cels and color charts but nothing else.
4. A subjective combination of cels, color charts, promotional images, and original broadcast.
5. Some other method (please share what that would be if you say this one)
I'm using color matching program now. There is an accuracy rating in color matching called average deltaE, the lower the number the better. An average deltaE of 3 means really good. And in this example, I got a 4.24 among the two character models used, which means the matching is pretty accurate.
Oh interesting. What things are you trying to match to each other? I assume one is DBOX (maybe a DBOX upscale) and the other image that is your color reference is..?
Yes, I'm matching the color on the dragon box image ikaos shared to the cel reference chart with the color design sheets for Goku and Piccolo.

User avatar
lansing
Beyond Newbie
Posts: 282
Joined: Mon Jan 06, 2014 1:49 am

Re: Color Correcting the Dragon Box - 3 Part Spectacular

Post by lansing » Fri May 06, 2022 4:35 am

jjgp1112 wrote: Thu May 05, 2022 6:04 pm
Please tell me you're being literal and pedantic on purpose. There is no "tuning process." I used tune as what I sure thought would be a simple way of just saying "They picked the colors that would lead to the results they wanted from the 16mm film prints."

(Actually you know what? There probably is a tuning process. Colorists put a lot of thought into what they do even on a low-budget production like Dragon Ball. So they most likely assessed how their junky film stock would alter the colors and then based their cel paint pigments on their findings)

Again, you seem to be stuck on ikaos' example and somehow think it's contradictory. It's not. Just because Toei wasn't meticulously going through different types of film stock to get some neat effect on the colors doesn't mean they're not aware of the fact that whatever they do use will ultimately change the colors.

Toei may not be ambitious, but they're not idiots either - they are a multi-billion dollar company pushing out a brand and want to keep a consistent look and presentation with that brand, which means making sure the primary color palette is correct across all materials. And so they're going to make sure that with whatever film stock they use, the colors that come out are the same as their promotional material...which means, everything I just wrote in the parentheses above. Trust me, I work in media and see first-hand how calculated all this stuff is.
What you're ultimately suggesting is that Toei would be using some kind of color chart to correct the altered color the different film stocks introduced so that the color would be consistent throughout the show, which I agreed. And with the altered color fixed, the original intended color captured can be restored, which is the cel color. You are actually defying what he and his group were trying to claim with the altered film color being the intended color of the show.

User avatar
jjgp1112
Born 'n Bred Here
Posts: 7478
Joined: Mon Jul 23, 2007 10:15 pm
Location: Crooklyn

Re: Color Correcting the Dragon Box - 3 Part Spectacular

Post by jjgp1112 » Fri May 06, 2022 6:14 am

lansing wrote: Fri May 06, 2022 4:35 am
jjgp1112 wrote: Thu May 05, 2022 6:04 pm
Please tell me you're being literal and pedantic on purpose. There is no "tuning process." I used tune as what I sure thought would be a simple way of just saying "They picked the colors that would lead to the results they wanted from the 16mm film prints."

(Actually you know what? There probably is a tuning process. Colorists put a lot of thought into what they do even on a low-budget production like Dragon Ball. So they most likely assessed how their junky film stock would alter the colors and then based their cel paint pigments on their findings)

Again, you seem to be stuck on ikaos' example and somehow think it's contradictory. It's not. Just because Toei wasn't meticulously going through different types of film stock to get some neat effect on the colors doesn't mean they're not aware of the fact that whatever they do use will ultimately change the colors.

Toei may not be ambitious, but they're not idiots either - they are a multi-billion dollar company pushing out a brand and want to keep a consistent look and presentation with that brand, which means making sure the primary color palette is correct across all materials. And so they're going to make sure that with whatever film stock they use, the colors that come out are the same as their promotional material...which means, everything I just wrote in the parentheses above. Trust me, I work in media and see first-hand how calculated all this stuff is.
What you're ultimately suggesting is that Toei would be using some kind of color chart to correct the altered color the different film stocks introduced so that the color would be consistent throughout the show, which I agreed. And with the altered color fixed, the original intended color captured can be restored, which is the cel color. You are actually defying what he and his group were trying to claim with the altered film color being the intended color of the show.
Holy FUCK you are just reading what you want to read, huh? Jesus Christ.

Seriously guys, just take a second to read stuff over again before responding.

The film is the final product.

I repeat: the film is the final product.

Not the cels.

The cels are a means to an end.

And since the film alters the colors of the cel, they're going to paint the cels in a way so that the final product, meaning the film, looks like what they want. This color is the color they want, so they paint the cel this color to get it. Again: THE FINAL PRODUCT IS WHAT THE FANS ARE GOING TO SEE.

The fans do not see the other side of the process. They only see the finished work. And so they make sure the finished work comes off as needed.

For example: My personal laptops monitor is shitty and distorts the colors because HP white balances their displays differently from literally everyone elses. Colors that should be purple look blue. So if I were to design something, the colors would look wildly different on other monitors. So I pick colors based not on how they look on my monitor, but how they'll look on other monitors. This is why professionals use color values instead of simply going off the colors they're looking at, because again: the final product is the goal!

Toei wasn't thinking about a restoration process back then, which is the main reason why the masters weren't preserved as well as they should have been. Home video for TV shows werent a thing. So no, this interpretation you somehow arrived at from my post is imaginary.

You're so stuck on the cel colors that you won't listen to reason and will just twist everyone else's words to keep rolling with what you think.

I mean for Christ's sake, ikaos already showed a quote 2 pages ago of a producer straight up saying that they paint the cels a different color to account for how the film changes it, but here you are.
Yamcha: Do you remember the spell to release him - do you know all the words?
Bulma: Of course! I'm not gonna pull a Frieza and screw it up!
Master Roshi: Bulma, I think Frieza failed because he wore too many clothes!
Cold World (Fanfic)
"It ain't never too late to stop bein' a bitch." - Chad Lamont Butler

User avatar
lansing
Beyond Newbie
Posts: 282
Joined: Mon Jan 06, 2014 1:49 am

Re: Color Correcting the Dragon Box - 3 Part Spectacular

Post by lansing » Fri May 06, 2022 9:52 am

jjgp1112 wrote: Fri May 06, 2022 6:14 am I mean for Christ's sake, ikaos already showed a quote 2 pages ago of a producer straight up saying that they paint the cels a different color to account for how the film changes it, but here you are.
The paints used in cel production are manufactured in slight variations of the desired colour in order to compensate for the extra layer of transparent cel added between the image and the camera; for example a slightly brighter colour would be used for a bottom layer cel to offset the dulling effect of the cels on top
https://japanesegallery.com/anime_and_m ... els-cel-ga

That is not what the article is talking about. It's talking about paint company offsetting the desired paint (a little brighter) so that when they were painted on the back of the cel sheet, the offset will compensate for the thickness of the sheet (darkening) so that the desired color will come out correctly when you look at it from the front. It has nothing to do with film.

Image
jjgp1112 wrote: Fri May 06, 2022 6:14 am Seriously guys, just take a second to read stuff over again before responding.
The exact statement right back to you. READ. Or own a damn cel to understand what it's talking about before building your whole fantasy around it.

If that article is the only source where you base your fantasy from, your whole story has now be debunked.

User avatar
VegettoEX
Kanzenshuu Co-Owner & Administrator
Posts: 17537
Joined: Sat Jan 10, 2004 3:10 pm
Location: New Jersey
Contact:

Re: Color Correcting the Dragon Box - 3 Part Spectacular

Post by VegettoEX » Fri May 06, 2022 11:15 am

I don't understand why you have to be such goddamn weirdos at each other about this. Just... talk? And be nice and cool?

That's a formal warning, by the way.
:: [| Mike "VegettoEX" LaBrie |] ::
:: [| Kanzenshuu - Co-Founder/Administrator, Podcast Host, News Manager (note: our "job" titles are arbitrary and meaningless) |] ::
:: [| Website: January 1998 |] :: [| Podcast: November 2005 |] :: [| Fusion: April 2012 |] :: [| Wiki: 20XX |] ::

User avatar
vanner64
Newbie
Posts: 36
Joined: Wed Jan 26, 2022 3:43 pm

Re: Color Correcting the Dragon Box - 3 Part Spectacular

Post by vanner64 » Fri May 06, 2022 6:22 pm

I think instead of swearing and name calling we would all be better off trying a more objective analysis of the images posted here. I'm going to go through some problems with one of the correction attempts and methods I've seen mentioned here. II think a ton of misinformation has come from people listening to others without evidence. I'm going to share my resources so you (yes you the reader) could follow along and try this method out.
One very handy resource I use is Just Color Picker, which you can download here for PC/Mac. This will show you the color and values of pixels on your screen, here is an example:
Those three values of 0 degrees, 0%, and 100% using the HSB/HSV readings under 7x7 Avg will be important later (remember this!). But first, I'll change it to RGB to get the Red, Green, Blue color values.

Iansing posted this attempt at color correction:
And also posted this color reference chart:
I had a feeling something was off, but maybe I was wrong to think this way? Is there a more objective/scientific way to confirm one way or the other? The thing that stood out to me the most was the color red in the chart compared to his color correction attempt. Let's take a look at this red color:
(FYI there is a second red color but this matches exactly, feel free to check yourself)
The color value we get from the chart red is: R=224, G=35, B=13

So how does that compare to the red in the color correction? To account for grain/noise, I changed the sample area to 7x7 (seven by seven pixels on your screen which is the highest setting), which will "average out" any grain/noise to get a fair assessment. I went for what I thought was the largest red area on Piccolo on that color correction. With the 7x7 sample and picking areas away from lines/edges we eliminate issues of "bad or unfair samples". This is what I got:
(Note, the yellow circle was added after I got the sample, just to share the approximate area my cursor was on)
The color value we get from the correction red is R=208, G=49, B=1

OK so these values are different, but how different are they? I wasn't sure if there was a good way to measure how close colors are, until Iansing shared this:
lansing wrote: Thu May 05, 2022 7:53 pm I'm using color matching program now. There is an accuracy rating in color matching called average deltaE, the lower the number the better. An average deltaE of 3 means really good. And in this example, I got a 4.24 among the two character models used, which means the matching is pretty accurate.
This is excellent! We can now use his own recommended verification method to see how well his correction matches the color charts.

I found an online tool that can calculate deltaE just by inputting the RGB values, and anyone is free to check it for themselves. This is my result from comparing the two RGB values:
In this instance we get a value of about 15.38. This is much higher than what was claimed to be "really good" (remember, lower is better). These colors are certainly no match.

"Now wait a minute Mr. 64!" you might proclaim. "Isn't it unfair to just take a single sample and conclude that the image does not match the color chart?" To that I reply "You're correct, that would be unfair... and that's why I took four more samples"

Sample 2
deltaE calculation 2
Sample 3
deltaE calculation 3
Sample 4
deltaE calculation 4
Sample 5
deltaE calculation 5
I tried my best to pick fair red areas and minimize the deltaE value, but there was not a single instance in which the deltaE is less than 8.0 which, according to this writeup on deltaE, ranges from the high end of "perceptible at a glance" to "colors are more similar than opposite".

By Iansing's own method, his own measurement, and his own standards, I have shown that the color correction is not up to par. But you don't have to take my word for it, all the resources and tools are there to easily check for yourself. The color red stood out to me the most as not matching, but anyone else is free to check the other colors as well.

That ends my part about color, now I want to discuss something else that bothered me about that correction attempt: the black and white of the image. Something was bothering about the way Goku looked, so I cropped the image to get a better look here:
I have two simple questions for anyone familiar with Dragon Ball reading this:
What color are Goku's eyes?
What color is Goku's hair?

I would hope we can all agree Goku has white eyes (not counting his pupils), as do most human-like anime characters. We should also all be able to agree Goku has black hair. If anyone has a problem with either of these statements, I can only conclude you have never seen Goku before or are trolling.

Now, when I noticed that image of Goku from the correction, something seemed off to me. The whites of his eyes and teeth didn't seem right and his hair didn't look as black as it should be. So, I did a white balance on just that part of the image with the following result:
Now let me get ahead of potential comments regarding the color of his skin, gi, and the sky. I know those colors are different now and may look "wrong" to someone. I am only doing this comparison to check the white balance of the image. Normally, we would do a white balance first and then hue shift/color correct, but I'm looking at this already color corrected image.

Is there a way we can objectively measure how white/black a color is instead of saying "Well, I feel like one looks better." Turns out we can with Just Color Picker. I said to remember the HSB/HSV setting of Just Color Picker since that is what we're going to use now. The first number, the degrees, represents hue, but is not very important when looking at blacks/whites. The second number, the saturation percentage, determines how "colorful" something is. 100% would be a rich saturated red/green/magenta whatever, and a 0% would be black/white/grey. The third number, the brightness or value number, determines how bright something is, with 100% being bright white and 0% being the darkest black. We can use the saturation number (the second one) to measure how black/white an image is, with lower saturation being better, and 0% saturation being perfect.

That being said, it's impossible to get exactly 0% saturation for an entire white area on something with grain since the grain and inconsistency in the film causes tiny changes to the color. The Seed of Might standard based on visual perception is 4% and lower saturation for whites unless absolutely impossible to avoid due to large white areas and film inconsistency. Our standards for blacks is under 20% saturation, but this one is even harder to achieve consistently based on Dragon Ball at least. (I can explain why if anyone asks but this has gone on long enough) Now, anyone is free to have their own standards on what is "good enough", but we can objectively say that lower saturation produces better blacks/whites.

With that out of the way, let's compare these two images. Iansing's will always be on the left and mine will always be on the right. The yellow circles are added after the fact to show anyone where I sampled the color from. Let's start with the white areas.

Iansing's left eye:
My left eye:
Remember, you're looking under the "7x7 Avg" rectangle and want to pay attention to the second number. Iansing has a saturation of 5% and has a bit of a creamy look, while I was able to get a saturation of 0%. Let's check some more areas.

Iansing's right eye:
My right eye:
Iansing's teeth:
My teeth:
OK that's it for white areas, but how do the black areas compare?

Iansing's forehead:
My forehead:
Iansing's behind the ear:
My behind the ear:
Iansing's left brow:
My left brow:
Iansing's right brow:
My right brow:
Iansing's hair spike:
My hair spike:
Now there's something pretty interesting about this last one. My saturation in the black hair is 24% and "fails" my standard. This ties back to what I mentioned about the black areas being inconsistent for saturation. If I fixed this spot, it would make a different spot look worse. We must accept some limitations. It does appear there is some blue light on the right side of the image anyway and on the Iansing version it is 41% saturation.

In all examples, my version had less saturation in the whites/blacks, so when jjgp1112 said:
jjgp1112 wrote: Thu May 05, 2022 3:12 pm And your colors are still oversaturated.
These may or may not have been the colors he was referring to, but he is absolutely correct.

Once again, if anyone doesn't believe me, you are free to check for yourself.

By two objective measures of color, deltaE for the chart and HSB/HSV saturation for blacks and white, we can conclude two things:
That color correction attempt that strives to match the color chart does not match the color chart.
The color correction attempt can be better black/white balanced.

My question now for everyone is this: If the person who is dedicated to correcting the image to match the color charts couldn't make a properly matched and balanced image, why should anyone even attempt to try using that method?


NOTES:
  • While I did try and share the software/websites used for this, I could not share my white balance tool (Vegas Pro 19) since that is paid software. My comparisons and readings should be enough to show that the image is properly white balanced.
  • If you try taking samples on your own and get something that is a tiny amount different (Ex: my RGB = 100, 100, 100 and yours = 99, 101, 102) please don't use that as some kind of "gotcha!". The color readings change just by moving a single pixel. I tried my best to be fair and sample away from edges due to the way color/light can bleed through a little bit. You'll notice in dark areas it changes even more.
  • If anyone has any objection, please respond with actual images and comparisons. Just writing "I used the chart/cels so you're wrong!" isn't an intelligent response, as I have just shown that using the color chart as a reference does not guarantee you were successful in matching that reference.

User avatar
lansing
Beyond Newbie
Posts: 282
Joined: Mon Jan 06, 2014 1:49 am

Re: Color Correcting the Dragon Box - 3 Part Spectacular

Post by lansing » Fri May 06, 2022 11:48 pm

vanner64 wrote: Fri May 06, 2022 6:22 pm This is excellent! We can now use his own recommended verification method to see how well his correction matches the color charts.

...

By Iansing's own method, his own measurement, and his own standards, I have shown that the color correction is not up to par.
I wouldn't be writing up a full on report to challenge something I don't know if I were you.

Normally deltaE should be measured in the LAB color space, not sRGB, so your testing were wrong from the start. But more importantly, what I said in my original quote was "average deltaE", not "deltaE". Measuring the deltaE of a single color when your goal is to try to match a whole bunch of them is pointless. And I have already told you the average deltaE for that sample image, I mean what are you doing??
vanner64 wrote: Fri May 06, 2022 6:22 pm now I want to discuss something else that bothered me about that correction attempt: the black and white of the image. Something was bothering about the way Goku looked
...
Now, when I noticed that image of Goku from the correction, something seemed off to me. The whites of his eyes and teeth didn't seem right and his hair didn't look as black as it should be.
Have you even spent the time to look at the black and white on the cel reference chart? That is what they look like, they are not bit perfect neutral black and neutral white. I'm just sticking to the rule of matching to those color. And yes I can easily "fix" them if I wanted, but that's not the purpose of showing the image.
vanner64 wrote: Fri May 06, 2022 6:22 pm My question now for everyone is this: If the person who is dedicated to correcting the image to match the color charts couldn't make a properly matched and balanced image, why should anyone even attempt to try using that method?
Wait what? It was one of your group member who asked for an example to see what the cel color looks like. My intention in this whole conversation here was on what you guy claimed to be the intended color of the show.
then by all means please show an example, using the color reference charts and raw Dragon Box footage, of how it should look in your eyes.

User avatar
jjgp1112
Born 'n Bred Here
Posts: 7478
Joined: Mon Jul 23, 2007 10:15 pm
Location: Crooklyn

Re: Color Correcting the Dragon Box - 3 Part Spectacular

Post by jjgp1112 » Sat May 07, 2022 12:57 am

If not even the regular blacks and whites on the chart are balanced and neutral, then...nah nvm :lol:
Yamcha: Do you remember the spell to release him - do you know all the words?
Bulma: Of course! I'm not gonna pull a Frieza and screw it up!
Master Roshi: Bulma, I think Frieza failed because he wore too many clothes!
Cold World (Fanfic)
"It ain't never too late to stop bein' a bitch." - Chad Lamont Butler

User avatar
Trachta10
Beyond Newbie
Posts: 191
Joined: Mon Aug 10, 2015 9:32 pm

Re: Color Correcting the Dragon Box - 3 Part Spectacular

Post by Trachta10 » Sat May 07, 2022 3:07 am

nward50987 wrote: Thu May 05, 2022 9:21 pm Here is a CC using a new method. Let me know how it looks. https://imgsli.com/MTA2NzE0
Really really good :thumbup:

User avatar
Trachta10
Beyond Newbie
Posts: 191
Joined: Mon Aug 10, 2015 9:32 pm

Re: Color Correcting the Dragon Box - 3 Part Spectacular

Post by Trachta10 » Sat May 07, 2022 6:41 am

An attempt to correct the colors in episode 30, of the new Dragon Box Selecta Vision upscale
Didn't try to be accurate with the colors, is just by eye, more or less what seemed good to me.

2 minutes video
https://drive.google.com/file/d/1yvI0Ae ... sp=sharing

https://imgsli.com/MTA2ODY2
https://imgsli.com/MTA2ODcx

Compared to other versions


User avatar
Trachta10
Beyond Newbie
Posts: 191
Joined: Mon Aug 10, 2015 9:32 pm

Re: Color Correcting the Dragon Box - 3 Part Spectacular

Post by Trachta10 » Sat May 07, 2022 6:45 am

lansing wrote: Fri May 06, 2022 2:16 am
vanner64 wrote: Thu May 05, 2022 9:18 pm
lansing wrote: Thu May 05, 2022 7:53 pm

I'm using color matching program now. There is an accuracy rating in color matching called average deltaE, the lower the number the better. An average deltaE of 3 means really good. And in this example, I got a 4.24 among the two character models used, which means the matching is pretty accurate.
Oh interesting. What things are you trying to match to each other? I assume one is DBOX (maybe a DBOX upscale) and the other image that is your color reference is..?
Yes, I'm matching the color on the dragon box image ikaos shared to the cel reference chart with the color design sheets for Goku and Piccolo.

Can you try to match this two screenshots with your color chart?

User avatar
lansing
Beyond Newbie
Posts: 282
Joined: Mon Jan 06, 2014 1:49 am

Re: Color Correcting the Dragon Box - 3 Part Spectacular

Post by lansing » Sat May 07, 2022 3:28 pm

Trachta10 wrote: Sat May 07, 2022 6:45 am
lansing wrote: Fri May 06, 2022 2:16 am
vanner64 wrote: Thu May 05, 2022 9:18 pm

Oh interesting. What things are you trying to match to each other? I assume one is DBOX (maybe a DBOX upscale) and the other image that is your color reference is..?
Yes, I'm matching the color on the dragon box image ikaos shared to the cel reference chart with the color design sheets for Goku and Piccolo.

Can you try to match this two screenshots with your color chart?
I don't have reference for the sky and the rock, if that is what you're looking for. So I can only match to the Goku model. We'll need a model with similar color at the scene to get the sky color relatively. Just a guess, maybe Bulma's hair when she arrived at the scene?

Image

User avatar
FPSSJ4_Goku
Regular
Posts: 534
Joined: Tue Aug 11, 2020 1:33 pm
Location: New York, US
Contact:

Re: Color Correcting the Dragon Box - 3 Part Spectacular

Post by FPSSJ4_Goku » Sat May 07, 2022 7:35 pm

Hey, would it be okay if I posted my own colored version of this cel?
So, you decided to read my signature, eh?

If you'd like, check out my YouTube channel, and maybe subscribe?

Locked