Color Correction Test for Editors

musicvid10 wrote on 1/30/2014, 12:09 PM
Rather than hijack Gaross' provocative thread, here is a basic color correction test for video editors.

This is called a "ringaround" in the photo industry. It is used as a cc reference, to calibrate optical printers, and I also used it to screen new employees at Pallas back in the day ;?)

In the spirit of a visual exercise, no eyedroppers, samples, or masks, please. You are not to judge the color of the reference (it's your monitor, remember?). The choices are C<M<Y<R<G<B in random order, starting at the lower left and going clockwise. There are no "tricks" (i.e., combination colors have been left out). Notice how flesh tones (mostly red and yellow) don't always behave the way we expect to color shifts. A few experienced graders will get it right away. Most won't. That's why we need a reference.

Comments

richard-amirault wrote on 1/30/2014, 12:31 PM
HEY, THAT'S SHIRLEY!! (supposedly the name of the model that was used for the first "test negative" from Kodak. The name stuck no matter who the model was)

(on a more serious note) I don't think you mentioned WHAT we are supposed to "choose" You say The choices are C<M<Y<R<G<B in random order, starting at the lower left and going clockwise. I don't know what that means.
musicvid10 wrote on 1/30/2014, 12:57 PM
Is the first example red, green, blue, cyan, magenta, or yellow when compared to the reference image in the middle?

Repeat the same question for the second through sixth examples, starting at the lower left, and going clockwise in numeric order. No cheating.

Clearer?
larry-peter wrote on 1/30/2014, 1:25 PM
From my desktop monitor (not my editing monitor) I would say 1-6 is G, R, C, B, Y, M.
john_dennis wrote on 1/30/2014, 2:57 PM
Thanks for posting this since I've been meaning to learn something (anything) about color. I thought today would be a good day to start since I'm recovering from shoulder surgery and I'm certainly not going anywhere.

I created a simple project in Vegas 9.0e to help me look at the Video Scopes in Vegas Pro.

Now, to figure out what it all means...

My main system:
Motherboard: Asus X99-AII
CPU: Intel i7-6850K
GPU: Sapphire Radeon RX480-8GB
RAM: Corsair Dominator (4 x 4 GB) DDR4 2400
Disk O/S & Programs: Intel SSD 750 (400 GB)
Disk Active Projects: 1TB & 2TB WD BLACK SN750 NVMe Internal PCI Express 3.0 x4 Solid State Drives
Disk Other: WD Ultrastar/Hitachi Hard Drives: WDBBUR0080BNC-WRSN, HGST HUH728080ALE600, 724040ALE640, HDS3020BLA642
Case: LIAN LI PC-90 Black Aluminum ATX Full Tower Case
CPU cooling: Corsair Hydro series H115i
Power supply: SeaSonic SS-750KM3 750W 80 PLUS GOLD Certified Full Modular Active PFC Power Supply
Drive Bay: Kingwin KF-256-BK 2.5" and 3.5" Trayless Hot Swap Rack with USB 3
Sound card: Crystal Sound 3 on motherboard. Recording done on another system.
Primary Monitor: Asus ProArt PA248q (24" 1920 x 1200)
O/S: Windows 10 Pro 22H2, Build 19045.2130

Camera: Sony RX10 Model IV

https://www.youtube.com/user/thedennischannel

Grazie wrote on 1/30/2014, 3:10 PM
That's great info MV10!! I just learnt heaps. Thanks for taking the time.

Grazie

larry-peter wrote on 1/30/2014, 3:49 PM
This is a good thread. I thank you too, MV10.
On the topic of skin tones, I mentioned in another thread that I think Caucasian skin is difficult for video to capture. I have shot chip charts along with talent and even after color vectors were aligned perfectly, still had to tweak skin tone. Well, to qualify a bit, for talent not using makeup skintone was difficult to get right. Most makeup removes the subtleties that make skin look alive.
Like to hear what others think. Of the three HD cameras I've had, I felt Panasonic was the worst at natural skin, Sony second, and JVC came closest to reality (GY110 CCD camera.)
robwood wrote on 1/30/2014, 5:11 PM
wow, good cc exercise!

had to increase saturation before making the choice
(i do this a ;
anyway, this is what i saw and why.

3 cyan - shadows on skin, washed out skin (cyan nulling red)
2 red - eyes, but mostly because i'd already picked the other 5
1 green - eyebrows / hair
4 blue - eyebrows / hair
5 yellow - skintone
6 magenta - right cheekbone
DiDequ wrote on 1/31/2014, 2:56 AM
This question is a trap as most of us know that :
- some colors look more saturated when near some other colors.
- some colors look less saturated when near some other colors.
- somme colors are not affected by any other colors.

Yes, musicvid10 : That's why we need a reference.

It is a little bit easier if we ask the question like this (having a neutral gray around each picture)

Above picture is the original







Generally speaking, people from the USA will give different results from Europeans or Asian people. (This does not mean YOU american will not give the same results as I european would)
And as we have no icc profile in our eyes, the colors I can see are not the colors you will see.

Again, yes, we need a reference.
We have the neutral reference in my way of asking that question is not enough but better than nothing.

Didier.
musicvid10 wrote on 1/31/2014, 8:21 AM
No, the task is to compare each example to the original, the way it's done in production. The task is not to compare each example to a gray canvas.
That is almost impossible the way you've layed them out. In fact, the only example that has a chance to make a fair comparison, is the top one. And once again, bringing up icc profiles is irrelevant to the discussion.

And my reason for not placing images from the same side of the spectrum adjacent to each other should be obvious.

So now that you've circumvented the purpose of the exercise, your answers are?
DiDequ wrote on 1/31/2014, 9:24 AM
musicvid, Ok, I shall do 6 pictures
my picture 1 beside my reference picture
my picture 3 beside my reference picture
my picture 5 beside my reference picture
and so on.
And tomorrow afternoon, french time, I shall answer (nobody answered yet)
I promise I shall not use any color picker / tool to check my answers.
So, please wait a little bit before posting the results .

If I have fully understood, you want us to say
1 C>M>J>N>R>V>B
2 C>M>J>N>R>B>V
3 ...
4
5
6
for example (no this is not my answer for 1 and 2)

Why I do not want to use YOUR picture ?
because I can be influenced by other pictures, this is what I tried to explain.
When you put a color beside another one, you might think it is a different color.
The only way not to be influenced is to have a neutral gray as background / background display between each pictures.

Didier.
larry-peter wrote on 1/31/2014, 11:09 AM
I guess we all have different methods of color correction, but mine has always been similar to the optometrist test, "Which looks better, A or B, C or D?" Musicvid's test method appeals to the way I work, and gives some good insight even for those who prefer to throw up an image and just correct away.
Tim Stannard wrote on 1/31/2014, 12:37 PM
When do we ge the correct answer? FWIW I've always struggled with colour, but without reference to anything else (and without reading ahead) I figured the same as atom12 and robwood.
However, this task was made somewhat easier by the knowledge that here was one of each. A trickier task would be to be given the reference and just ONE of the samples to test. Not only would this be trickier (IMO) but it would be much more akin to a real world situation where you have two images (perhaps from different cameras or under slightly different lighting conditions) to match.
musicvid10 wrote on 1/31/2014, 12:43 PM
didler,
The test is fairly simple.
Each of the examples varies by ONE color from the reference image. So there is only one answer for each of the six examples :?o
Read the directions and follow the examples of the people who have already answered. And please don't read too much into it, OK?

tim,
Give it a couple of more days. I'll post the answers along with one deductive approach to getting there.

In production, we use the ringaround to match the first test print out of the processor as closely as possible with one of the examples. Of course there are in-between colors, and that will be the subject of Part II of this topic.
john_dennis wrote on 1/31/2014, 3:04 PM
O.K. I'm going to get my ignorance on record. Something I used to try to do early in college so I could get the shame out of the way and move on.



I'll show my work.

From my indecision and the time I spent on it, I would starve to death if this is what I did for a living. I do want to learn about the subject, though.

My main system:
Motherboard: Asus X99-AII
CPU: Intel i7-6850K
GPU: Sapphire Radeon RX480-8GB
RAM: Corsair Dominator (4 x 4 GB) DDR4 2400
Disk O/S & Programs: Intel SSD 750 (400 GB)
Disk Active Projects: 1TB & 2TB WD BLACK SN750 NVMe Internal PCI Express 3.0 x4 Solid State Drives
Disk Other: WD Ultrastar/Hitachi Hard Drives: WDBBUR0080BNC-WRSN, HGST HUH728080ALE600, 724040ALE640, HDS3020BLA642
Case: LIAN LI PC-90 Black Aluminum ATX Full Tower Case
CPU cooling: Corsair Hydro series H115i
Power supply: SeaSonic SS-750KM3 750W 80 PLUS GOLD Certified Full Modular Active PFC Power Supply
Drive Bay: Kingwin KF-256-BK 2.5" and 3.5" Trayless Hot Swap Rack with USB 3
Sound card: Crystal Sound 3 on motherboard. Recording done on another system.
Primary Monitor: Asus ProArt PA248q (24" 1920 x 1200)
O/S: Windows 10 Pro 22H2, Build 19045.2130

Camera: Sony RX10 Model IV

https://www.youtube.com/user/thedennischannel

Stringer wrote on 1/31/2014, 7:52 PM
The ear jewelry seems to provide a key .. Will see if I am wrong.

I'm definitely not a pro, and have been biased by earlier answers...
musicvid10 wrote on 1/31/2014, 9:27 PM
Stringer,
Yes, the glass ear jewelry is a good neutral reference.
Going back to 1970's color correction theory, the whites of the eyes provide a good cool/neutral reference point, because they are pretty consistent among individuals and races. Age and health are factors, though.
DiDequ wrote on 2/1/2014, 9:09 AM
And here is my suggestion.
1 Red / 2 Red / 3 Red / 4 Red / 5 Red / 6 Red.
Answer based on my printing experience where all colors are reproduced on 4 monochromatic plates.
When you see a human face, on those plates, you are not influenced by colors. Please do not laugh if I'm completely wrong - I have not used any tool to help me.
This is probably why you only show the lady's face on pictures 1 to 6, because you did not want us to see the background.

I did work with people that knew how many % change for each color they had to apply to get a given result. I have never been able to do this !
They would have answered: 1 X 12% / 2 Y 5 %, etc... (12 and 5 choosen to illustrate) and would have passed your test.

Didier.
musicvid10 wrote on 2/1/2014, 10:47 AM
The correct order is G, R, C, B, Y, M.
Although atom 12 was the first, several got it correct independently.
I'm sure everyone used a different rationale, but here is one "classroom" example:The main components of caucasion skin tones are red and yellow. So let's concentrate on those in comparison to the reference, since they are easily identifiable, and most people will recognize them easily.



The most common mistakes in correcting Caucasion and Asian flesh are mistaking green for yellow (1 and 5), and magenta for red (6 and 2). That is because the natural red and yellow components can mask cyan and blue bias, thus the need for at least one other near-neutral reference to get our brains on track. As mentioned, the whites of the eyes are often the secondary reference, and usually register as slightly cyan/blue. Fabrics can be tricky as references, because of spectral reflection / absorption that registers differently than it looks to the eyes.

Now, one might ask, how would it look if the error was green and yellow, or magenta and red, or some other combination, rather that a pure primary additive or subtractive? These are the tricky situations that can baffle the most experienced colorists, and where a second neutral or "memory" color in the scene is an absolute necessity. That exercise will be in the next round of discussion, coming soon . . .
;?)

Remember, the additive primaries are red, green, and blue, not what they taught you in second grade (yellow light is a subtractive primary).



Note: I was hired straight out of college as a QC tech by Technicolor. That was in 1970. I was field-trained by Frank Ishihara, one of the legendary greats in the business, whose name still gets mentioned in phototechnical circles. I continued in the industry, interspersed with long road tours as a musician, until 1989, when I was benched by occupational illness (systemic developer poisoning). Following major surgery in 1990, I never returned to the industry, but continue to scan, print, and edit in Photoshop and Vegas in my home office. I've taught math and music, and run my own nonprofit musical theater company since retiring.

larry-peter wrote on 2/1/2014, 12:04 PM
This thread really took me back to the time I was learning the trade. My first job as an editor in the 80s involved hand carrying our negatives to Atlanta or Dallas for transfer to tape. There was one colorist who wasn't annoyed by my constant questions and told me when adjusting skin tones to always look at the transition area between highlight and shadow. In this case it was the area between the cheekbone and shadow in front of the earring. I wasn't certain of my answers, but those were the prominent colors I saw in that area.
DiDequ wrote on 2/2/2014, 2:39 AM
So, I was completely wrong, probably because I did not really understand the question.
I checked with Gimp and this tool :

You can only use The R settings to get something like pictures 1 to 6, but as I promised not to use any tool...
There are some color questions related to colors eating some others but here, there was no "trap" (piège)
because for example when you do a R setting, your eyes wil think green is different (Red eats green). I I I mentionned in my first post here.
I always try to make things difficult before going back to basics. Good lesson.
Grazie wrote on 2/2/2014, 6:04 AM
Musicvid10 : I was hired straight out of college as a QC tech by Technicolor.That's some fine pedigree. We are blessed to have you here on the VP Forum.

Cheers

Grazie

musicvid10 wrote on 2/2/2014, 7:57 AM
Pedigreed mutt, I would say. Just happened to apply at the right time.
If only it was that easy these days.

Care to try your hand at the next round? You have excellent color vision.