APPLE FINAL CUT PRO: Apple Final Cut Pro X FCPX Debates FCP Legacy FCP Tutorials

The Color Wheels Mystery - Demystified

COW Forums : Apple Final Cut Pro X Debates

<< PREVIOUS   •   VIEW ALL   •   PRINT   •   NEXT >>
Simon Ubsdell
The Color Wheels Mystery - Demystified
on Jan 5, 2018 at 1:27:30 pm
Last Edited By Simon Ubsdell on Jan 5, 2018 at 1:38:50 pm

Here is a follow-up video that brings together some of the points Oliver and I have tried to explain on this forum and also offers a clearer explanation of the how and why of what we are seeing.







Edit: Something I keep forgetting to mention is that although I have only looked at the Y values (what Apple in typically eccentric fashion is calling "Brightness" here), the same considerations apply to the RGB component of the color wheels. And that has important consequences for using the wheels in Rec.709 for color grading. But I didn't want to over-complicate the story at this point.

Simon Ubsdell
tokyo productions
hawaiki


Return to posts index

Jeremy Garchow
Re: The Color Wheels Mystery - Demystified
on Jan 5, 2018 at 4:51:15 pm

So why add an offset to the value? How is that good for the user?

Couldn't the Master be used as a better offset control instead?


Return to posts index

Simon Ubsdell
Re: The Color Wheels Mystery - Demystified
on Jan 5, 2018 at 5:00:54 pm

[Jeremy Garchow] "So why add an offset to the value? How is that good for the user?"

I was using the "offset" there as an illustration for why the mapping doesn't match between Rec.709 and Rec.2020. It's not a true value ... and I haven't given enough thought yet to what it would actually be.

I'm not saying that there is a control offset component in this. It's just that the Wheels are reading the input pixel values AS IF they had an offset applied, and that is the root cause of what we are seeing.

[Jeremy Garchow] "Couldn't the Master be used as a better offset control instead?"

The Master is indeed a standard Offset control or Brightness control, which is the exact same thing by a different name.

The maths for that is simply: pixelValue+controlValue

I think I might have managed to confuse you by my use of the term "offset" (though I did use it correctly).

... Can you expand on your second question because I'm not sure how you are meaning it to apply to this?

Simon Ubsdell
tokyo productions
hawaiki


Return to posts index


Jeremy Garchow
Re: The Color Wheels Mystery - Demystified
on Jan 5, 2018 at 5:43:56 pm

First, thanks for attempting to demystify this mystery.

[Simon Ubsdell] "... Can you expand on your second question because I'm not sure how you are meaning it to apply to this?"


My direct question is, why do it this way? :)


Return to posts index

Simon Ubsdell
Re: The Color Wheels Mystery - Demystified
on Jan 5, 2018 at 5:50:11 pm

[Jeremy Garchow] "My direct question is, why do it this way? :)"

To that, my friend, I have no answer ☹

Or rather, I need to make clear that they didn't "do something funny".

This is not some exotic color science that they have actively decided to implement. (Sorry, Bill, but it really isn't this time.)

They simply decided (?) not to normalise the Color Wheels to the Rec.709 scale. Which is to say the the Color Wheels are effectively seeing the "wrong" input values.

(Just as the Color Board is seeing the "wrong" input values in Rec.2020 for the inverse reason.)

Simon Ubsdell
tokyo productions
hawaiki


Return to posts index

Jeremy Garchow
Re: The Color Wheels Mystery - Demystified
on Jan 5, 2018 at 7:17:19 pm

[Simon Ubsdell] "Or rather, I need to make clear that they didn't "do something funny"."

And by funny, you mean clever?

I know you said not to use the word “bug” but doesn’t this feel like a mistyped or misread or misused calculation?


Return to posts index


Simon Ubsdell
Re: The Color Wheels Mystery - Demystified
on Jan 5, 2018 at 7:41:22 pm
Last Edited By Simon Ubsdell on Jan 5, 2018 at 7:43:33 pm

[Jeremy Garchow] "I know you said not to use the word “bug” but doesn’t this feel like a mistyped or misread or misused calculation?"

That's a very good question, I think.

The word "bug" can cover so many different things, from simple mistyping, as you say, up to much more fundamental design decisions.

I have a problem with calling the latter "bugs", but that's a personal thing.

This is where we need our good friend, Walter.

In this case, I would say there is not even the remotest chance that this is a mistyped line of code.

I simply cannot believe that the team were not aware of the results. My impression was that they thought this would be "good enough".

But while I can troubleshoot the problem, I simply have no idea how they arrived at the decision to let this through. That's a people thing ...

Simon Ubsdell
tokyo productions
hawaiki


Return to posts index

Walter Soyka
Re: The Color Wheels Mystery - Demystified
on Jan 5, 2018 at 9:55:26 pm

[Simon Ubsdell] "The word "bug" can cover so many different things, from simple mistyping, as you say, up to much more fundamental design decisions. I have a problem with calling the latter "bugs", but that's a personal thing. This is where we need our good friend, Walter."

I'll go one step beyond that. Our little studio has been doing a good bit of development work, and I have come to hate the word "bug" for anything. It's the connotation. Bugs sound external. Who would actually put a bug into something? They just crawl in on their own.

I think the term "bug" lets the developer off the hook. I prefer the words "defect" or "flaw."

While "bug" deflects blame, "defect" states simply and matter-of-factly that there's a problem with the product. Defects don't have agency, and they don't spring into being through abiogenesis. We create them, and if we care about quality, we must find and repair them.

Simon, these threads remind me of our conversations about the FCPX light wrap [link] and the FCPX keyer [link] -- it sounds like an engineering/QA miss.

Walter Soyka
Designer & Mad Scientist at Keen Live [link]
Motion Graphics, Widescreen Events, Presentation Design, and Consulting
@keenlive [twitter]   |   RenderBreak [blog]   |   Profile [LinkedIn]


Return to posts index

Simon Ubsdell
Re: The Color Wheels Mystery - Demystified
on Jan 5, 2018 at 10:00:10 pm

[Walter Soyka] "I'll go one step beyond that. Our little studio has been doing a good bit of development work, and I have come to hate the word "bug" for anything. It's the connotation. Bugs sound external. Who would actually put a bug into something? They just crawl in on their own. "

I couldn't agree more.

(It's great to have your input after all this time. How we have missed it!)

Simon Ubsdell
tokyo productions
hawaiki


Return to posts index


Jeremy Garchow
Re: The Color Wheels Mystery - Demystified
on Jan 5, 2018 at 10:48:14 pm

[Walter Soyka] "I think the term "bug" lets the developer off the hook. I prefer the words "defect" or "flaw.""

Perfect.

So, the FCPX Color Wheels and Color Board are flawed depending on what color space the Library is in.

They also appear to have a defect in the way values are read and/or calculated.

The term I like the best? Busted.

The FCPX color wheels are busted. :-D


Return to posts index

Michael Gissing
Re: The Color Wheels Mystery - Demystified
on Jan 5, 2018 at 10:52:21 pm

[Walter Soyka] "I think the term "bug" lets the developer off the hook. I prefer the words "defect" or "flaw."

I always imagined it as short for 'buggered'. As in "we really buggered up the code here". Maybe that's just an Australian sensibility where we use words like buggered to mean lots of different things.


Return to posts index

Nick Meyers
Re: The Color Wheels Mystery - Demystified
on Jan 5, 2018 at 11:15:48 pm

[Michael Gissing] "Maybe that's just an Australian sensibility where we use words like buggered to mean lots of different things"

and then abbreviate them !


n


Return to posts index


Simon Ubsdell
Re: The Color Wheels Mystery - Demystified
on Jan 5, 2018 at 7:08:41 pm

I just wanted to point out that if you're eagle-eyed with your maths, you will have spotted that I didn't complete the "offset" calculation, which should have read like this:

pow(x+.1, 1/gamma)-pow(.1, gamma)

It doesn't make any difference to the validity of the demonstration. (I have used 0.1 as the variable but, as I have mentioned, I am guessing at that - again, for the purposes of the demonstration it's close enough.)

I just wanted to keep things simple enough not to confuse - if possible. (Not sure I've managed that, but I tried!)

Simon Ubsdell
tokyo productions
hawaiki


Return to posts index

Michael Gissing
Re: The Color Wheels Mystery - Demystified
on Jan 5, 2018 at 11:17:42 pm

And thanks Simon for taking the time and effort to present this.


Return to posts index

Shawn Miller
Re: The Color Wheels Mystery - Demystified
on Jan 6, 2018 at 12:39:04 am

[Michael Gissing] "And thanks Simon for taking the time and effort to present this."

+1 - I don't have a horse in the race, but it's still really interesting stuff.

Shawn



Return to posts index


Simon Ubsdell
Re: The Color Wheels Mystery - Demystified
on Jan 6, 2018 at 11:17:25 am
Last Edited By Simon Ubsdell on Jan 6, 2018 at 11:44:34 am

[Shawn Miller] "it's still really interesting stuff."

Thanks, Shawn and Michael.

But the credit really goes to Oliver who not only identified that there was something worth investigating in the first place, but then subsequently came up with the explanation as well.

All I have tried to do is join up the dots in between. And hopefully add some interesting background.

I just love the breathtaking power and simplicity of these basic pixel functions and I hope I have managed to communicate some of that.

Where would we be without the "magical" properties of One and Zero?

Simon Ubsdell
tokyo productions
hawaiki


Return to posts index

Michael Gissing
Re: The Color Wheels Mystery - Demystified
on Jan 6, 2018 at 12:08:07 pm

[Simon Ubsdell] "Where would we be without the "magical" properties of One and Zero?"

As long as you don't divide by zero.


Return to posts index

Simon Ubsdell
Re: The Color Wheels Mystery - Demystified
on Jan 6, 2018 at 12:18:17 pm
Last Edited By Simon Ubsdell on Jan 6, 2018 at 2:43:20 pm

[Michael Gissing] "As long as you don't divide by zero."

Indeed. Suddenly the magic turns sour!



Simon Ubsdell
tokyo productions
hawaiki


Return to posts index


Herb Sevush
Re: The Color Wheels Mystery - Demystified
on Jan 7, 2018 at 3:35:02 pm

[Simon Ubsdell] "I just love the breathtaking power and simplicity of these basic pixel functions and I hope I have managed to communicate some of that. "

I have learned more about the design and workings of the color wheel from these threads than in all I had read previously.

Thank you, and Oliver.

Herb Sevush
Zebra Productions
---------------------------
nothin' attached to nothin'
"Deciding the spine is the process of editing" F. Bieberkopf


Return to posts index

Jeremy Garchow
Re: The Color Wheels Mystery - Demystified
on Jan 7, 2018 at 5:31:16 pm

[Herb Sevush] "I have learned more about the design and workings of the color wheel from these threads than in all I had read previously.
"


This is true of almost any of Simon’s videos. Amidst the “how this button works”, there’s lessons, real lessons, about technique of the broader subject. Watching any of Simon’s tutorials on Fusion or Motion brings a lot of insight in to compositing. Real technique and knowledge that can travel with you, no matter what program you use, or level of interest in compositing (or color correction, or keying/matte generation) which is why I find these conversations so valuable as it is a discussion of root issues that could benefit anyone who is interested and follows along.


Return to posts index

Walter Soyka
Re: The Color Wheels Mystery - Demystified
on Jan 7, 2018 at 7:20:24 pm

[Jeremy Garchow] "Amidst the “how this button works”, there’s lessons, real lessons, about technique of the broader subject. Watching any of Simon’s tutorials on Fusion or Motion brings a lot of insight in to compositing. Real technique and knowledge that can travel with you, no matter what program you use"

+1000. Simon is a really excellent teacher, and he's so open with his knowledge.

Thanks to both Simon and Oliver for making this forum so interesting to read (again).

Walter Soyka
Designer & Mad Scientist at Keen Live [link]
Motion Graphics, Widescreen Events, Presentation Design, and Consulting
@keenlive [twitter]   |   RenderBreak [blog]   |   Profile [LinkedIn]


Return to posts index


Simon Ubsdell
Re: The Color Wheels Mystery - Demystified
on Jan 7, 2018 at 8:58:44 pm
Last Edited By Simon Ubsdell on Jan 8, 2018 at 11:54:52 am

Thank you so much, Jeremy and Walter and Steve, for your kind words.

Tim would not forgive me if I failed to point out that I only started out on the journey of making tutorials because of all of you here and the fascinating discussions we have had on this very forum dating back quite a few years now.

There is nothing remotely as interesting as a good question and that is what this forum has so often thrown up.

So my thanks are owed to you. When this forum is good, it is very, very good.

Simon Ubsdell
tokyo productions
hawaiki


Return to posts index

Andrew Kimery
Re: The Color Wheels Mystery - Demystified
on Jan 9, 2018 at 5:47:11 am

Yes, thank you to everyone that's contributed to this discussion. You all have taken something that can be pretty complex and broken it down so that the info is accessible to non-experts as well as experts. And that's not always an easy thing to.


Return to posts index

Aindreas Gallagher
Re: The Color Wheels Mystery - Demystified
on Feb 7, 2018 at 10:16:40 pm

Reading through all of this was worth it, if only to watch Bill Davies stick his fingers in his ears and scream "I CAN'T HEAR YOU, HAVE YOU TRIED XINTWO MY MARVELLOUS TRAINING SERIES."

Good times guys, good times. Well worth the dip back in. Also that CC colour science operation is clearly fubar yes? Yes.

https://vimeo.com/user1590967
producer/editor.grading/motion graphics


Return to posts index


Michael Gissing
Re: The Color Wheels Mystery - Demystified
on Feb 9, 2018 at 1:11:32 am

[Aindreas Gallagher] "Good times guys, good times. Well worth the dip back in."

Ah we've missed your epic sprays Aindreas. Good times indeed.


Return to posts index

Steve Connor
Re: The Color Wheels Mystery - Demystified
on Jan 7, 2018 at 7:21:41 pm

Can I just add it's great to see Simon back, I was about to stop visiting but this debate and the return of Tim has kept me here :)

"Traditional NLEs have timelines. FCPX has storylines" W.Soyka


Return to posts index

Claude Lyneis
Re: The Color Wheels Mystery - Demystified
on Jan 10, 2018 at 6:57:10 am

Beautiful demonstration of the math right and wrong behind the gamma curves in the color wheels. Obviously not a bug. As a professor of physics at Stanford said many years ago, that is not a mistake it is a blunder. I am hoping Apple sees the light and fixes it, because the color wheels have great promise and are much more convenient to use than the available plugins.
Thanks Simon. This is the best explanation I have seen as to how gamma should work. For me the math is easier to grasp than to do great looking color grades.


Return to posts index


Jeremy Garchow
Re: The Color Wheels Mystery - Demystified
on Jan 10, 2018 at 8:00:17 pm

[Claude Lyneis] "Obviously not a bug. As a professor of physics at Stanford said many years ago, that is not a mistake it is a blunder."

I guess I would group "bug" in to "blunder", but I guess that is incorrect nomenclature.

I still feel like this particular problem is something that was overlooked and not QC'd.

I'm not sure what the 'proper' terminology would be, but I know as a user, I would file a bug report as it is something that behaves correctly sometimes, but not other times. Obviously, as nearly everyone agrees, it has nothing to do with full and legal RGB ranges as the Facebook crew seemed to espouse.

Wikipedia describes a bug as an error, of either human or computer, in a program:

"A software bug is an error, flaw, failure or fault in a computer program or system that causes it to produce an incorrect or unexpected result, or to behave in unintended ways...Most bugs arise from mistakes and errors made in either a program's source code or its design, or in components and operating systems used by such programs. A few are caused by compilers producing incorrect code. A program that contains a large number of bugs, and/or bugs that seriously interfere with its functionality, is said to be buggy (defective). Bugs can trigger errors that may have ripple effects. Bugs may have subtle effects or cause the program to crash or freeze the computer. Other bugs qualify as security bugs and might, for example, enable a malicious user to bypass access controls in order to obtain unauthorized privileges."

I would guess that we would never have any bugs if all software problems were "simple" blunders, mistakes, or miscalculations. I guess "Bug reports" if using the tall tale origin of the word "bug" should be saved for hardware, where a bug literally crawls in to the system and short circuits the entire operation? Does this mean we should change what we know as "bug reports" to "blunder reports"?

https://developer.apple.com/bug-reporting/
http://www.adobe.com/products/wishform.html

The magnetic timeline is a decision. The skimmer is a decision. Roles and other tags are decisions. Having a compound clip the grinds the system to a halt, as in the very early days of FCPX, is not a decision, that's a bug or at least what I would call a bug.

Having luma values that scale out of a typical range in one color space, but not another while using the same tool seems like it would not be a decision. It would be a "bug" that produces an incorrect result depending on what tool and color space (color wheels or color board, 709 or 2020PQ). I would also think it would be something that Apple (the landlord?) would fix in an update.

Bill Davis, if you want to see the RGB range in action look here, please note not a color wheel was touched: https://forums.creativecow.net/thread/344/49874


Return to posts index

Claude Lyneis
Re: The Color Wheels Mystery - Demystified
on Jan 10, 2018 at 10:13:32 pm

Jeremy:
Sorry if I didn't take the discussion of whether it is properly labeled a bug or not too seriously. Whatever it is labeled it needs to be fixed by Apple and I think Simon's explanation of what is going wrong is crystal clear. Shouldn't take too much effort for it to be sorted out. This is not rocket science.

So, bug, blunder, mistake whatever, I hope Apple fixes it soon.


Return to posts index

Jeremy Garchow
Re: The Color Wheels Mystery - Demystified
on Jan 11, 2018 at 12:16:34 am
Last Edited By Jeremy Garchow on Jan 11, 2018 at 12:21:25 am

[Claude Lyneis] "Simon's explanation of what is going wrong is crystal clear."

Absolutely.

I was just trying to divert this discussion in to another debate, as always happens in this forum, to:

What is a bug if it’s not a bug? :)


Return to posts index


Shawn Miller
Re: The Color Wheels Mystery - Demystified
on Jan 11, 2018 at 1:52:31 am

[Jeremy Garchow] "What is a bug if it’s not a bug? :)"

An undocumented feature. ☺

Shawn



Return to posts index

greg janza
Re: The Color Wheels Mystery - Demystified
on Jan 11, 2018 at 4:35:05 pm

As someone who has been following this thread simply out of curiosity, I want to thank Simon ( and Oliver) for the very clarifying information and explanations of the FCPX color wheel anomaly.

This demonstrates how helpful these forums can be for the entire professional community.

I Hate Television. I Hate It As Much As Peanuts. But I Can’t Stop Eating Peanuts.
- Orson Welles


Return to posts index

Bill Davis
Re: The Color Wheels Mystery - Demystified
on Jan 11, 2018 at 8:31:47 pm

And please don't forget you have a DIRECT LINE to the Apple FCP X Team in the form of the report plumbed directly into the FCP X menu bar.

If this bothers you, I would expect you to have reported that fact directly to the X development team.

The more reports that arrive at the OFFICIAL problem reporting system, the faster things get fixed.

Let your voice add to others and be heard.

Creator of XinTwo - http://www.xintwo.com
The shortest path to FCP X mastery.


Return to posts index


Mark Smith
Re: The Color Wheels Mystery - Demystified
on Feb 5, 2018 at 7:59:49 pm

I was messing around with some R3d files I shot over the weekend prior to setting a library for the project. I imported a couple of R3d files into my sandbox project and was making a few color adjustments and I noticed tht color wheels worked as expected with redcode 4 files. I was a little surprised and had to make sure my eyes weren't deceiving me, but the color wheels in 10.4 played nice with my footage.


Return to posts index

Michael Gissing
Re: The Color Wheels Mystery - Demystified
on Feb 5, 2018 at 11:19:32 pm

[Mark Smith] "...and I noticed tht color wheels worked as expected with redcode 4 files."

Was your color space Rec 709 or Rec 2020? The bug is in Rec 709. Rec 2020 works as expected.


Return to posts index

Mark Smith
Re: The Color Wheels Mystery - Demystified
on Feb 6, 2018 at 3:09:22 am

I understand about the Rec2020 color space and color wheels working correctly. What stands out for me is that I didn’t set the project up as Rec 2020 . I imported a couple of my Red clips into my sandbox library, Picked on and threw it in a new project without specifying Rec 2020 wide gamut color. When I went to explore rough grading, color wheels worked as they should which I found surprising, so maybe FCPX is getting some info from Redcode 4 files that filters into the project setting behind the scenes.


Return to posts index

<< PREVIOUS   •   VIEW ALL   •   PRINT   •   NEXT >>
© 2018 CreativeCOW.net All Rights Reserved
[TOP]