FCP 10.4: Color Boards vs. Wheels

Not being a colorist (poor score in the Pantone test on color vision accuracy), I had ingested the whole manuals and bought the whole Ripple/Lynda video trainings for both Apple Color and Resolve 9. Resolve was my tool of choice if I wanted to really polish a video. Over the years though, I came to like the seemingly ridiculous Color Board of FCP X, and for most tasks it was sufficient- and fast.

With the Color Wheels, I thought I could make them my primary CC tool (cmd + 6). I found it difficult to get similar results in the same time with them, compared to the Color Board. So I thought I just had to get used to them. But a few minutes ago, I stumbled over this clip by Simon Ubsdell:


He admits he isn't a colorist either. Is this a bug? One comment says, no, highlight-, shadow- and midtone wheels are not supposed to do the same as lift, gamma, gain.

So whoever is right, the least we can say is that the Color Board is not just another representation of the same functions. And that it obviously remains to be number one for primary CC. Can anyone explain?

What's going on?
The video now doesn't exist anymore???

It demonstrated that by rising the highlights only, you would also raise the shadows and vice versa. And that midtones as well affected both shadows and highlights and barely produced a gamma curve. The control behaved almost like the MASTER (=offset).

Has Apple contacted Simon and urged him to delete the clip?

Crazy!

I see some heads rolling in Cupertino ...

In a german fcp forum someone has saved the contradicting comment from the Youtube side. Here it is:

Lift, Gamma and Gain (LGG) have NOTHING to do with Shadows, Midtones and Highlights (SMH). He is correct in his definition of LGG but his assumptions about SMH are misguided. What's happening with SMH is that there is an ADDITIONAL weighing factor (a transfer function) that splits the Luma range into three distinct ranges that sum up to 1 (think of the three ranges as luma masks). What you're seeing when adjusting the the Luma level in the SMH wheels is an OFFSET (additive) transformation weighed by a Luma xfer function. The behaviour is absolutely normal. If the effect was the same, why would you need both the color board and the wheels ? These are two different methods. The primary use of the SMH wheels is to colour balance luma ranges differentially. No bug there and no reason to be alarmed. I would add that it is WRONG to think of Lift, Gamma and Gain as methods to control Shadows, Midtones and Highlights; they are more appropriately, methods to control CONTRAST, where the contrast pivot is either pegged to black, white or both. Gamma can also double up as a classic "brightness" control, that doesn't clip or dim specular whites and doesn't lift or crush blacks. The xfer functions used to split the luma range into SMH can have many shapes. What you're seeing with the curve is simply a reflection of the SMH weighing functions chosen by the FCP team. These are valid choices, but not the only ones possible. But still perfectly valid.

Be it as it may, then obviously everybody has missed the fact that the wheels are of NO USE to efficiently adjust contrast, and our trusted tutors on FCP X (all of them, as far as I can see) have missed this. See here how Spencer (2'31") adjusts the highlights and (watch the RGB parade!) destroys his shadow adjustment whilst doing so:



I don't believe in this explanation. Of what use is any selection, if there is little limitation to it? Ubsdell was right.

A small glitch, still a very useful tool, doesn't stop you from achieving your goals quickly. Yes, a small glitch. No, not a huge deal killer.

And yes, Lift and Gain are equivalent to Shadow and Highlight, just with a different algorithm. Mids and Gamma are somewhat similar, but definitely not "the same". Midtones control works the center of the luma range, where Gamma creates a curve favored to one end of the spectrum or the other.

Axel wrote: What's going on?
The video now doesn't exist anymore???
… ...

no, still alive in Simon's account:
youtube.com/watch?v=8lU35F1Q2Q4

he had switched off the commentaries … probably the topic 'LGG vs SMH' should be discussed first …

Karsten Schlüter wrote:

Axel wrote: What's going on?
The video now doesn't exist anymore???
… ...

no, still alive in Simon's account:
youtube.com/watch?v=8lU35F1Q2Q4

he had switched off the commentaries … probably the topic 'LGG vs SMH' should be discussed first …

The video was offline a few hours yesterday. Simon deleted the comments and added the line:
(Hint: It has nothing to do with the SMH model.)

The SMH-model vs LGG is explained here . And Simon is right: in a correct SMH-model (I believe the old 3-way Color Corrector of FCS, labelled Blacks, Mids and Whites, used that, it had additional white-point and black-point controls for adjusting contrast - I may err though), the shadows would even be less affected by any highlight changes, that is: not at all!

The whole thing would make some sense if the Color Wheels behaved like the LOG Wheels in Resolve:


I'm not sure if this exactly describes the SMH model. Can someone comment?

The issue with the wheels only applies to 709 projects. It works normally for 2020. Stay away from the wheels if you are grading in 709. Or else with the next FCP update you may have lost all your corrections for existing projects!

Vow! This is really quite unexpected. I wonder what will be in 10.4.1 - are they going to change the wheels or the colour board, to match the other? Considering that the board has now been out for years, I would think they should not be changed otherwise an awful lot of projects are going to be impacted.

I am not only NOT a colourist, but I am also not very experienced, so this is just my perception:

I saw the exact behaviour as described - that is that Shadows wheel effects the 100% bright point and the Highlights wheel effects the black point. BUT, if that is not to your liking, then use the Colour Board, which works more as expected, or the Luma Curve. From a PERCEPTION point of view, I actually like the behaviour of the wheels, even though they are not what is commonly accepted. I think this would be more like Lightroom for stills.

CoreMelt's Chromatic works traditionally, whereas Color Finale Pro does so for Lift but more like FCPX Wheels for Gain! So go figure!

I think ultimately the user will just need to decide which method they like the best, but it is very good to bring this to light!

I've just done a little experiment. I took a clip and put on it all the various colour tools and then unticked all. I then selected each, one at a time and then tried to see how easy/quick I could get a result that was pleasing to my eye. Clearly this is a subjective "test" and for only one clip, but this is what I found, in this case:

In order of preference, starting with the best results and descending to the least liked:

1. Color Wheels
2. Color Finale Pro
3. Color Board
4. Chromatic

Fascinating experiment. I would urge others to do something similar.

pszilard wrote: I've just done a little experiment. I took a clip and put on it all the various colour tools and then unticked all. I then selected each, one at a time and then tried to see how easy/quick I could get a result that was pleasing to my eye. Clearly this is a subjective "test" and for only one clip, but this is what I found, in this case:

In order of preference, starting with the best results and descending to the least liked:

1. Color Wheels
2. Color Finale Pro
3. Color Board
4. Chromatic

Fascinating experiment. I would urge others to do something similar.

Really, the whole discussion is a mess! Please, PLEASE, don't just trust me citing experts from Creative Cow. I did try the gradient in 2020, and for me the wheels still don't work as expected (like in Resolve). Right now*:

1. Color Board exclusively for fast CC (rarely need more).

or

1. Color Board for primaries.
2. HSL curves for secondaries (then in Resolve, now more in FCP).

(*subject to change)


There is a (small) chance that we are missing an ingenious new paradigm regarding the Color Wheels. Time will tell.

I would disagree that the discussion is a mess. We are discussing Apple's implementation of the color wheels, and I for one, did my own experiment to formulate my own pragmatic approach, comparing the four methods available to me.

I hadn't checked 2020, as it is not of significance to me, but I am interested to read about anyone's further analysis.

There is also the question: what is more important to an editor or colorist. Is it to have mathematically correct tools, or tools that make it easiest to create a visually pleasing result? For me, it's the latter, but I respect that others may think differently. It really comes down to choosing the tool/method that you like to do your job. Choice is good, as is an understanding of the tools on offer.

pszilard wrote: I would disagree that the discussion is a mess.

I made it a mess, because I added contradicting statements of others as soon as I found them elsewhere. That's okay if you take every "fact" I cite with a grain of salt.

pszilard wrote: It really comes down to choosing the tool/method that you like to do your job. Choice is good, as is an understanding of the tools on offer.

This!

This issue isn't stopping anyone from doing great color grading with the new tools. Some small tick in the function, well, I don't notice it when grading my TV shows. I still get great results.

FWIW, I find the color tools in Chromatic to be more sensitive and accurate than Finale. I also find the new color tools in FCPX to be far, far superior to either. But I'm keeping Chromatic for the trackable masks and balance tools I use a lot.

So, the two apps work different, but is anyone UNABLE to achieve a great grade with either?

Simon Ubsdell updates with math .
One might say the Color Board is designed for 709 and the Wheels for 2020 color spaces.

cseeman wrote: Alex4D updates with math .…

Simon did … credit where credit is due

And for those, like me, who mention of 'math' leads to NoGo territory: lots of illustrative graphs and curves. Love the comparison of DaVinci's SMH handling with an audio-equalizer … thought, this is how old Color-board is handling it (vertical sliders …) - Not so! Bummer …

so, actually we have, at last, three 'systems' how to handle colors … <head spinning>

And they all achieve the same result...

Hooray to answers!

I have a question. I tried to replicate Simon's experiment but I am unable to generate a gradient with liner waveform.
This is FCPX generator i.imgur.com/S03oR9I.png
and this is gradient from Affinity Photo i.imgur.com/rFkptW8.png
Is there any generator which generates a gradient with linear waveform? Thank you

Either it's a bug that 709 works with the color board and breaks in color wheels and 2020 does the opposite or they have not made clear the intention that you should use the wheels with 2020 and the board with 907. Hopefully this gets addressed soon before people do to much work correcting and it will change later on forcing you to redo your adjustments.

EDIT: I submitted feedback on this at apple.com/feedback and linked to Simon's video, hoping for a clarification or that the intent is more obvious in FCPX.