[Simon Ubsdell]

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

[Simon Ubsdell]

[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

[Simon Ubsdell]

[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

[Simon Ubsdell]

[Axel Schreiter] “Before having seen Simons video, I was already irritated that the new Color Wheels seemed to be less powerful than the pathetic Color Board. It’s like you say, I thought I must be missing some new ingenious paradigm. Is there one? I wish your color scientists could help us with a plain explanation.”

I hope this new video will help you understand the answer:

https://youtu.be/bESH7_p6aRk

Simon Ubsdell
tokyo productions
hawaiki

[Simon Ubsdell]

[Oliver Peters] “What isn’t getting mentioned a lot in a “wait and see” approach is the real world ramifications of the issue.”

To expand on that a bit in relation to the RGB performance of the wheels, here is the Color Board result when pushing the midtones of the greyscale test ramp to yellow:

Entirely as predicted by the conventional model.

And here is what the Color Wheels do with the same correction:

You can see red “whites” being pushed above one (100%) and the green and blue “blacks” being pushed below zero, very heavily so in the case of the blue. Similarly the green and blue “whites” are reduced, the blue dramatically so, rather than being pegged at 100% at the top of the curve.

I think it’s fair to say that this is not really desirable.

You can start to get back to the desired result by pulling the “Shadows” and “Highlights” away from yellow, but that’s two extra corrections to get to the same place, and even then it’s not exactly the same place.

Incidentally, this is not a separate issue to the one we have been highlighting up to now – it’s merely the repercussions of the exact same issue as they affect the RGB.

That is to say, there is one answer to this single mystery, not several separate mysteries. I think …

Edit: There is an added complication when dealing with RGB in that we need to factor in the luma coefficients but it’s best not to add more difficulty to the explanation at this point. Suffice it to say that because green, red and blue are not equally “bright” (to use a non-technical term for it) RGB corrections need to factor in those relative “brightness” values and if the input pixel values are not being correctly interpreted (as is the case with the color wheels in Rec.709) that can make life quite tricky.

Simon Ubsdell
tokyo productions
hawaiki

[Simon Ubsdell]

[Scott Witthaus] “Is this really that big of a deal?”

I would say that the current set-up is entirely workable (if curious) and nothing to get too distressed about.

But because I am interested in this kind of stuff, I wanted to share my findings in the hope that it would interest others and perhaps lead to a better understanding of some basic color science principles that it does pay to be aware of since they can help with your work.

That’s all really.

Simon Ubsdell
tokyo productions
hawaiki

[Simon Ubsdell]

[Tim Wilson] “And in Chrome, you can also just copy-paste and drag or drop images.”

Thank you!!!!

I had no idea about this, but I love it 🙂

Simon Ubsdell
tokyo productions
hawaiki

[Simon Ubsdell]

Hi Michael,

My advice, if you want to grade Alexa LogC (without a LUT) using the color wheels, is not to touch the Shadows control.

Set the black level with the Midtones control and adjust the Highlights as required.

If you try setting the blacks with the Shadows control first, you will make life more difficult than it needs to be.

And of course, use a first instance of the CW to balance the shot and only apply your look to another instance or instances.

Edit: I don’t mean to suggest there is anything anomalous about this. In fact, it shows that the Midtones correction is indeed a true gamma correction, and it’s gamma that is going to flatten out the Log curve and get you to a balanced place.

Simon Ubsdell
tokyo productions
hawaiki

[Simon Ubsdell]

[Bill Davis] “This might well be a BUG.”

Dear all,

Please, please, please stop calling it a “bug”.

There is nothing more guaranteed to get developers wanting to ignore you (and worse) than when you call it that without having the full picture.

Simon Ubsdell
tokyo productions
hawaiki

[Simon Ubsdell]

Dear Bill

Thank you for sharing your chat with Philippe.

To my mind there are two ways in which one can go about imparting this kind of information:

a) you can try and make it as easy as possible for your readers/audience/viewers/listeners to follow along, even if they don’t get to grips with the full complexity of the subject;

or b) you can dazzle everyone with your superior knowledge by copying and pasting impressive-sounding passages from the standard texts.

Both approaches are valid, but my preference is for the former.

Although the first option is recommended when you have been caught out in an egregious error.

Simon Ubsdell
tokyo productions
hawaiki