Simon Ubsdell
6
COW Coins
Forum Replies Created
-
[Bill Davis] “Now sowing distrust is simply this eras MOST effective strategy. “
How true that is.
And how very damaging to the fundamental fabric of our social interactions – at every level.
Would it were not so. The internet has a lot to answer for.
Simon Ubsdell
tokyo productions
hawaiki -
[Jeremy Garchow] “Except the Color Board. “
True. But there really is no reason why it shouldn’t be there except that they decided not to implement it. I don’t think we can read anything into that decision.
[Jeremy Garchow] “I’m sure it is, but it is odd that this behavior is present on the wheels and not the board. It suggests (and I have no idea what I am talking about, by the way) there is different math at play, or perhaps different ways of determining (or in this case, ignoring) the shadow, midtone, highlight range instead of pivoting on 0 (gain) and 1(lift) and sweeping the middle (gamma).”
Have a look at the screenshots I posted on my “Solution?” post above and see what you think.
I am certainly still leaving open the option that Apple have adopted a different model here but I am starting to wonder if the solution isn’t quite a lot simpler.
I haven’t really had enough time to look at this properly though so I am not going to commit to any definitive theory …
My problem is that I can’t see the virtue in the way that the Wheels behave in Rec.709, whereas in Rec.2020 they make complete sense and conform to what I would expect.
Simon Ubsdell
tokyo productions
hawaiki -
[Tim Wilson] “Wishing you all a bit more normalcy and a lot more joy in 2018,”
A very Happy 2018 to you, Tim.
Wishing you all the normalcy you could hope for in the coming year. You certainly deserve it.
Simon Ubsdell
tokyo productions
hawaiki -
[Bill Davis] “Already this thing that started out as “Apple Screwed Up!!!!” – is rapidly backpedaling into “Apple has approached the need for modern color grading somewhat differently.” “
Can I just go on record yet again as saying that I have not claimed or wanted to claim that “Apple Screwed Up”?
Thank you.
Simon Ubsdell
tokyo productions
hawaiki -
[Jeremy Garchow] “but couldn’t it be that Apple isn’t using a Lift/Gamma/Gain calculation for the “brightness” parameters on Master, Shadows, Midtones, and Highlights?”
“Brightness” is an offset control. You’re simply adding the control value to the pixel value:
pixelValue+controlValue
Offset is very useful for grading but it’s not going to replace the Lift/Gamma/Gain concept …
Edit: I should add that just because Apple give a familiar name to something, we should always be very cautious about assuming it shares the same function. Changing the meaning of names is what they like to do.
One prominent voice on that FB thread argued to me that what Apple were doing here was SMH – presumably because he was misled by the naming convention.
Simon Ubsdell
tokyo productions
hawaiki -
[Jeremy Garchow] “There is also a “Mix” parameter on each control of Master, Shadows, Midtones, and Highlights (MSMH).”
This is a really basic function.
You can see it in any Apple filter.
It’s basically asking “how much of this effect do you want to apply, between 100% and zero?”
Again very simple maths.
Simon Ubsdell
tokyo productions
hawaiki -
Just to add some illustration for Oliver’s very ingenious theory (which I strongly suspect is correct), here are some screenshots:
Expected gamma curve using the Color Board in Rec.709:
Unexpected gamma curve using the Color Wheels in Rec.709:
Expected gamma curve using the Color Wheels in Rec.2020:
Unexpected gamma curve using the Color Board in Rec.2020:
Simon Ubsdell
tokyo productions
hawaiki -
[Bret Williams] “I wish you’d repost. It really clarifies the situation.”
I’ve made it Public again, should anyone care to take a look. But I’ve made sure to disable comments to avoid the spread of FUD and misinformation which is really unhelpful here.
Incidentally Oliver P. has just suggested to me a very intriguing and plausible explanation for what’s happening here, but I won’t steal his thunder.
In fact, I’d never have noticed what was happening here at all if it weren’t for his initial prompting.
Clever fellow.
Simon Ubsdell
tokyo productions
hawaiki -
[Tony West] “Seems to me that if it’s a “feature”, then people might look toward Hawaiki Color, the product that you worked on.
If it’s a “bug”, then apple will likely fix it in an update and people might not be as likely to seek out HC.”
Disclaimer: Although I am very proud to have collaborated on Hawaiki Color and very grateful to the discriminating users who have sought it out over the years, I no longer have any commercial involvement with the product.
Simon Ubsdell
tokyo productions
hawaiki -
[Robin S. Kurz] “Neither (max/min) highlights nor shadows stay completely unaffected when adjusting the midtones, even with the color board as you describe.”
Yes, I did notice a little of that when doing the video.
Obviously the maths says that this shouldn’t actually happen.
0^gamma (or more practically speaking: 0^1/gamma) is always going to be zero whatever the input value. Similarly with 1^gamma (or 1^1/gamma) – the result should always be 1.
But the discrepancy is very small and is unlikely to be a problem in practice.
Simon Ubsdell
tokyo productions
hawaiki
