FORUMS: list search recent posts

Possible bugs...

COW Forums : DaVinci Resolve

<< PREVIOUS   •   VIEW ALL   •   PRINT   •   NEXT >>
Yoram Tal
Possible bugs...
on Dec 27, 2012 at 8:59:29 pm

Hi all.

I am working on my brand new full version 9.0.4 and finding a few things that might be bugs... Please chime in if you are encountering the same.

I am using iMac 27' with black magic thunderbolt 3d video card and Avid Artist control board.

1. Auto correct stops working on occasion and will not come back until reboot.
2. Curves will start acting in reverse to their normal behavior ) creating a point pulling it down will brighten picture up will darken)

Odd. Anybody else seeing this?

Thanks.

Tal


Return to posts index

Joseph Owens
Re: Possible bugs...
on Dec 29, 2012 at 12:11:45 am

[Yoram Tal] "1. Auto correct stops working on occasion and will not come back until reboot.
2. Curves will start acting in reverse to their normal behavior ) creating a point pulling it down will brighten picture up will darken)"


Sorry, point (1) I cannot comment on as I have never used auto correct.

Point (2). DaVinci Custom curves are a behaviour modification, not a correction, and this is the misapprehension that someone else who is trying to get the developer to change the application into Photoshop is labouring under. The curve is *multiplied* by the gamma adjustment, so if you have a negative gamma factor, then "up" is "down" and vice versa.

jPo

"I always pass on free advice -- its never of any use to me" Oscar Wilde.


Return to posts index

Laco Gaal
Re: Possible bugs...
on Dec 30, 2012 at 11:33:46 pm

Joseph,

what is the best way to reproduce the curves that I did in Davinci in Photoshop? (Davinci is the source, and I'd like to do the same grade in Photoshop, only curves were used)

thanks in advance.


Return to posts index


Joseph Owens
Re: Possible bugs...
on Dec 31, 2012 at 5:48:47 pm

[Laco Gaal] "what is the best way to reproduce the curves"

It would probably have to be a two-layer process in Photoshop. Draw the same shape curve -- do not move the endpoints, and then use a histogram multiplier to emulate the gain, black and gamma factors. It will be inexact because the internal matrixing will be completely different between the sRGB application, which is NOT operating in Rec709, and Resolve, and because CustomCurves, at least as far back as its introduction on Renaissance 8:8:8 was never designed to be a correction method on its own, and thematically would better fit in the YSFX section of the manual.

Another path to explore might be to try to simulate the image values with a LUT that might be exportable. But because the target values are dependent on the base grade for a specific clip, it might not be universal. Which is why anti-log LUTs for field-acquired images are not a magic filter -- cinematographers will be up and down the response curve of the overall exposure behaviour of their imaging device, just like Original Camera Negative, so there are non-linearities within the non-linearities. The classic puzzle wrapped in an enigma posed as a riddle.

And don't we all love a mystery?

Happy New Year, tous le monde.

jPo

"I always pass on free advice -- its never of any use to me" Oscar Wilde.


Return to posts index

Laco Gaal
Re: Possible bugs...
on Dec 31, 2012 at 6:08:44 pm

Joseph, thanks for the detailed answer.
At the end I went this way:

- I tried to draw the same exact Red, Green, and Blue curves as I did in Davinci (including the endpoints)
- after that I saw that the picture in Davinci has more contrast - as you mentioned, color channel curves are "multiplied" with the gamma curve - so I let the R, G, B curves alone, and draw a gamma-like curve on the unified RGB curve. This did the trick:)

Happy New Year!


Return to posts index

<< PREVIOUS   •   VIEW ALL   •   PRINT   •   NEXT >>
© 2017 CreativeCOW.net All Rights Reserved
[TOP]