Not that much (65 536 per channel when image is saved in 16 bits - only 256 when image is saved in 8 bits) but as I said above I think this sound to be enough (16 bits) in any color space to represent any subtlety. 8 bits per channel is far from enough (see banding problems even in srgb space) !
And all that is display medium resolution related.
With 8 bits you only can represent a gradient of 256 pixel wide in one channel from black to max value (I deliberately forget dithering - or adding “grain”).
And with 16 bits a gradient of 65 536 pixels wide.
Those are those in between colors that matters.
The importance isn’t what we can see. It’s the editing. Smaller steps prevents posterization or banding. At the end we all see it in the gamut and bit depth of the output device.
George
Well it’s a bit of both.
Deepprime AI searches for noise like looking (color)patherns which are removed before WB is applied and the image is demosiaced to pixels. To minimise the problem of noise being pixelised and got a part of the color rgb values which makes it more hidden.
The XD part is the invent part.
Rebuilding lost detail out of nearby looks the same-isch paterns.
2 Likes
8 bits per channel is usually enough to represent the image, even most fine gradients, but its not enough to survive heavy editing or delicate image elements, like fine gradation. So you need more, mostly for editing, rather than preservation, contrary to popular opinion.
Gamut on the other hand only matters if the image itself contains specific color, but with say grayscale image or black and white if you like, gamut is a moot point. So its really, specific to the image and device on which you are viewing/reproducing the image.
2 Likes
Hi,
I would suggest this one which is very clear and easy to understand :
8 Likes
Wow!!! Lightbulb moment!!! Finally, I now understand.
2 Likes
Patrick,
Many thanks for sharing this. I will need to watch it again (and again?) and experiment, and think about this - but I do feel I’m starting to get an understanding of this topic and how and why to use the PL features.
Thanks Again.
Paul
Great explanation + demonstration, why some saturated colours look better when converted from DxO Wide Gamut WCS instead from DxO Classic-Legacy WCS.
2 Likes
He is saying at 3:30 that the banding in his example is due to the compression of a wider color space into a smaller color space. The example is with a jpg. As far as I know banding is due to a to small bit depth when editing. sRGB in a jpg is 8 bit.
George
Banding appears when, in a given area of the image, colors that are initially different (but out of gamut) are replaced with a unique color. This may happen during various operations, including conversion to a smaller color space.
As explained in this thread, this is the reason why the higher accuracy of the calculations made in a wide working color space and a higher bit-depth are often decisive.
In this specific case there’s no out of gamut. The jpg is in sRGB and 8 bit. And a rather extreme editing is done. To me it’s a typical 8 bit problem.
https://www.cambridgeincolour.com/tutorials/bit-depth.htm
George
Thank you for sharing.
I already had a good idea of how things works, but that is a simple but brilliant visual explanation 
1 Like
Hello,
we have discussed this 10 months ago and there was a link to https://youtu.be/Hx-l7Avm8Bs?list=PLjZ7Y0kROWitoJtnw0pdvjPmS8mYGvrBR by James Ritson, also explaining why to work in wider color space.
Good start into the week by
Günter
These are the type of essential insights - for whatever reason - are not provided by DxO. It would avoid so much misunderstandings, confusion and useless discussions.
5 Likes
Actually, an in-depth book about Photolab would be welcome. But it seems that nobody is able to write it or that no editor is wanting to invest in such a project. Maybe Photolab is evolving too fast…
1 Like
6 Likes
Affinity manages to produce masses of videos on its programs. Why DxO can’t do something like it on a much more limited number and range of programs is a pity especially as the “manuals” are not very good. There are as in this topic a number of very good presentations by others but its had to be users who are finding and point others to them not DxO. Given the long running discussions and clearly high level of confusion on the new colour space in v6 if they do actually read what’s here its even more appalling its been left to non DxO people to make it so much clearer. I will be the first to admit these videos have made me understand what it is and hopfully be able to use it now.
3 Likes
Actually a very good point. There must be some DXO engineers who could make a You Tube training video like James Ritson?
1 Like
Indeed.
But about any software I use does not explain the basics of the business for which they are designed.
(Anyone who process digital images should know what a digital color is).
Anyway, happy to see that people understand that this new color space is really a major improvment for photolab and a must be for any serious demosaicer.
And it seems that the choice for this working color space is very clever and photography oriented : so good job on this for DxO team !!!
1 Like