Color/Luminosity separation

Started 3 months ago | Discussions
Currantos Senior Member • Posts: 1,448
Color/Luminosity separation

Hi,

I've read early a general statement about more advanced workflow where one separates the color and luminosity on different layers and then works on each separately.

I've looked up a few tutorials and experimented, kind of getting inconsistent results since there are different 'luminosities'.

Anyone has helpful simple way to do it?

In that tutorial they separated luminosity and color AND did frequency separation, so there was no 'color contamination' and color stayed consistent throughout.

At that time when I read it I did not have enough PS knowledge to fully understand the details and commit it to memory, only got the vague idea. Now that I am a little further ahead in PS I'd like to recreate that workflow and I see advantages to working separately on luminosity/color.

Thanks everyone

Sam B
Sam B Veteran Member • Posts: 6,297
Re: Color/Luminosity separation
1

Here are some tutorials from Piximperfect on the subject:  https://www.youtube.com/c/PiXimperfect/search?query=frequency%20separation

 Sam B's gear list:Sam B's gear list
Apple iPhone 12 Pro Max
OP Currantos Senior Member • Posts: 1,448
Re: Color/Luminosity separation

Thank you

I see a lot of Frequency separation onces, don't see one about separating color/luminosity.

Are you pointing to a specific one?

gordonpritchard Veteran Member • Posts: 4,886
Re: Color/Luminosity separation
7

If you want to separate luminosity from color then just convert your image mode from RGB to Lab mode. Your images will still be three channels: "L" the luminosity channel, and two color channels: a and b. Select (make active) the L channel and work on it. When finished change the mode back to RGB.

Alternatively you can duplicate your image to a new layer and set that layer's blend mode to "Luminosity" Work on that layer and when finished flatten the image.

-- hide signature --

———————————————————————------
"View their gallery before accepting their comments."
———————————————————————------
My photos: http://www.gordonpritchard.blogspot.com/

 gordonpritchard's gear list:gordonpritchard's gear list
Sigma 10-20mm F4-5.6 EX DC HSM Sony a77 II Sony SLT-A57 Sony 500mm F8 Reflex Tamron 16-300mm F/3.5-6.3 Di II VC PZD Macro
Pictus
Pictus Veteran Member • Posts: 6,543
Re: Color/Luminosity separation
2

Currantos wrote:

Hi,

I've read early a general statement about more advanced workflow where one separates the color and luminosity on different layers and then works on each separately.

Anyone has helpful simple way to do it?

Hi, to proper separate color/luminosity in Photoshop https://www.modelmayhem.com/forums/post/894058/2#post18305720

AKoblentz Regular Member • Posts: 159
Re: Color/Luminosity separation
1

If you are working in Lightroom or ACR, you may use enhanced profiles LAB Color Vol. 7 - Chrominance Blend, just google

Mark Hollister Regular Member • Posts: 482
Re: Color/Luminosity separation
1

Currantos wrote:

Hi,

I've read early a general statement about more advanced workflow where one separates the color and luminosity on different layers and then works on each separately.

I've looked up a few tutorials and experimented, kind of getting inconsistent results since there are different 'luminosities'.

Anyone has helpful simple way to do it?

In that tutorial they separated luminosity and color AND did frequency separation, so there was no 'color contamination' and color stayed consistent throughout.

At that time when I read it I did not have enough PS knowledge to fully understand the details and commit it to memory, only got the vague idea. Now that I am a little further ahead in PS I'd like to recreate that workflow and I see advantages to working separately on luminosity/color.

Thanks everyone

The simplest way is: Ctrl-J twice to make two copies of your layer. Set one to color blending mode, set the other to luminosity blending mode. I'd rename them as COLOR and LUMINOSITY to make it easier to keep straight which one is which.

You can do the same thing with adjustment layers, use the appropriate adjustment layers for color and luminosity.  Set their blending modes as I described above.

Tom Axford Veteran Member • Posts: 8,537
Re: Color/Luminosity separation
1

Currantos wrote:

Hi,

I've read early a general statement about more advanced workflow where one separates the color and luminosity on different layers and then works on each separately.

That can work fine provided you keep the colour and luminosity separate.

However, you have to eventually convert back to RGB to view your image.  The problem is that by working on colour and luminosity separately you can easily go outside the RGB gamut (which is much less than colour + luminosity) and so the colours get distorted when you convert back to RGB.

Tom Axford Veteran Member • Posts: 8,537
Re: Color/Luminosity separation
1

Just to explain in more detail what I meant in my previous post:

Take a colour like the purple shown below. In RGB it is (128, 0, 255), but its luminosity (L* in LAB colour) is only 39.3% (pure white is 100%).

If we are working in the luminosity channel and double the luminosity to 78.6%, that seems perfectly reasonable. But for this purple colour, doubling the luminosity would change the RGB values to (256, 0, 510) which is way outside the gamut of RGB (each R, G and B value must be between 0 and 255).

I'm not sure exactly what the software does when converting out-of-gamut LAB values to RGB, but supposing it simply clips the channels to the maximum 255. Then the colour would be shown as (255, 0, 255) which is a very different colour.

RGB = (128, 0, 255)

RGB = (255, 0, 255)

I use HSV Value instead of Luminosity if I wish to change the lightness of the image in a way that is guaranteed to preserve the colours exactly. Unfortunately, HSV colour is not supported in Photoshop and most editors, so I use GIMP, where it is supported.

Joachim B Contributing Member • Posts: 938
Re: Color/Luminosity separation

Currantos wrote:

Hi,

I've read early a general statement about more advanced workflow where one separates the color and luminosity on different layers and then works on each separately.

I've looked up a few tutorials and experimented, kind of getting inconsistent results since there are different 'luminosities'.

Anyone has helpful simple way to do it?

In that tutorial they separated luminosity and color AND did frequency separation, so there was no 'color contamination' and color stayed consistent throughout.

At that time when I read it I did not have enough PS knowledge to fully understand the details and commit it to memory, only got the vague idea. Now that I am a little further ahead in PS I'd like to recreate that workflow and I see advantages to working separately on luminosity/color.

Thanks everyone

If you wanted you could read up a bit on the L*ab color mode.

Joachim B Contributing Member • Posts: 938
Re: Color/Luminosity separation
1

gordonpritchard wrote:

If you want to separate luminosity from color then just convert your image mode from RGB to Lab mode. Your images will still be three channels: "L" the luminosity channel, and two color channels: a and b. Select (make active) the L channel and work on it. When finished change the mode back to RGB.

Alternatively you can duplicate your image to a new layer and set that layer's blend mode to "Luminosity" Work on that layer and when finished flatten the image.

If you don't want to switch all the time and keep things non-destructive and dynamically adjustable, and you use PS, you could make the layer a Smart Object. You can open that up as an indivudual document, change it to Lab mode, and make your adjustments. When you close that, you can use the layer in your RGB document as you are used to, and you can always go back to the Smart Layer and edit the settings.

Pictus
Pictus Veteran Member • Posts: 6,543
Re: Color/Luminosity separation

Tom Axford wrote:

That can work fine provided you keep the colour and luminosity separate.

However, you have to eventually convert back to RGB to view your image. The problem is that by working on colour and luminosity separately you can easily go outside the RGB gamut (which is much less than colour + luminosity) and so the colours get distorted when you convert back to RGB.

Use a very large Color Space like ProPhoto RGB and there will be no problem.
With smaller Color Space we may have problems like loss of brightness in the girl dress https://www.modelmayhem.com/forums/post/894058/2#post18305828

GIMP has the grain extract trick and I have to check with Affinity Photo

Tom Axford Veteran Member • Posts: 8,537
Re: Color/Luminosity separation

Pictus wrote:

Tom Axford wrote:

That can work fine provided you keep the colour and luminosity separate.

However, you have to eventually convert back to RGB to view your image. The problem is that by working on colour and luminosity separately you can easily go outside the RGB gamut (which is much less than colour + luminosity) and so the colours get distorted when you convert back to RGB.

Use a very large Color Space like ProPhoto RGB and there will be no problem.

Unfortunately, the problem is more fundamental than that and changing the colour space doesn't solve it.

It occurs because mathematically Lab colour has a much larger gamut than RGB.  However, the difference is not because some colours in Lab cannot be represented in RGB.  Instead, it is because Lab can represent saturated colours on a much wider brightness scale than is possible in RGB (whatever RGB colour space is used).

Pictus
Pictus Veteran Member • Posts: 6,543
Re: Color/Luminosity separation

Tom Axford wrote:

Unfortunately, the problem is more fundamental than that and changing the colour space doesn't solve it.

It occurs because mathematically Lab colour has a much larger gamut than RGB.

Are you sure?
http://www.brucelindbloom.com/index.html?WorkingSpaceInfo.html

However, the difference is not because some colours in Lab cannot be represented in RGB. Instead, it is because Lab can represent saturated colours on a much wider brightness scale than is possible in RGB (whatever RGB colour space is used).

From https://www.dpreview.com/forums/post/55571367
"Keep in mind that CIELab was just an attempt to create a perceptually uniform color space where equal steps correlated to equal color closeness based on the perception of a viewer. The CIE didn't claim it was prefect (because its not). Most color scientists will point out that Lab exaggerates the distance in yellows and consequently underestimate the distances in blues. Lab assumes that hue and chroma can be treated separately. There's an issue where hue lines bend with increase in saturation perceived by viewers as an increase in both saturation and a change in hue when that's really not supposed to be happening. Further, according to Karl Lang, there is a bug in the definition of the Lab color space. If you are dealing with a very saturated blue that's outside the gamut of say a printer, when one uses a perceptual rendering intent, the CMM preserves the hue angle and reduces the saturation in an attempt to make a less saturated blue within this gamut. The result is mathematically the same hue as the original, but the results end up appearing purple to the viewer. This is unfortunately accentuated with blues, causing a shift towards magenta. Keep in mind that the Lab color model was invented way back in 1976, long before anyone had thoughts about digital color management."

Anyway, in Photoshop by staying in 16 bits+RGB color mode+ProPhoto RGB  color space when doing this https://www.modelmayhem.com/forums/post/894058/2#post18305720 to separate color/luminosity, I see no loss of colors.

Tom Axford Veteran Member • Posts: 8,537
Re: Color/Luminosity separation

Pictus wrote:

Tom Axford wrote:

Unfortunately, the problem is more fundamental than that and changing the colour space doesn't solve it.

It occurs because mathematically Lab colour has a much larger gamut than RGB.

Are you sure?
http://www.brucelindbloom.com/index.html?WorkingSpaceInfo.html

However, the difference is not because some colours in Lab cannot be represented in RGB. Instead, it is because Lab can represent saturated colours on a much wider brightness scale than is possible in RGB (whatever RGB colour space is used).

From https://www.dpreview.com/forums/post/55571367
"Keep in mind that CIELab was just an attempt to create a perceptually uniform color space where equal steps correlated to equal color closeness based on the perception of a viewer. The CIE didn't claim it was prefect (because its not). Most color scientists will point out that Lab exaggerates the distance in yellows and consequently underestimate the distances in blues. Lab assumes that hue and chroma can be treated separately. There's an issue where hue lines bend with increase in saturation perceived by viewers as an increase in both saturation and a change in hue when that's really not supposed to be happening. Further, according to Karl Lang, there is a bug in the definition of the Lab color space. If you are dealing with a very saturated blue that's outside the gamut of say a printer, when one uses a perceptual rendering intent, the CMM preserves the hue angle and reduces the saturation in an attempt to make a less saturated blue within this gamut. The result is mathematically the same hue as the original, but the results end up appearing purple to the viewer. This is unfortunately accentuated with blues, causing a shift towards magenta. Keep in mind that the Lab color model was invented way back in 1976, long before anyone had thoughts about digital color management."

Anyway, in Photoshop by staying in 16 bits+RGB color mode+ProPhoto RGB color space when doing this https://www.modelmayhem.com/forums/post/894058/2#post18305720 to separate color/luminosity, I see no loss of colors.

We are talking about somewhat different things.

If you go increase the luminosity up to the maximum for all pixels, when you convert back to RGB (whatever colour space), the colours will often be washed out, possibly even to pure white.  Exactly what happens on conversion from Lab to RGB will depend on the software, but usually the colour is not preserved if the luminosity is greater than RGB can represent for that colour.

Try it for yourself, I'd be interested to know what result you get.  Create a pure white luminosity layer (i.e. maximum luminosity) and then combine your original colour layer with the white luminosity layer.  When I have done that in GIMP, the result has not preserved the colours.  There is no guarantee that Photoshop will do exactly the same.

An example:

Original image

Luminance set to maximum in GIMP

HSV Value set to maximum in GIMP - this truly preserves colour (hue and saturation) while increasing luminosity to the maximum possible for each pixel individually

Pictus
Pictus Veteran Member • Posts: 6,543
Re: Color/Luminosity separation

Tom Axford wrote:

We are talking about somewhat different things.

If you go increase the luminosity up to the maximum for all pixels, when you convert back to RGB (whatever colour space), the colours will often be washed out, possibly even to pure white. Exactly what happens on conversion from Lab to RGB will depend on the software, but usually the colour is not preserved if the luminosity is greater than RGB can represent for that colour.

Try it for yourself, I'd be interested to know what result you get. Create a pure white luminosity layer (i.e. maximum luminosity) and then combine your original colour layer with the white luminosity layer. When I have done that in GIMP, the result has not preserved the colours. There is no guarantee that Photoshop will do exactly the same.

In RGB color mode, a color layer above a white layer results in a 100% white image
In LAB color mode, a color layer above a white layer results in this

Tom Axford Veteran Member • Posts: 8,537
Re: Color/Luminosity separation

Pictus wrote:

Tom Axford wrote:

We are talking about somewhat different things.

If you go increase the luminosity up to the maximum for all pixels, when you convert back to RGB (whatever colour space), the colours will often be washed out, possibly even to pure white. Exactly what happens on conversion from Lab to RGB will depend on the software, but usually the colour is not preserved if the luminosity is greater than RGB can represent for that colour.

Try it for yourself, I'd be interested to know what result you get. Create a pure white luminosity layer (i.e. maximum luminosity) and then combine your original colour layer with the white luminosity layer. When I have done that in GIMP, the result has not preserved the colours. There is no guarantee that Photoshop will do exactly the same.

In RGB color mode, a color layer above a white layer results in a 100% white image
In LAB color mode, a color layer above a white layer results in this

Thanks for posting that.  It is not the same as what GIMP does, but similar in that it certainly doesn't preserve the colours accurately if the luminosity is higher than can be represented in RGB.

Of course, this example is very extreme, but even much more modest operations on the luminosity layer can result in colour changes.  The changes do not normally affect the whole image, but only those areas where the colours are highly saturated.  Highly saturated blues are the most susceptible.

Joachim B Contributing Member • Posts: 938
Re: Color/Luminosity separation

Do you guys keep in mind that Value, Brightness, Lightness and Luminocity are different things, and are not interchangeable?

Tom Axford Veteran Member • Posts: 8,537
Re: Color/Luminosity separation

Joachim B wrote:

Do you guys keep in mind that Value, Brightness, Lightness and Luminocity are different things, and are not interchangeable?

Yes, they are different ways measuring how bright a pixel is, and are certainly not interchangeable.  However, each of them provides a measure of the brightness of a pixel independently of its colour (hue and saturation).

Pictus
Pictus Veteran Member • Posts: 6,543
Re: Color/Luminosity separation

Tom Axford wrote:

Thanks for posting that. It is not the same as what GIMP does, but similar in that it certainly doesn't preserve the colours accurately if the luminosity is higher than can be represented in RGB.

Of course, this example is very extreme, but even much more modest operations on the luminosity layer can result in colour changes. The changes do not normally affect the whole image, but only those areas where the colours are highly saturated. Highly saturated blues are the most susceptible.

You are welcome, it is an interesting test to check the LAB Color Mode behavior.

Keyboard shortcuts:
FForum MMy threads