Luminosity

[BigDingus]

Hi folks.

I've done the usual Google and YouTube search, but can't find a good tutorial on Photoshop Luminosity that isn't a platform for selling another piece of software.

Can anyone recommend a good tutorial, even on Adobe site?

[Moderator edited title to fix typo.]

[AxelMatt]

I've searched with Google and Youtube and found tons of tutorials.

A Simple But Effective Way to Create Luminosity Masks in Photoshop - YouTube

How To Create Luminosity Masks In Photoshop - YouTube

Creating luminosity masks in Photoshop | Photo Tools Weekly | lynda.com - YouTube

Understanding Luminosity Selections and Luminosity Masks Basics - YouTube

[Trevor.Dennis]

BigDingus  wrote

Hi folks.

I've done the usual Google and YouTube search, but can't find a good tutorial on Photoshop Luminosity that isn't a platform for selling another piece of software.

Can anyone recommend a good tutorial, even on Adobe site?

Can you give us some context? 

Are you asking what it is, or how to use it?

Are you thinking of masks, or blend modes?

If you don't feel that the work people put into developing useful software is worth rewarding, then Sven Stork's Luminosity Mask extension is free unless you want to make a small donation.

https://svenstork.com/interactive-luminosity-masks/

https://www.adobeexchange.com/creativecloud.details.12307.html

[BigDingus]

Thanks Axel. I too found loads on YouTube but they then go into how to use their own addons etc.

Trevor if I was wanting to use their sofetware I would deffinately reward them. However, all I want is the knowledge to do it myself and not have someone use an underhand method of selling their software by pretending to help.

[BigDingus]

Axel, been looking at your links. You found better ones than I did. Thanks

[Norman Sanders]

If your objective is altering and controlling luminosity (discrete from color), consider switching to Lab Color Mode where the L channel (luminosity) is separate from the magenta/green and yellow/blue channels that build the image. An excellent book devoted to Lab Color is "Photoshop LAB Color" by Dan Margulis. For starters I recommend the 2006 edition.

[BigDingus]

Thanks for that Norman

[D Fosse]

Photoshop Luminosity is identical to the Lab L channel. You don't need to switch to Lab mode just for that.

Lab is always at work in the background in Photoshop. Color management in Photoshop uses Lab as Profile Connection Space, so Lab values are always readily available. This is what Luminosity blends and selections are based on. When you load a luminosity selection, you get Lab L, and not a desaturated RGB version, which is different.

[Stephen Marsh]

I would suggest that Luminosity (RGB) and Lightness (L of Lab) are indeed different, however in practice they are similar in how they are used.

[D Fosse]

Stephen_A_Marsh  wrote

I would suggest that Luminosity (RGB) and Lightness (L of Lab) are indeed different, however in practice they are similar in how they are used.

It should be easy enough to test. In my own limited testing I have found them to be identical. And it would be the logical way to do it, since the Lab version is always present in the background. Those data are easily pulled up.

Note, however, that if you display a luminosity mask as a single-channel grayscale image, it will be displayed according to your working gray. Your working gray is assigned.

So you need to keep gamma identical. In practice that means you must test with Adobe RGB and Gray Gamma 2.2 as working gray. All other combinations will change the tone response curve in the mask representation on screen.

[Stephen Marsh]

Replying To:

D Fosse “It should be easy enough to test. In my own limited testing I have found them to be identical. And it would be the logical way to do it, since the Lab version is always present in the background. Those data are easily pulled up.

Note, however, that if you display a luminosity mask as a single-channel grayscale image, it will be displayed according to your working gray. Your working gray is assigned.So you need to keep gamma identical. In practice that means you must test with Adobe RGB and Gray Gamma 2.2 as working gray. All other combinations will change the tone response curve in the mask representation on screen.”

____________

It is easy to test – and test results with confirm my previous statement. As I said they are not exactly the same thing. They are even called different things, the L in Lab is lightness and the L in RGB Luminosity is not exactly the same as Lab lightness. The HSL L(ightness) value is also not exactly the same thing as the L(ightness) value in Lab.

If the RGB luminosity and the Lab Lightness were exactly the same thing, you could paste the L channel from Lab over the same RGB image and blend it in luminosity mode without any visible change to the underlying RGB image.

Conversely if they were the same thing, the luminosity component of an RGB file could be pasted into the L channel of the same file in Lab mode and there would be no visible change to the image.

In both cases this is not so, there is a visible change.

As I said, people use the term loosely and interchangeably to indicate a method of editing that separates colour edits from tonal edits – however technically they are not the same data and they can’t simply be swapped.

[D Fosse]

OK. I'm travelling with only a laptop right now, no Photoshop, so I'll have to look into it later.

[Stephen Marsh]

Question: What does the RGB composite curve actually represent?  Answer: RGB Luminosity.

RGB luminosity is often “disguised and hidden” in Photoshop, here is a solid fill of 128r128g128b in sRGB. The Lab L value is 54. The RGB composite curve is value is 128…

How about with a bright colour that does not have equal channel values? A lime green sRGB of 151r255g0b has a Lab L value of 91, however the RGB composite curve has a marker of 196. Notice that in Lab, the L value on the curve precisely lines up with the histogram value. In the RGB version, the RGB histogram value is at 151 (median value) and not 196 (luminosity value).

[D Fosse]

Well, that's the whole point of using L (or so I thought). The lime green is an excellent example. Composite RGB would show this as a drab middle gray because it just averages the channels. L maintains the inherent brightness of the color.

[Stephen Marsh]

My main point is that RGB Luminosity and Lab’s Lightness are very similar in their underlying data, but they are not exactly the same thing. The concept may be similar in splitting tonal data from colour, however again they are not exactly the same thing.

So how does one create true RGB luminosity data, if copying the L channel from Lab into the RGB file and blending in luminosity blend mode changes appearance (however small)?

It was mentioned earlier in the thread that desaturating RGB was not the answer, however this leads the way to the correct answer. Luminosity is the component of the RGB image that does not contain Hue or Saturation (Color) data. So desaturating without affecting luminosity is the way to create RGB Luminosity data. There are many ways, one can fill with any neutral value (white, black or shades of gray) using color, hue or saturation blend modes.

[D Fosse]

Stephen_A_Marsh  wrote

My main point is that RGB Luminosity and Lab’s Lightness are very similar in their underlying data, but they are not exactly the same thing.

I finally got to my workstation so I could test this.

Luminosity is Lab - but with a twist: it's not identical to the L channel, but to a desaturated Lab image.

When you display the Lightness channel in a Lab image, you are representing a single channel on screen. And that's a grayscale image. What always happens then, is that your working gray is assigned. It's Lab data, but represented in your working gray.

Of course, the same happens to a mask displayed directly (alt-click).  You can see this for yourself by changing working gray, the display changes in real time. So that should even out, but there is still a very slight difference. However, the difference appears to be only a slight gamma change.

Here's a screenshot demonstrating what I found.

1 original Adobe RGB

2 converted to Lab and desaturated

3 original Adobe RGB with luminosity mask displayed directly (alt-click)

4 original Adobe RGB desaturated

You can overlay 2 and 3 and they are absolutely identical. But note that no.4, the desaturated RGB image, is a different animal altogether. Here the bright yellow-greens turn into a dull gray, because the values are averaged.

So the takeaway is that Luminosity is fundamentally Lab L data, but with some subtle change in the tone response curve (gamma) that I haven't figured out yet. I have a hunch that the working gray has a part in that too.

[D Fosse]

Or a simpler way to put it: Luminosity is the image converted to grayscale. That's a perfect match.

A grayscale conversion also goes through Lab, and again the crucial implication of that is that the inherent lightness of a color is preserved. Same arrangement:

An old screenshot I've used many times, which explains itself:

[Stephen Marsh]

Converting to grayscale in modern versions of Photoshop uses the Lightness values from Lab, however this is not exactly the same thing as RGB Luminosity.

I agree, desaturating without taking luminosity into account is not RGB Luminosity… Which is not what I am suggesting. If one desaturates while blending to color; hue or saturation blend modes, then one preserves the RGB luminosity. This maintains the distinction between hues, however Lightness is not exactly the same as Luminosity, as we agree the tonal response is different, so different numbers.

Annotated images below to illustrate, the right hand side was converted from RGB to Lab and back to RGB for the mockup, however the measurement samples expressed in Lab were each taken in the native colour space. Visually and numerically, they are not the same, but they are achieving the same thing in their respective native colour space. This is why one can’t simply swap the Lightness channel from Lab for the RGB Luminosity:

Here is a good PDF that explains what RGB Luminosity is and what it is not. Ignore the fact that the PDF is from a (dated) commercial tool, judge the content for yourself and test, test, test to verify the information presented from all sources. One does not need this commercial tool to create a true monotone RGB Luminosity layer, I have detailed how to do so:

Manual Section Two - LUMINOSITY

[Ussnorway]

in my book the best use for them is turning summer photos into winter

here is one that covers the basics but the trick is as old as Photoshop itself and there are many fine-tune options so look around

Summer To Winter - Snow Photoshop Tutorial - YouTube

[BigDingus]

When I first posted here, it was to learn about Luminosity. I knew absolutely nothing. However your discussions between you are revealing loads for me to soak up like a sponge. Excellent.