Originally posted by undo
When using Split HSL, it separates away the layers of the image (if there are any) based on Hue, Saturation and Lightness.
That is the reason I still do not believe in this method.
Splitting to HSL is the same as splitting to RGB or CMYK, it decomposes the image in their components using different systems, not the layers of the
original image.
I made some images that I hope will explain what I want to say.
This is the original image.
It was made using the text tool on PaintShopPro 6, that creates a text layer. Then all layers were joined to the background and saved as PNG to avoid
those nasty JPG artefacts.
The first sentence is written with the same Hue as the background colour but fully saturated.
The second sentence is written with the same Hue as the background but with a different Lightness, that is why is whiter than the first. The
saturation is also at the maximum.
The third sentence is written in a different Hue but with the same saturation (fully saturated) and lightness than the first sentence.
These are the "H", "S" and "L" images resulting from the "split to HSL" command in PainShopPro 6.
"H"
As you can see, and because the pixel values of this image are mapped to the Hue of the pixels of the original image, the first and second sentences
have exactly the same Hue value as the background, so they cannot be seen in this image.
The third sentence, being in a different Hue, is clearly seen over the background.
"S"
In this image all sentences look the same because they were all written with full saturation, that is why they look white over black when seen only as
saturation values over the background, which has zero saturation.
"L"
The lightness image shows the difference in lightness between the second sentence and all others. The second sentence was the only one written with a
different lightness value, so its the only one to show over the background. The pixels around the first and third sentences can be seen because the
text was applied with anti-aliasing, creating a smooth contrast between the letters and the background.
What I think this means is that if some image is applied to the (digital) original but we keep its Hue, Saturation and Lightness the same as the
surrounding area, then the separation to HSL can not show that the original was changed.
Also, as this only applies to colour images, its useless in greyscale images, if that image to which SteveR pointed us to had not the yellow text and
the arrows, as the original images from John Lear, this technique of splitting to HSL would probably show nothing, as you can see if you split the
original John Lear images to HSL.
This is also useless in greyscale images they had a tint applied to them, like those two images that
zorgon posted above.
PS: PaintShopPro was a very user friendly program before Corel bought it.