I think you’ll find that most people don’t change the colour mode all that often in any one image. At most, there’ll be one change from RGB to CMYK right at the end of the image manipulation process, as well as maybe 16 bit to 8 bit, before the image is to be sent off to the printers.
Changing the Color Mode several times is not a very professional way of working, IMHO. No matter what version you are using the image will degrade when you do that. Why do you need to do this?
Absolutely wrong – no lossing we have in CS! Even after decades of convertings!
As for your "IMHO", bear in mind that Photoshop is rather universal program, and there are many non-standard tasks where it could be useful – it is long to explain! – and where one really does need to convert image from RGB to Lab, and back, especially when you want to repair very bad color and to prevent color noise increasing at the same time. In such methods you need to separate color information from lightness, and it is possible only by going to Lab.
And imagine: who is that here Adobe’s programmer who cannot do something new without crashing some previous achievements?
It’s only natural that you would get effects similar to that, because RGB-8 and Grayscale both use 8-bit integers to represent values, so there are rounding errors when converting from RGB to Grayscale and the colorspace profiles are applied.
Converting from RGB to Grayscale does not simply use the average of RGB to produce Grayscale, which would produce no changes in either direction in the case of a desaturated white-gray-black gradient such as post #4. Instead, the default conversion uses a weighted combination of R, G, and B for the black level in grayscale.
Photoshop uses the formula K = 59% * G + 30% * R + 11% * B when doing a simple conversion from RGB mode to Grayscale. Since the three coefficients add up to 100% this won’t cause any changes going back and forth in an image that is completely desaturated such as the example. (There is a minuscule possibility of very slight rounding errors when the floating point intermediate products are added together and converted back to integers, but I doubt there would be any differences in the integer results from converting due to this.) In a Grayscale -> RGB conversion, it simply uses the grayscale K level as the coefficient for R, G, and B.
However, that’s not all that happens. Photoshop also takes your image’s working space (color profile) into account in each conversion, and the working spaces for your RGB and Grayscale intermediate conversions are unlikely to have exactly the same color curves. The ripples or ridges in the result of your repeated conversion undoubtedly result from the effect of the repeated profile changes.
Do you read the message with closed eyes? There is no effect in CS, with any bit-depth! Many words – low sense! Koshmar!..
Alexandra: If you are going to be rude to the kind people here who are trying to help you, you might just find any future requests are completely ignored.
Alexander, what are your workspace profiles for RGB and Grayscale in CS and CS2? If there are differences, that would explain different results in the two versions.
All profiles are the same, but it cannot be the reason of the effect.
Note that in the end of 90s Adobe announced that it remains no problem with color mode convertions because the intermediate color space bit depth had become much better, and no rounding effects we will get forever. And it was true. Till CS2 had come…
Dear friends, I do not want to be rude to some kind fellows but I like not when onesome is so hurrying up to return. It is very common place to veil clear facts by well-shaped words.
And I need no help – simply I use Photoshop 8 and am getting no troubles in those mentioned points of work. I am leaving CS2 being installed on my comp too, but using it only for some additional operations (geometric transformings, for example) where it is mighty.
And humbly waiting for proper updates…
