Assign Profile vs Convert to Profile

I knew this once upon a time but have now forgotten and can't seem to find an explanation anywhere. Could someone please explain, as simply as possible, the difference between Assign Profile and Convert to Profile in CS2. Thanks.

Convert to Profile will allow you to preserve the appearance of the image you're working on (using the Adobe Perceptual engine). It will change from, say, ProPhoto RGB to sRGB without the huge ugly color shift if you just assign a new profile.

Think of it as the difference between decoding and encoding. If you assign a profile, PS decodes the file as though the data comes from the chosen profile (whether that is correct or not). If you convert to profile (choosing both a source and destination space), PS changes (encodes) the data in the file to correspond with the chosen profile.

In both cases you have to know what color space the file was originally created in.

I just did this, so I hope I don't mess up the answer. Assign means that one is not only changing the profile, but the image also changes (one can see the changes to the image itself, at least I could). Convert means that one is only changing the profile for that image-the image itself doesn't actually change.

In general, you convert to a profile to move your entire image into a newly defined colorspace while preserving the color relationships.

By contrast you generally assign a profile to alter the colors and correct them for a specific device in a given colorspace -- like a profile for a digital camera.

a colour profile defines the colour space in which your are working. Your image (in 8bit colour) has 256 shades of each of Red, Green and Blue. When you open the colour pallete you can see the values defined in decimal for each colour against the R G and B values.

(Also below the RGB values, you can see the colour defined in base 16 which is hexadecimal and hexadecimal uses values 0-9 and A-F which gives the 16 characters of base 16. The first two characters of the hexadecimal value are the red value, the next two are the green and the last two are the blue value. I have added this only because in many applications you only see the hexadecimal values and it usually easier to work with)

back on track, when you ASSIGN a profile, each of the colour values in your image is mapped directly from the new profile. When you CONVERT to a profile it takes each of the colour values in your image and finds the nearest corresponding colour value in the new profile so that converted image will only have very small localised colour shifts which often you don't even notice on the screen, but they are still there.

For example, the colour value of pure blue (R)000 (G)000 (B)255 (hex 0000FF) may be very different in one profile compared to another. Assigning a profile may show a significant colour change whereas converting to a profile will show very little if any colour change. However, if you started with hex 0000FF and covert to a profile and then go back and check the hex value afterwards, you will find that the hex value has changed to a new value. When you assign a profile the hex values of your colours do not change.

Therefore, you should know that CONVERT physically alters your file and retains colours as best as possible using the the new profile for output but maybe with some clipping. Assign does not alter the file but the output using the new profile may look very different colourwise.

[edit]
Also note that converting from one profile to another and then converting back to the original profile is not guranteed to give you the same values as you started with, so be careful with convert when experimenting.
[/edit]

A simplistic way to think of it:

Assign Profile leaves the image data the same, but (usually) alters the appearance.
Convert to Profile alters the image data, but (more or less) keeps the appearance the same.

For a more elaborate explanation, re-read what robc wrote.

Thanks very much, I think I've got it.