[argyllcms] Re: Could someone help me understand something applying printer profiles.

On Thu, 2008-01-24 at 19:18 +0000, Alastair M. Robinson wrote:
> Hi :)
> 
> Leonard Evens wrote:
> 
> > By the way, the reason I prefer not to use photoprint is that the
> > interface is very confusing and the documentation is not much clearer.
> > I am not sure I really know how to apply a printer profile using it.
> 
> In the Options -> Colour Management dialog, simply set the Printer 
> Profile to the profile you've made for your printer, and set the Default 
> RGB profile to sRGB or whatever you prefer.  Images will now be 
> transformed into the printer's colour space when printing.
> 
> PhotoPrint could definitely use some decent documentation!

Thank you. I did figure that out myself.  Using cctiff helped understand
the process better, so when I went back today to look at photoprint, it
became clear what to do.

I'm still not sure what Image>Set color profile is for.   When I set
something there, and the original file had an embedded profile, it warns
me it will override it.  I don't know what the purpose of the before and
after is.  And how does all this interact with the default image profile
set under options? 

> 
> > But what would photoprint do with an 8 bit per channel image that
> > cctiff, for example, wouldn't do.
> 
> Nothing really - you just avoid losing a small amount of colour 
> information by quantizing to 8-bits again after applying your printer 
> profile.   PhotoPrint keeps the data in 16-bit depth from the output of 
> the colour transformation right through to the driver.  How much that 
> matters depends a lot on how well-behaved your printer is, and thus how 
> well your printer profile makes use of the 256 codes available in 8-bit 
> data.

I thought it might be something like that.  So, is it true that I can't
do essentially the same thing with cctiff or lcms's tifficc?  Do they
perform calculations in lower precision?

If I have it right, then the comment that I was losing information was
not quite accurate.  It would be more accurate to say I was processing
the information at lower precision.  It would be the difference between
adding rounded off floating point numbers or adding first at higher
precision and rounding off afterwards.  I think I understand that sort
of thing pretty well as a mathematician, and I see that it could make a
difference, but I would be surprised if the difference were significant
compared to all the other variation present in real life color
management and color editing.

> 
> All the best,
> --
> Alastair M. Robinson

Other related posts: