I confess, I still do not understand what the harm of soft proofing to sRGB might be and I would like to learn why.
Sure. The idea is to soft proof to a color space that isn't the one you're currently viewing (an sRGB image soft proofs in sRGB). The only reason to soft proof to sRGB is if the output is to an sRGB device, meaning an emissive display; the internet, a mobile device. There is no such thing as an sRGB printer! I tried telling that to Ms. Snider, she clearly didn't get it. In fact she wrote that:
Sure, soft-proofing in sRGB is pointless for printing to your own inkjet printer, or for something printed on a commercial press, but it's crucial for using an online printing source, including Apple. All the online labs mentioned in the article confirmed that print in sRGB colorspace. My understanding is that they print in a space similar to sRGB. But when I emailed all of those companies, they emphatically stated that yes, the images are printed in something very similar to sRGB. Not sure how to argue with that?
So either the labs were confused or she is (or both). I provided gamut maps for several contone chromagenic printers that show they are not like sRGB, sent her the profiles and suggested she soft proof to see the difference. More push back from her.
OK, so we have an image in Adobe RGB (1998) like this one I recommend you download for a later test:
http://www.digitaldog.net/files/2014PrinterTestFileFlat.tif.zipWe want to see how it will appear when we print it. We select a printer output profile, say one for a lab using a Lightjet. We invoke the soft proof, the simulation check boxes and the idea is we see a simulation of how the print will appear. This has nothing to do with sRGB other than sRGB may be the source color space for conversion to that printer output color space. Just as in her article, one might be in Adobe RGB (1998) and need sRGB because the lab demands that. Well guess what? Just convert to sRGB, you're now soft proofing sRGB but NOT the printer output color space. I tried to explain to Ms. Snider, printer profiles have multiple rendering intents while sRGB provides just Colorimetric. I suggested she try toggling with an actual printer profile between RelCol and Perceptual then sRGB to see that they don't appear the same. More push back.
Someone mentioned the "preserve RGB numbers" choice. What was the implication of that comment?
This is where Ms. Snider really screwed the pooch in her article
. What
Preserve Color Numbers does is answer this question: What would the output look like if I sent it with the current color space as is,
not convert to the output color space? IOW, show me how ugly Adobe RGB (1998) will look if I send it to the web without converting to sRGB, or show me how ProPhoto RGB would look if I sent it to my Epson 3880 for Luster paper but did not convert to that Epson 3880 profile. It looks awful. What Ms. Snider's huge, egregious error in her article did was recommend people pick the soft proof option:
View > Proof Setup > Internet Standard RGB (sRGB). That DOES invoke the preserve RGB numbers option! See for yourself: download the image above. It's in Adobe RGB (1998). Use
View > Proof Setup > Internet Standard RGB (sRGB) and see how it appears. Now select sRGB in the
Customize Proof setup instead. TURN OFF
Preserve Color Numbers if on. See the difference? Ms.Snider screwed this up big time and it's quite clear she didn't even test the silly recommendation she made! She tells her poor readers that by soft proofing to sRGB, they will get some idea of how the print will appear: wrong. But worse, she provided a recipe to have them see a soft proof (
View > Proof Setup > Internet Standard RGB (sRGB)) that's totally wrong and
she clearly doesn't understand why. I tried to point this all out in a series of emails to her and I assume her editor she CC'd (Jay Nelson). I suggested they fix the article, simple to do. Here's the last email I got back from Ms.Snider:
On Dec 5, 2015, at 12:30 PM, Lesa Snider <**> wrote:
BUGGER OFF
L e s a S n i d e r
So much for peer review, or the idea that Macworld's author’s, certainly Lesa Snide
r have any desire to disseminate useful or accurate information to their readers. Very sad.
** it's an easy email to find from her blog but I'll leave it out due to proper internet forum respect, not that she deserves it. Email me if you want to contact her.
