BIG difference between ABW and color printing with profiles. There are profiles pretty readily available for a variety of printers and papers via paper manufacturers' websites. Not so with ABW.
Well, in that there is no "manufacturer" in this case I would agree with you but there are a lot of people willing to help out someone with no access to a densitometer (see the last paragraph below).
As far as my comment not saying or adding much at all; that response does exactly what I outlined. It patently ignores the fact that a lot of people are not in a position to make their own custom profiles. Finding some through the The Print Yahoo group is fine but even then the variety available is unlikely to be close to what is available for color printing.
I said that the ABW driver is no more a "black box" than any other. Both need/should be used with profiles. It's a real shame that Epson missed a beat and didn't deploy this extension of their great ABW driver - and make available profiles for at least their papers and the standard settings (warm, cool etc). But then they've always been focused elsewhere and not on B&W. ABW was a good step forward but it's far, far away from current "state-of-the-art" B&W printing. It is, however, when coupled with QTR ICC profiles, tremendously convenient for the average user.
WRT a and b being unmanaged in the print process, should it matter? ABW expects to receive data in true grayscale, R=B=G. Given that, all that should be required is L. Or are you saying that a and b are required as well because the printer still mixes some colour ink even when printing via ABW?
I think you miss the point a little bit but I will attempt to answer your question. In crude terms, as you undoubtedly know, with a conventional colour profile and colour printing, we profile (as opposed to calibrate) a printer by measuring its colour response to a set of colour numbers
and then try - as much as possible within the available gamut - to get the printer the produce the right colours by altering the image file numbers "on the fly" as it is sent to the printer. Various methods manage any necessary gamut compression including the rendering intent and black point compensation.
When black and white printing (when not using a colour workflow), we use the driver to select hue and, typically, only send it a single channel file (or, as you note, an RGB file where R=G=B). In a RIP such as QTR, managing hue is done by selecting which inks (they may be colour inks or a graded B&W ink set which in turn may or may not have warm and cool inks for toning) are used, when and in what amounts. In ABW, this is managed with the (very convenient) hue picker (or standard settings) which drives how the ink channels are used/mixed. I may want a warm print (via either driver) and so not desire a=b=0 across the gradation from paper white to ink black, even though my image in PS appears as a (neutral) greyscale file (whether in an RGB or single channel workspace). If a QTR ICC profile tried to manage a* and b* then it would attempt to alter file values to reverse the selected hue in ABW. So in outbound management of the file, a* and b* are ignored. Rather we are only concerned with fitting the image's luminance ramp into the narrower print space (for, particularly for matt papers, the printer doesn't get close to a perfect black or even a perfect white). We record the luminance response of the printer for various stimulus numbers build a curve for that response and embed it into an ICC profile shell. Together with BPC, the profile manages the file to print space "luminance gamut" transition by changing the files numbers (either single channel or, in the case of an RGB image, with always R=G=B) to produce the appropriate, corrected L*. So this is what I mean by a* and b* aren't managed. a* and b* are left to fall where they may according to the shift in L*.
The "freebie" that comes along with using an ICC profile for this curve adjustment is that we can, however, record the a* and b* response of the printer - for soft-proofing purposes only. So you can create profiles for a given paper, printer, ABW settings and soft-proof the hue your image will have when printed. (Take a close look at a QTR ICC profile with Colorsync Utility.)
Epson's ABW driver still uses some colour when you ask it to print a single channel greyscale file, just as any driver uses its colour set to produce a colour when asked to deliver neutral grey (it just may not do a very good job as was the problem pre ABW and, some would argue, still today). But, importantly, ABW uses dramatically less yellow and no dark cyan or dark magenta. It also uses a lot more light-light black. Here are some numbers Roy Harrington recorded when ABW first came out:
"Since there's been talk about the two Epson driver modes I thought
it would be worthwhile seeing how they compare in ink usage.
The following "prints" were done to a file and analyzed to see what
commands were being sent to the printer.
Comparison of ink usage from the Epson R2400 with Epson driver
Image - grayscale 21 step wedge
Paper - Velvet Fine Art
Quality - Best Photo (1440x1440)
Both prints use 2 dot sizes, percents are weighted based on drop size
Epson ABW driver - neutral
total 100% dots 9,763,341 small 2,532,646 medium
K 31.27%
LK 26.15%
LLK 21.03%
LM 12.76%
LC 7.92%
Y 0.87%
Epson standard driver - Epson Velvet Fine Art profile
total 100% dots 10,393,768 small 2,694,351 medium
K 32.84%
LK 49.40%
LLK 1.29%
M 3.25%
LM 4.39%
C 2.90%
LC 2.37%
Y 3.54%
Some observations:
ABW uses a lot more LLK --- 16 times!
ABW uses a lot less Y --- 1/4 the amount
ABW uses no dark C and M
Interestingly this really supports the idea of replacing the Y
ink with a LLK ink with either driver.
Roy"
You don't need a density curve to see the differences, Alan. The difference between printing through ABW with and without a profile can be easily seen in regular prints of b&w images. And the differences are not always that subtle.
Agreed and that's why we like it.
While one needs a spectrophotometer to get the soft proofing part of QTR Create ICC, one only needs a densitometer to get the L* management. So it really can be extremely cost effective to adopt. And if one doesn't have even that, there's a community out there which is willing to help those that desire a profile for a particular ink, printer, ABW hue combination. I'm not suggesting that every imaginable combination is available, but there's a lot of help out there for the most commonly used papers and printers.