Optimized subsampling: bug and workaround

Question

(CS3 and CS5. Mac).
InDesign appears to have some really bad defaults for the printer internals of "Send Data: Optimized subsampling" in the Graphics pane of the print dialog.
We moved from an HP 8150DN to a Xerox Phaser 7500DX. On the Xerox, 1bpp images come out looking abysmal when Optimized subsampling is turned on. On the HP, they were fine. Here's a scan of the difference. The left image is optimized subsampling on the Xerox (in case you couldn't tell!); the right is "Send Data: All" on the Xerox:
Why does this happen?
Well, Adobe's documentation isn't so good, but Real World InDesign CS5 (Olav Kvern, David Blatner, and Bob Bringhurst; who is Olav, anyhow?) tells us:
Black-and-white (bi-level images). When you're printing bi-level, black-and-white images, and have selected Optimized Subsampling, InDesign matches the images it sends to the resolution of the output device. … FOR THOSE WHO REALLY WANT TO KNOW, INDESIGN GETS THE PRINTER'S RESOLUTION FROM THE DefaultResolution KEYWORD IN THE PPD.
[EMPHSIS mine]. It appears the PPD situation with the Xerox printer is fairly weird. Fundamentally, Xerox ships a PPD file that lacks a DefaultResolution keyword. So it appears that InDesign has a very bad default -- it presumes that a PPD that specifies no resolution means it should assume a resolution of 25dpi or so.
If we edit the PPD file to add:
*DefaultResolution: 600dpi
*SetResolution 600dpi: ""
then suddenly Optimized Subsampling works fine.
It turns out the reality is a bit stranger. In Xerox's Mac drivers for the 7500DX, they ship "Xerox Phaser 7500DX.mpkg", an OSX package file. Inside that file they ship several Contents/Archive.pax.gz files (basically tar files), one in Contents/Packages/xeroxPrintDriver.pkg and one in Contents/Packages/xps104.pkg. The xps104 version is the one that gets installed, it's a 724440-byte file that ends with "*% End of XR7500DX.PPD, 724439 bytes."
But the xeroxPrintDriver version is different -- among other things, it has a *DefaultResolution line:
*DefaultResolution: 600dpi
*SetResolution 600dpi: ""
*% === Halftone Information ========
*DefaultHalftoneType: 9
*ScreenFreq: "60.0"
*ScreenAngle: "45.0"
*DefaultScreenProc: Dot
*ScreenProc Dot: "{180 mul cos exch 180 mul cos add 2 div} bind"
*DefaultTransfer: Null
*Transfer Null: "{ }"
*Transfer Null.Inverse: "{1 exch sub} bind"
*DefaultColorSep: ProcessBlack.60lpi.600dpi/60 lpi / 600 dpi
*% === For 60 lpi / 600 dpi  ===============================
*ColorSepScreenAngle ProcessBlack.60lpi.600dpi/60 lpi / 600 dpi: "45.0"
(Its a 617603-byte file with "% Date: 07/06/05" in the initial headers and ends with "*%end of PPD file for Xerox Phaser 7500DX")
Maybe we should just switch to using the xeroxPrintDriver version, I don't know.
Really, this situation is quite annoying. Xerox installs a print driver that lacks a *DefaultResolution, screwing over InDesign users who use the default setting of Optimized subsampling. InDesign has a horrible default that chooses 25dpi instead of 300dpi or 600dpi if no resolution is specified.
Does anyone know who the right person at Xerox is to get their side fixed? I suppose we'll submit a wishform bug report to Adobe on this (grr, I hate being a volume license customer sometimes. Maybe if I buy a single copy of InDesign at street rates then can I submit real bug reports and get useful responses and tracking numbers? Anyone know?).
Oh, some of you will ask why we use optimized subsampling at all? The short answer is that it often makes a material difference in the time to print proofs of high-resolution jobs. Of course real jobs are not printed in-house on these printers, but we do proofs all the time. And a casual benchmark showed one job going from 18 minutes to 36 minutes, and another going from 5 minutes to 18 minutes; that's from switching off Optimized Subsampling. So it does have some real value for us. (I think, though, there might be a network problem with the printer those benchmarks were done on. Not sure. But the person doing the benchmarks did them on our slowest hardware to get a worst-case estimate.)
Any thoughts?

Answer 1

Some excellent sleuthing work.
As for
I suppose we'll submit a wishform bug report to Adobe on this ...
-- the wishform is nothing more than that, "I wished ..", and probably (likely ) Adobe only goes through this when they have absolutely nothing else left to do.
..  (grr, I hate being a volume license customer sometimes. Maybe if I buy a single copy of InDesign at street rates then can I submit real bug reports and get useful responses and tracking numbers? Anyone know?).
I wouldn't put my faith in Adobe's Customer Support either. They would probably give it a half-hearted try at their end (which of course works fine because they don't use a Xerox), tell you they can't reproduce the problem, then tell you not to use optimised sub-sampling, advise you to test with some other printer, and finally state "it's not our problem, complain to Xerox" and hang up on you.