Digital BW, The Print

Hi

Earlier I communicated directly with Carl Schofield about a problem I observed using 
OPM. His response called attention to this thread and pointed out that Diana 
apparently encountered the same problem. Our correspondence is shown below.

As a result of his inputs I did a test that  responds to Antonis's suggestion to Diana. 
Here is what I reported to Carl.

" I did print your Milkweed file (Wonderful Photograph). Printed at the original 8-Bit/
360 dpi configuration there was serious "stairsteps" in the text - particularly in the 
legs of the "M". The photo itself was OK to the eye, but with a loupe even there some 
starstepping was evident. I reprinted the image at 16-Bit/360 dpi and the results 
were near perfect even when examined with the loupe. I don't understand what 
happens when an 8-Bit file is changed in PS to a 16-bit file. It's not clear whether 
artificiality is introduced as during res-ing up.  But apparently, it does make a big 
difference to OPM.

 I did a series of prints of a particularly troublesome section of my test file varying 
Bit-Depth and Resolution. The test file was originally 8-Bit/240 dpi (which by the way 
printed very well through the Epson 2200 driver but poorly through OPM). I printed 
the section all the way up to 16-Bit/720 dpi. My conclusions were that, to the naked 
eye, increasing the resolution to 480 dpi at 8-Bit or just converting from 8 to 16-Bit 
yields a satisfactory print. With  an 8x loupe, the criteria increases to 8-Bit/720 dpi or 
16-Bit/360 to 480 dpi.  I think I could still see improvement even up to 16-Bit/720 
dpi, although only marginal improvement.

The results indicated that changing from 8 to 16-Bit yields roughly the same 
improvement as doubling the resolution. So, for the Milkweed print at 16-Bit, I am 
seeing about the same fidelity as when I print at 8-Bit/720 dpi. The results with my 
test file are comparable."

Doug Nielsen
Epson 2200
MacOSX 10.2.6/Classic 
Blue&White G3
____________________________________________________________

Here is the previous correspondence with Carl:

From: Douglas Nielsen <douglasnielsen@...>
Date: Fri May 16, 2003  11:21:15  AM America/Phoenix
To: scho@...
Subject: Bowhaus Inkjet Control and OPM Software

Carl

I have seen your posts in the Canon SLR Talk, Epson Talk and Mac Tools dpreview 
forums regarding the Bowhaus software. I have downloaded  the OJC and OPM 
software and have successfully printed a number of samples  using your profiles. I 
share your enthusiasm with the results. No color castes and extremely little 
metamerism if any. Particularly impressive to me is the ability to blend your warm and 
cool profiles to the mix of my choice. However, I encountered a significant snag.

I am not knowledgeable enough to understand the cause, but prints produced with 
OPM and your profiles appear to be of very low resolution - ie "jaggies" are clearly 
visible. Diagonals in the images as well as in text text are unacceptably 
"stairstepped". Since OPM doesn't provide any user control of resolution, maybe OPM 
just does not print at a high enough resolution. However I suppose the problem could 
also be that OPM uses a dither pattern that produces an apparent limited resolution, 
or possibly printing using only four of the inks could leave "holes" in the dither 
pattern. Finally the problem could just be with me or my system. (I am running OSX 
10.2.6 and Classic on a Blue and White G3) In any event this is a show stopper for me.

All that said my questions are 1. Have you observed similar results? and 2. Do you 
know of any solution?

Your thoughts would be appreciated.

Doug Nielsen
Albuquerque NM


From: Carl Schofield <scho@...>
Date: Fri May 16, 2003  1:04:27  PM America/Phoenix
To: Douglas Nielsen <douglasnielsen@...>
Subject: Re: Bowhaus Inkjet Control and OPM Software

Doug,

I did find one thread in the Digital BW archive where an OPM user ran into a similar 
problem with stairstepping.  I copied the content of the thread below along with the 
link.  Suggestion to her was use 16 bit mode and increase resolution to 720 dpi.  As I 
said earlier, I haven't found it necessary to use high resolution (more than 360 dpi) 
but there may be some strange things going on with particular systems where this 
helps.  Also, Bowhaus told me that there is an upcoming upgrade for both bug fixes 
and OS X compatibility so when the newer version is ready it might be worth trying.  
Finally, I uploaded an 8 bit greyscale tif file to my public filesharing folder (http://
homepage.mac.com/scho/FileSharing2.html) that contains both an image and some 
text with strong diagonals (it is actually a cropped section from a larger triptych 
image file) that prints fine with OPM and my profiles on my 2200.  The file name is 
"milkweed.tif".  You might want to try printing this file and see if you still have 
problems with jaggies.

Carl

http://groups.yahoo.com/group/DigitalBlackandWhiteThePrint/message/29265


--- In DigitalBlackandWhiteThePrint@yahoogroups.com, "Antonis Ricos" 
<antonisphoto@y...> wrote:
> Hello Diana;
> 
> I tried to restore this thread under its original subject heading and so repeated 
> our previous messages below for anyone who wants to follow this. It's easy to 
> hit reply and end up with "Digest" as your subject!.... 
> 
> I am glad you saw an improvement in your results with the resing-up tip. There 
> is continuing development on OPM, with a beta version currently being 
> tested for the 2200 (no new features, just tweaks to address some of the 
> issues users have reported). Should you buy IJC, you will find that it is 
> capable of controlling up to 7 ink channels with smooth spline curves that 
> can accept 20+ points each (normally 2 or 3 are plenty). That gives 
> unprecedented controll over the amount and combination  of inks to be laid 
> down for each digital value in your image file.  As has been discussed here 
> before, for the full benefit  you would need a densitometer, but good results 
> are also possible by visually comparing your output to a good print of the 
> 26-step grayscale.
> 
> A note about comparing printing software: Comparisons between OPM and 
> any other bw printer driver would generally have to assume perfect profiles 
> for each, so that the drivers are in fact compared, not the profiles. This gets 
> tricky if drivers that rely on canned profiles are fixed to a certain performance, 
> while OPM uses custom profiles (made in IJC) that can take into account 
> every nuance of your particular set up. This is less of an issue with printers 
> like the 2200 using OEM carts, and a bigger problem with CIS-equipped 
> desktops and large formats.
> 
> I hope you can let us know what you think of the results your are getting.
> 
> 
> Antonis
> 
> 
> ------------------------------------------------------------
> Hi Antonis,
> 
> I took your advice about rezzing up a file to 720 dpi and changing to 
> 16 bits and I got MUCH better results with OPM, at least for 1 
> particular image with lots of thin strong diagonals. There had been 
> noticeable (but fine) stairsteps when the file was 300 dpi and 8 
> bits. The same file printed with Roy's QuadToneRIP under OSX didn't 
> have the stairstep problems though - maybe his driver can rez up the 
> file like the Epson driver.
> 
> I think both programs/drivers have alot of potential. I haven't 
> purchased Inkjet Control (yet) but the fact that their software lets 
> you tweak every ink channel is very nice. With QuadToneRip, you tweak 
> 4 channels in CMYK mode, regardless of printer model (at least that 
> is the case for the 2200 curves). At least with both programs I'm 
> seeing no metamerism with the Matte Black ink on EAM and Eclipse 
> Satine, whereas the Epson driver always gives me greens and/or 
> magentas!
> 
> By the way, I have a Mac G4 and am running OPM under OS 9.1. I run 
> Roy's QuadToneRIP under OSX 10.4.2.
> 
> Diana
> 
> 
> 
> 
> 
> 
> 
> --- In DigitalBlackandWhiteThePrint@yahoogroups.com, "Antonis Ricos" 
> <antonisphoto@y...> wrote:
> > Diana,
> > 
> > I have discovered that some images look sharpest at 720 dpi when printed 
> > with OPM  on the 2200. If you open your image in Photoshop and check the 
> > resolution it should read 720 dpi. If not, you can res it up using bicubic 
> > interpolation. Then try printing again.  
> > 
> > This may or may not make a big difference in your print depending on how 
> > close you were to 720 dpi before you resed up.  It also depends on what 
> kind 
> > of image it is - how grainy, how sharp, if it has sharp diagonals... etc.
> > Also, if you still see problems, it's worth  setting the image to 16 bits and 
> > comparing prints.  
> > 
> > Testing tip:you can crop a small area where a problem shows up and print it 
> > over and over on the same sheet with different settings. You set the x,y 
> > positions in OPM to avoid overlaps.
> > 
> > Just to be clear about the 720dpi "preference": It is unrelated to the amount 
> of 
> > information a scan or digital camera file contains; it has to do with matching  
> > pixels-to-dots and driving the printer at its native hardware resolution. The 
> > Epson driver (for the 2200 at least) does this interpolation internally, so no 
> > matter what res you throw at it, the image will be made to look its "best" as 
> far 
> > as the printer goes.
> > 
> > I am very interested in your comparisons between Roy's RIP and OPM - 
> > especially since they share some DNA between them ! (and didn't that 
> > originate at MIT?....) The tonal gradations in OPM are mostly the result of a 
> > particular profile and whether your particular system performs as well as 
> mine 
> > (since I made the first 2200 profiles).
> >  In theory, you are supposed to have IJC and be able to tweak the profiles 
> > both for color and for smoothness. This system is really aimed at putting an 
> > end to canned profiles of any kind (including its own) ! 
> > 
> > As we discuss this further, it would help if you mentioned your operating 
> > system and other details of your setup.
> > 
> > If you are experiencing problems with the 2200, you may want to write to 
> > bowhaus and see if they will let you try the current beta which has some 
> fixes 
> > for the 2200. Their email for OPM / IJC is: software@b...
> > 
> > Antonis
> > 
> > 
> > --- In DigitalBlackandWhiteThePrint@yahoogroups.com, Diana Park 
> > <dip@m...> wrote:
> > > Antonis,
> > > 
> > > I've just started experimenting with OPM on my 2200 and I'm not sure 
> > > what you meant below - should the file actually be set to 720 dpi or 
> > > should the printer output be set to 720 (as opposed to 1440 or 2880)? 
> > > I haven't purchased IJC so I can't control the curves supplied with 
> > > OPM but I do like the Blender option where I can mix warm and cool 
> > > curves. I've found 70% warm and 30% cool to give a nice neutral (to 
> > > my eye) tone.
> > > 
> > > I've also FINALLY successfully installed Roy Harrington's QuadToneRIP 
> > > (many thanks Roy for all your troubleshooting help!) and while his 
> > > neutral curve is warmer than the above mix in OPM, the tonal 
> > > gradations seem smoother. Now for some curve tweaking!
> > > 
> > > Diana Park
> > > dip@M...
> > >