A note on the Canon interface
02:48

My friend Abhijit Menon-Sen writes:

The [Canon EOS d30 and 40D] cameras (and, as far as I know, all other EOS cameras) share one bad habit that has always annoyed me. In aperture-priority autoexposure mode, the wheel behind the shutter button controls the aperture and the dial on the back controls exposure (via a ±2 stop compensation). In manual exposure mode, these functions are gratuitously reversed, and the front wheel controls exposure directly.

It's clearer to think of it this way:

• In manual mode, the rear dial controls aperture, and the front wheel controls shutter speed.
• In autoexposure modes, the rear dial controls exposure compensation, and the front wheel controls the prioritized parameter; that is,
  • In shutter-priority, the wheel controls shutter speed.
  • In aperture-priority, the wheel controls aperture.
  • In program, the wheel "swaps" shutter speed for aperture, changing each so that the exposure is maintained.

The first prints.
02:49

This photo courtesy of my new Samsung Alias 2 phone, which is pretty sweet. On the other hand, boo to Apple: Mac OS X 10.5 doesn't support Bluetooth BIP, so I couldn't dispatch this photo to $coworker's Mac and post it earlier today. Instead I read it off the MicroSD using my netbook's builtin SD reader (about which more below).

I see a long USB cable in my future.

I made a few prints on normal office paper and they're pretty awful, but I also bought some Epson Ultra Premium Photo Paper Glossy (figuring it'd be good to use up Epson's inks), so I thought that and Epson's canned profiles would give me better results.

Initially I just plugged in the printer, printed "Hello world!" and was happy. But then I saw that Epson's documentation says to install the software first, so I spent a few hours screwing around with that, because the first time I installed, Epson's canned profiles showed up in Lightroom's print module as color management options. But then they didn't show up upon the subsequent reinstalls. I finally hunted them down and... somehow... Got them reinstalled.

Finally I set up soft proofing using the glossy profile, and that warned me that the blues in my picture of Russian Hill are way out of gamut, so instead I'm going to print this picture:

I stuck two copies on the same sheet; the right-hand copy had brightness +35.

This print follows the guidelines in Epson's managing color document for the R2880:

File > Page Setup to select the 1400 and the appropriate paper (8x10 borderless). Then Print, and "Photoshop Manages Colors". The profile should be "SP1400 1410 UPGP". Rendering set to "relative colorimetric", "black point compensation" checked. Then hit "print", and in the Epson dialog... "Print Settings", select highest quality. It's already got the right paper. Turn off "high speed". Then select "Color management", turn it off... And let 'er rip.

So, while waiting for that to print... I installed my new SSD. I hope this one will last longer than an afternoon (it's already survived a couple of hibernate cycles, so...)

Eeebuntu 3.0 installed painlessly. I like it, despite GNOME.

(Ah, that took about 8 minutes to print.)

So, the print came out with a lot of wet ink beaded on its surface. That's... Not good. I'll try printing again; this time lightening the whole image 45, and I'll try with quality Photo instead of Photo RPM.

(4-5 minutes later)

Ok. That's much better. Note, don't use Photo RPM.

The paper instructions say dry for 15 minutes and then put it under a sheet of normal paper for 24 hours; if the normal sheet's wavy, replace it with a new one and check it again 24 hours later. So I'll have a look at these tomorrow morning and afternoon, I think. I'll probably also buy a pack of 13x19 Epson paper.

Christmas day two.
12:49

Ah, I just received my inks. And now Canon tells me my repaired 28mm f/1.8 has shipped. So good news all around!

Printing, round two.
22:53

So this time, I'm using a picture from my second trip to Alcatraz:

So I soft-proofed it, and reduced the hue of the yellows and reds until they were within gamut. The other test print is a fair bit more saturated than the same image on-screen, so I think think the result won't be as washed-out on paper. OTOH this demonstrates that soft-proofing has limited value... Maybe a custom profile would help. But not today.

Anyway, I exported this pic from Lightroom with its standard sharpening for glossy paper. I should print a few "proofs" of this image on an 8x10: Straight, brightness +15, brightness +30, contrast +15, contrast +30.

Also, some sharpening tests - a detailed part at 100%, then a couple more with USM or smart sharpen or something. I doubt I can fit that all on one sheet though.

If I do a "final" of this image I think I'll the sky could do with a bit more blue.

And here... we... go...

> Z
Time passes.

Ok, the grid I printed is quite a bit more saturated than the image on-screen (thanks for not much, profiles). The shadows in all of them are kind of an ugly green, and fairly blocked up. Brightness adjustments are more useful than contrast.

For fun I converted the image to grayscale, curved up the shadows significantly, and printed that. Fresh out of the printer it looked pretty good - but after drying for 10-20 minutes there are obvious color casts: Green in the shadows, bit magenta in the midtones. But it's smooth! Tonal range looks pretty good.

Finally, I made an 8x10 color print. Again, I used a curve to bring up the shadows a fair bit; the detail's ok, but the shadows are really greenish So, I'm not sure quite what's going on there.

Printing, round three.
22:53

I decided to give Mike Reichmann's From Camera to Print tutorial a look, and after watching a preview clip, I thought that maybe the shadows in my Alcatraz pictures were so ugly because I lowered the yellow and red saturation before printing (to get those colors back in-gamut). That, it turns out, wasn't it.

It's still ugly. And the detail rendition in a lot of areas is poor. In the past, e.g. using Shutterfly, I've had pictures come back that matched the screen colors (though brightness was a bit of a problem, unsurprisingly). So I think I'm doing it right... But these results are not satisfying. I suspect this is a profile problem (although I usually first suspect myself of doing something wrong). Maybe there are some other guys on yahoo who have some feedback...

Printing, round four.
08:57

I've done some googling and there are some interesting threads out there - like this one - and so I thought to myself... Let's see how the printer handles color management, in case Photatochop has bad profiles.

The answer is: At first glance, much better. But I had better let it dry and see.

15 minutes later, it's not significantly better than the other prints: The shadows block up with an ugly greenish color.

This is disappointing. I wonder if there's any reasonable way to find out if I'm suffering from double-color management transformation...

Later

I just did a print head alignment, which took a couple cycles of recalibration. I'll try a last color managed print and then swap in the blacks... I'm already tired of diddling around with this.

Yeah, that didn't help too much. It's still too dark. I'll let it dry down and see how the colors are.

In the meantime I'm going to get the MIS inks in there.

(Ok, I gave in to temptation and ran a copy of my brides photo. That also looks terrible - way too dark. Are the canned profiles that bad?)

Looks like this is the process:

• Swap cartridges.

  Now that was amusing, since the prefilled cyan cart was DOA. So I also learned how to load a cart. Hope I did it right.

• Nozzle check 0

  This has some gaps, so I'll check it again after the cycle.

• Cleaning cycle.

• Nozzle check 1

  This has different gaps. So.

• Cleaning cycle.

• Nozzle check 2

  Another gap.

• Cleaning cycle.

• Nozzle check 3

  Another different gap.

• Cleaning cycle (wow, that's used almost as much ink as I used printing...)

• Nozzle check 4

  Ahh, pass!

• Head alignment.

  And that's great.

• 5 pages of MIS's purge pattern.

Unfortuntately I don't have a paper suitable for the UT14s - the closest I have is the office supply junk I bought. Ohwell. I'll waste a sheet of photo anyway.

Christmas day three.
15:56

Woo, I got my repaired 28mm f/1.8 back. Finally... I need some paper...

Duh.
01:05

Earlier I wrote

• 1400-UT14-UltraPremGlossPP-N-1-rgb.icc Not sure what this is.

Ya think it might be Epson's Ultra Premium Glossy Photo Paper? Apparently the older name for what Epson now calls Ultra Premium Photo Paper Glossy.

Hi, my name is Mike, and I love to overlook the obvious.

Still, using that results in dark prints. I finally did what I should have done at first: printed one of Paul Roark's step wedges. The bottom four patches are all black. So I guess the best plan is to try to find a curve that matches that, then apply the inverse, and that should help.

Later

I have a sample image from The Light Room which includes a step wedge, so I can visually compare that to my printed results. Basically I printed the wedge, used curves to make the on-screen image look like the printed result, and then reopened the original, and used curves again (with the eyedropper) to sample the original and modified images and build an inverse curve. Essentially, profiling by hand.

So a few tests and adjustments later and the printed wedge looks fairly close to the Light Room's sample. I just printed a sample image with the compensator applied, and, yeah, that's fairly close to the screen image - there's not enough separation in the brighter tones, so the image's tonality suffers. But this is the best result so far, anyway.

But this all seems... Almost silly, really. The whole point of color management is to avoid this kind of nonsense. I think I must be doing something wrong. So I think I'll ask the guys over at Yahoo how much compensation they have to do.

The Way of the Cardboard Tube.
23:31

So, there's this thing called the Cardboard Tube Fighting League. Sunday they had a tournament or duel or... showdown, or whatever it's called. Naturally I took some photos:

They're on Flickr.

Latest developments.
18:10

So, on Paul's advice, I decided to see if maybe there was a driver screwup on my Mac. Here's what I did.

First I switched the Epson inks back in and ran a few purge pages; I did a test print or two from the Mac, and they match what I got before.

Installed Eye-One Match on my PC and profiled my monitor (I didn't actually change any settings, just profiled it - the settings were still correct from my last bout of profiling, although the luminance is 114 instead of 100, which is curious).

Installed Photoshop Elements 6 on my PC (neat, the printer included a copy).

Installed the Epson driver plus other software. I carefully did not plug in the printer until the software installation told me to do so. That appeared to go well.

Fired up Photoshop Elements and loaded a test image - it's in Lab color, which Elements doesn't support, so I converted it to RGB.

I followed Epson's instructions to print it. Yes, I chose the correct printer profile in the Elements dialog, and the correct media type in the printer driver dialog.

It came out exactly the same as the prints from my Mac - dark, oversaturated.

Then I tried printing again, but with Photo RPM quality instead of Photo quality. And, once again, the print came out with blotches of wet ink on it.

Gr.