More On Print(er) Sharpness

By Ctein

In my column on January 5th, "How Sharp Is Your Printer? How Sharp Are Your Eyes?", I described a relatively easy and robust way to test how sharp your printer was and how sharp you needed your prints to be. I recommend that readers reread that column and the comments before reading the rest of this one.

One of the questions I got from a couple of readers was whether or not printing at a printer's "magic" resolutions (360 / 720 ppi for Epson printers) would make a difference in the results I got. While I suspected it wouldn't, suspicion is not the same as experimental data. So, this week it was back to generating resolution test pages and pixel peeping.

I generated four more pages to accompany the ones I made for the last article. That provided me with comparison pages at 300, 360, 450, 600, 720, 900, and 1200 ppi. I printed everything out on my Epson R800 printer on Ultra Premium Luster paper at the highest quality settings. Time to start squinting. In the set of four running from 300 through 600 ppi, the results were unequivocal and monotonic; even with the naked eye I could see that each jump in file resolution resulted in a sharper print. Comparing the 360 ppi print to the 300 and 450 ppi prints, I could see single-pixel-scale artifacts in the latter two that didn't appear in the 360 ppi print.

These two illustrations, above and below, show highly magnified sections of the test prints produced as described in the article. From left to right, they're from the 300 ppi, 360 ppi, and 450 ppi files. Observe that there's nothing "magic" about 360 ppi; it's in between the other two in sharpness. the 600, 720 and 900 ppi files show similar results.

As I mentioned last time, I have pathological test targets that can demonstrate that difference. As I also mentioned, it's not relevant to printing real-world images. The 450 ppi print definitely looked sharper than 360 ppi, which looked sharper than 300 ppi, and the improvement in sharpness with each jump in resolution seemed just about right. In the real world, gross resolution trumps single pixel artifacts any day.

(Sidenote: There's no point in looking at intermediate resolutions to these—you'd be unable to see any differences in sharpness with such small resolution changes.)

As for the test prints between 600 and 1200 ppi, the same monotonic relationship held, but now you're down to the pixel-peeping area. prints get minutely sharper with increasing resolution, but I'll never be able to notice it in real life. I'm not convinced there is any improvement, even microscopic, going from 900 to 1200 ppi.

On a purely practical level, it's clear to me that there are folks out there who've missed the (real) forest for the (ideal) trees. They argue that everything should be printed at a magic resolution, ignoring the reality that your photographs don't magically come at those magic numbers. Resampling a photograph that's going to print out it 300 ppi at the size you want up to 360 dpi is not going to make it look visibly better. Resampling a 450 ppi photograph down to 360 ppi will cost you sharpness and fine detail.

Now, all my test photographs were printed out natively through Photoshop. Start bringing third-party RIPs into the picture, and all bets are off; the difference in appearance between a "straight" print and a RIPped print has a lot more to do with the rendering engine than the ppi it is rendered at. If the RIP wants to render at some magic resolution, more power to it. Just understand that the improved quality you see is much more a consequence of how was the data is being massaged for the printer than the resolution it's being output at.

By the way, no, I am not going to compare the various third-party RIPs versus Photoshop. Doesn't interest me. Photoshop works well enough for me. Just so you know. Your mileage may differ.

I think that's it for this subject (barring more intriguing questions that I haven't thought of, of course).

Next week, I hope to be able to revisit the Snow Leopard color management issue and the week (or two) after, an interesting take on the question, "How big can I print this?"

Ctein

Ctein's regular weekly column appears every Thursday morning.