Printer Resolution and Stuff

The variable dot size is defeated.

Thank you!

Ok, but AlanSh posted earlier he is happy with his prints and sells some, even after they have been resampled by the printer driver.

Many people buying and using consumer or prosumer level printers are very happy with the prints they output with printer driver resampling.

I’ve been shaving for more than 50 years now. During that time, not much changed. Sure, I’ve always used the latest razors — usually from Gillette. The shaving cream, always from an aerosol can, has changed a bit over the years. Stainless steel blades came along, and they lasted longer than carbon steel; double-edged blades morphed into single-sided razors with ever-increasing number of blades, and the shaves got a little closer, but, for 50 years the quality of my shave didn’t change much. For most of the last 25 years, I used one shaving cream, Edge. A couple of years ago, I had a cancelled flight and had to spend the night without my checked luggage. The hotel desk clerk supplied shaving cream, which I liked a little better than Edge.

That one little change ended 25 years of shaving stasis. I started randomly buying different kinds of aerosol shaving creams; some were good and some weren’t so good, but there wasn’t all that much difference between them. Finally, I happened upon a shaving cream from Italy that came in a old-fashioned metal tube (Proraso, if you’re in the market). It felt different, it smelled different, and it worked a heck of a lot better than all the aerosols. I went through a couple of tubes before I started to wonder if I was using it right. I bought a book on shaving (Imagine that! What’s next? A book on brushing your teeth?). It turns out that I was indeed using it wrong: I was supposed to froth it up into a lather with a brush before putting it on my face. I bought a badger brush and a little stainless steel rack so that I could hang it upside down and let it dry out between uses. It took some practice to get the ratio of water to shaving cream right, but I persevered until I could consistently achieve a rich, thick lather. The result: a closer shave than I had ever imagined; a difference far greater than 50 years worth of progress in razors. I’ve since found another shaving cream that performs almost as well (Musgo Real, made in Portugal), and I suspect there are many others out there.

You’re driving down the freeway in your car, and you’re thinking about changing lanes. You glance at your mirrors and you don’t see anything. But you know there’s a blind spot there, so you look over your shoulder before you turn the wheel. We’re talking about blind spots here, but they’re more insidious than the ones you deal with in your car. You know where they are. My ignorance about shaving was a blind spot I didn’t know was there. Call it a double-blind spot. I had been going along for most of my life thinking I knew everything I needed to know about shaving. It never crossed my mind that there might be a serious gap in my knowledge. Not only did I not know how to make my shaving experience better, I had no idea that it could be made better. For all that time, there was an opportunity for improvement that was completely unknown to me.

Donald Rumsfeld has taken a lot of heat for his didactic ruminations on “unknown unknowns”, but the concept accurately generalizes what we’re talking about here.

If unknown opportunities for improvement can persist in something as simple as shaving, they are far more likely in a complex and varied activity like photography. It is almost certain that there are things we can do to make our photography better, but we don’t know where to look for them.

As I posted earlier, people are entitled to make their own informed decisions based on their print quality threshold and needs.

Maybe some reading this thread might look into it a bit more and maybe adjust the way they prepare images for printing.

Others might conclude it's not going to bring a noticeable improvement in print quality.

It is their choice to make.

And here's some information that can help them make an informed choice:

blog.kasson.com/technical/resampling-for-printing-basic-alternatives/

You started with "PPI is used mainly for setting the physical print size for an image file." Pity if that is so. I was expecting you to debunk this misuse, btw, like you are trying to do with another subject, exposure.

I took that remark as a kind of shorthand. It would be less pithy if the dots were connected.

The print size in each dimension, in inches, for square pixels, is the number of pixels in the image in that direction divided by the image ppi metadata field.

Perhaps you could start a poll with a list of uses for ppi and let's see which turns out to be the most popular use.

That's a misuse. Take it from someone who used a rather large pre-press operation, with two Dolevs ;)

I don’t think it would be a misuse.

I see.
This is how we do it: we know that the optimal ppi (XResolution / YResolution) for a given printer / imagesetter / media (paper, film, ...) is N; we resize the image by scaling - changing pixel count along both axes (pixel dimensions, ImageWidth & ImageLength) so that for the necessary output dimensions N ppi are submitted to the driver/RIP. Thus, we don't use ppis to set the image size, ppis are fixed.
Why I mentioned Dolev: Dolev RIPs allowed to do the resizing on the RIP itself, based on the desired image dimensions, image ppis, and output dpis, and even to select the resizing algorithm / pattern (up to stochastic raster). Resizing was done at the rasterization step, theoretically the best point to do it. It was rather fast, but Photoshop worked better.

ImagePrint does the error diffusion at the resolution of the image you provide it, quantizing the dot locations, so it doesn't need a fixed pixel pitch in the input file. Sounds like the Dolev RIPs work the same way. I never did like the way that Epson did it, forcing the contone image to a pitch that's the resolution of the marking engine divided by an integer.

@IliahBorg

Ok, I sort of understand what you are describing but not completely.

The bottom line of my understanding is that to change the physical output print size for a given image file, the ppi value must be adjusted either directly or indirectly by inputing the print size dimensions, length and width, and the app then calculates the ppi value for that image size.

Is this not correct?

Depends on the app. Ps does it that way (sort of: you can't independently specify the print size in both directions if you want square pixels) if you tell it not to resample, but why wouldn't you let it resample if you're going to all the trouble of opening up the resizing dialog? Then you have to enter the ppi value.

Ok, so that is the point I was making in that the main use of ppi is to set the output print size.

It certainly is the only purpose for which I need to look at the ppi value as described earlier.

Actually, what you described is the reverse: You set the output size, and the editor picks the ppi meadata.

The way I do it, the ppi metadata number is set by the printer driver's native resolution. It doesn't affect the print size, since in Ps, I specify that, and Ps resamples the image.

Under the "Document Size": enter the target ppi ("Resolution"), next enter the target dimensions ("Width" & "Height", locked together to avoid distortion) in length units (inches/mm/picas/...), like this: