Better to resample or change resolution?

"Larry Bud" wrote in message

For print quality, am I better off keeping the pixel dimensions the same and just adjusting the resolution, or should I resample the image? Or does it not matter?

For example, I want to print a 5×5" image. The image I have is 500×500, 100 dpi. Should I resample to 1500×1500, 300dpi?

A 500 x 500 (I’m assuming pixels) *is* a 5" x 5" image.

For print quality, you want to de-select resampling and adjust the resolution. In your case, your 5 x 5 image will now print at 1.6" x 1.6" high-quality (300 dpi).

If you have a 500 x 500 image, and make it 1500 x 1500, what makes you think there will be more pixels per inch than what you started out with, even if you did specify a higher resolution? Where did all of the extra pixels come from?

In contrast, if you do not resample and adjust the image… This has the effect of ‘compacting’ all of the pixels in your image to a higher density (higher dpi).

—
SamMan
Rip it to reply

In article ,
says…

For print quality, am I better off keeping the pixel dimensions the same and just adjusting the resolution, or should I resample the image? Or does it not matter?

For example, I want to print a 5×5" image. The image I have is 500×500, 100 dpi. Should I resample to 1500×1500, 300dpi?

You want to scale (keep pixels the same instead of resampling) anytime it is halfway convenient to do that, that is, when the results are usable. I’d define usable as being 150 to 350 dpi, with the optimum target ballpark being 300 dpi whenever possible. I’m not saying 150 is great, but since that is all the pixels you have, then you have no choice, just print them (or better, print it smaller).

Your example of resampling 3x larger is undesirable – it cant help photo detail. That is, it certainly cannot magically give you 300 dpi of detail in this case if all you have is 500×500 pixels. Printing at 100 dpi is pretty low however (this image is too small), so conceivably upsamping a bit, say 150 dpi, possibly might make the pixels appear less visible, but it cannot help add detail, and so generally isnt worth considering. Because, this example case is simply too small, and wishing it were not so really wont change anything.

But when the image is too large, say if scaling would print it at say 1000 dpi, that’s excessive too, and resampling it SMALLER to say 300 dpi is a good thing to do then (followed by modest USM sharpening). Resampling smaller is never a problem when the need is smaller.

—
Wayne
http://www.scantips.com "A few scanning tips"

In article <BtG5e.9876$>,
"SamMan" wrote:

A 500 x 500 (I’m assuming pixels) *is* a 5" x 5" image.

Only if it is 100 pixels per inch.

If it is 200 pixels per inch, it is a 2.5" x 2.5" image. If it is 50 pixels per inch, it is a 10" by 10" image.

For print quality, you want to de-select resampling and adjust the resolution. In your case, your 5 x 5 image will now print at 1.6" x 1.6" high-quality (300 dpi).

If you have a 500 x 500 image, and make it 1500 x 1500, what makes you think there will be more pixels per inch than what you started out with, even if you did specify a higher resolution? Where did all of the extra pixels come from?

If the image is resampled, the extra pixels are made up out of thin air by a mathematical process that ‘guesses" what they should be. However, inventing pixels out of thin air does not add detail not present in the original; nothing–no program, no algorithm, no technique–can increase the number of pixels in an image and have the result contain more detail than the original.

The simple answer to the original poster’s question is:

You can’t. You can not take a low resolution image, make it be high resolution, and end up with a high-quality image. It cannot be done–not by Photoshop, not by anything. The hardest lesson for people to learn when they enter the world of digital imaging is that the gorgeous low-res image you see is utterly useless for printing in high resolution. You have to start over form scratch and create the image properly to begin with.

—
Art, photography, shareware, polyamory, literature, kink: all at http://www.xeromag.com/franklin.html

"Tacit" wrote in message

In article <BtG5e.9876$>,
"SamMan" wrote:

A 500 x 500 (I’m assuming pixels) *is* a 5" x 5" image.

Only if it is 100 pixels per inch.

Correct… I was using his specifications as the example…

—
SamMan
Rip it to reply

The simple answer to the original poster’s question is:

You can’t. You can not take a low resolution image, make it be high resolution, and end up with a high-quality image.

That’s not really the question, though. In my example, a 500×500 pixel image at 100 dpi will print at 5"x5". If I resample to 1500×1500, 300 dpi, it will also print at 300 dpi.

Now, in both cases, is the printer throwing down the same amount of ink to paper? If so, WHERE is the additional information coming from in the 100dpi example if the printer is a 300dpi printer? Is the Photoshop still resampling, or is the printer driver doing something?

Let’s take this a step further. If I have a 10×10 pixel image, at 10dpi, I will get a 1"x1" image when printing. What is doing the logic to get 10 dots to cover 1" wide?

Your example of resampling 3x larger is undesirable – it cant help

photo

detail. That is, it certainly cannot magically give you 300 dpi of detail in this case if all you have is 500×500 pixels.

No, I understand that, but I just ran a little test and there is a difference…

Take a 10×10 pixel image at 10dpi, and do a gradient from black to white from left to right and print. You pretty much get 10 equal bands when you print.

Now, resample this to 300×300 pixels, 300dpi, and you’ll see photoshop soften the edges. When I print this (with an epson c84), I get a much more even banding on the page. I would certainly say that it is of better quality than the 10×10 10dpi image, even though it came from the same data. The printout is much more uniform.

So I suspect the printer makes up for the additional information when printing at lower resolutions than what the printer prints at.

In article ,
says…

Take a 10×10 pixel image at 10dpi, and do a gradient from black to white from left to right and print. You pretty much get 10 equal bands when you print.

Now, resample this to 300×300 pixels, 300dpi, and you’ll see photoshop soften the edges. When I print this (with an epson c84), I get a much more even banding on the page. I would certainly say that it is of better quality than the 10×10 10dpi image, even though it came from the same data. The printout is much more uniform.

So I suspect the printer makes up for the additional information when printing at lower resolutions than what the printer prints at.

That does not seem meaningful to photo images. Take the converse case, of an actual real photo image containing the normal sharp edges of detail. Even better, a hard-edged detailed graphic. Interpolate it 2x or 3x larger in size. Interpolation is a blurring/blending operation. Now you have no sharp edges, the image is overall rather soft, probably objectionably soft. (yes, this probably does help your gradient).

My opinion is that it is better to simply print what you had first in this case, as being best case (unadulterated pixels). One possible exception is if it were such low resolution that the pixels are objectionably visible, then this softening may be desirable (however you still wont have much, as compared to the proper size image for the purpose).

If this is all you have, best of all is to just print it smaller.

—
Wayne
http://www.scantips.com "A few scanning tips"

In article ,
"Larry Bud" wrote:

That’s not really the question, though. In my example, a 500×500 pixel image at 100 dpi will print at 5"x5". If I resample to 1500×1500, 300 dpi, it will also print at 300 dpi.

Now, in both cases, is the printer throwing down the same amount of ink to paper?

Yes. It is laying down a five-inch by five-inch area of ink.

If so, WHERE is the additional information coming from in the 100dpi example if the printer is a 300dpi printer?

You are confusing two separate, completely unrelated things–the resolution of the image and the resolution of the printer. These things are not directly related to one another at all.

Let us say that you have a very, very low resolution image–it is 5" by 5" but it is just one single pixel; that is, it is 0.2 pixels per inch.

You print it. Your printer prints at 300 dpi. Your printer uses many printer dots to make one pixel. It’s really that simple.

Is the
Photoshop still resampling, or is the printer driver doing something?

Neither. The printer is using many dots to create one single image pixel.

Let’s take this a step further. If I have a 10×10 pixel image, at 10dpi, I will get a 1"x1" image when printing. What is doing the logic to get 10 dots to cover 1" wide?

The printer is saying "Hmm, let me se here. The first pixel I am getting is one inch wide. okay, i will now lay down one inch of ink that is the right color. I print at 300 dpi, so i will now print 300 dots of the same color."

—
Art, photography, shareware, polyamory, literature, kink: all at http://www.xeromag.com/franklin.html

Tacit wrote:

In article ,
"Larry Bud" wrote:

That’s not really the question, though. In my example, a 500×500

pixel

image at 100 dpi will print at 5"x5". If I resample to 1500×1500,

300

dpi, it will also print at 300 dpi.

Now, in both cases, is the printer throwing down the same amount of

ink

to paper?

Yes. It is laying down a five-inch by five-inch area of ink.

If so, WHERE is the additional information coming from in the 100dpi example if the printer is a 300dpi printer?

You are confusing two separate, completely unrelated things–the resolution of the image and the resolution of the printer. These

things

are not directly related to one another at all.

They are separate, but they certainly are related

Let us say that you have a very, very low resolution image–it is 5"

by

5" but it is just one single pixel; that is, it is 0.2 pixels per

inch.

You print it. Your printer prints at 300 dpi. Your printer uses many printer dots to make one pixel. It’s really that simple.

Yes, but the pattern that it uses is important in the quality of the print that you get. See my sample in this thread.

Let’s take this a step further. If I have a 10×10 pixel image, at 10dpi, I will get a 1"x1" image when printing. What is doing the

logic

to get 10 dots to cover 1" wide?

The printer is saying "Hmm, let me se here. The first pixel I am

getting

is one inch wide. okay, i will now lay down one inch of ink that is

the

right color. I print at 300 dpi, so i will now print 300 dots of the same color."

Try the sample that I gave. The results are not the same.