Reading mode: Light Dark

dpi/ppi resolution query

Go to last post
R

reck242358

Active member
Messages
55
Reaction score
0
Location
Santa Fe, New Mexico, US
Vincent Oliver at photo-i.co says in his review of the Epson 925:

"Print quality in both modes is excellent... At 5760 dpi, the 925 is able to generate maximum quality from 240dpi files. A 10 x 8 requires a file of around 13MB. To achieve similar results on other printers would require a 1200 x 1200dpi file which would be over 300MB."

This comment stretches my understanding of image resolution and suggests I don't quite understand how to make my first choice of high-end photo printer. I would appreciate anyone's clarifying comments.

He is making the point that the printer performs a sort of 'interpolation' by supplying so many dpi for every ppi of the image that the actual image dpi can be fairly small. This is intriguing and I have never come across the point elsewhere. But for me it asks more questions than it answers. To wit:

--What constitutes "maximum quality" (how is he defining this)?

--What "other printers" are we talking about? Is it clear to anyone from his math what printers he is thinking of? 4800dpi? 2400dpi? 1440dpi?

--I don't get the '1200dpi x 1200dpi = 300MB' equation. Does anybody?

Or does anyone have some general comment on the relationship of printer dpi to the recommended size in ppi of the digital image?

Thanks
 
reck said:
Vincent Oliver at photo-i.co says in his review of the Epson 925:

"Print quality in both modes is excellent... At 5760 dpi, the 925
is able to generate maximum quality from 240dpi files. A 10 x 8
requires a file of around 13MB. To achieve similar results on other
printers would require a 1200 x 1200dpi file which would be over
300MB."

This comment stretches my understanding of image resolution and
suggests I don't quite understand how to make my first choice of
high-end photo printer. I would appreciate anyone's clarifying
comments.
Click to expand...
Stretch????---To me, the quote BREAKS all credibility.

There is no inkjet printer on earth that would know what to do with a 600PPI file, much less 1200. Note the PPI, not DPI

File sizes are in Pixels per inch (PPI). Printers are often spec'ed in DPI, which does NOT relate to file size, and ---for typical high-end photo printers--can be effectively ignored.

Printers have differences in resolving power that are inherent to the printer. The paramater of interest is droplet size and the dpi rating. If a high-resolution printer gives a certain quality at a 240PPI file size, a lesser printer CANNOT match that performance by simply sending a larger file.

Without any context--based only on the quote--I give vincent oliver a credibility score of 0
 
mherring02 said:
reck said:
Vincent Oliver at photo-i.co says in his review of the Epson 925:

"Print quality in both modes is excellent... At 5760 dpi, the 925
is able to generate maximum quality from 240dpi files. A 10 x 8
requires a file of around 13MB. To achieve similar results on other
printers would require a 1200 x 1200dpi file which would be over
300MB."

This comment stretches my understanding of image resolution and
suggests I don't quite understand how to make my first choice of
high-end photo printer. I would appreciate anyone's clarifying
comments.
Click to expand...
Stretch????---To me, the quote BREAKS all credibility.

There is no inkjet printer on earth that would know what to do with
a 600PPI file, much less 1200. Note the PPI, not DPI

File sizes are in Pixels per inch (PPI). Printers are often
spec'ed in DPI, which does NOT relate to file size, and ---for
typical high-end photo printers--can be effectively ignored.

Printers have differences in resolving power that are inherent to
the printer. The paramater of interest is droplet size and the dpi
rating. If a high-resolution printer gives a certain quality at a
240PPI file size, a lesser printer CANNOT match that performance by
simply sending a larger file.

Without any context--based only on the quote--I give vincent oliver
a credibility score of 0
Click to expand...
Okay.

Still, it seems there is an interesting point in the quote trying to get out. For instance, what IS the maximum effective file size in ppi for a particular printer, beyond which is superfluous or impossible for the printer to handle? And does the extremely high resolution of the Epson make a smaller ppi file as effective as a larger one on a lesser printer? Obviously, like software interpolation, a printer's 'high-dpi interpolation' would only be good up to a point. But what would be the rules in thinking about this?

Perhaps questions like this only start to come out as printers begin to accomplish the goal of continuous-tone images with 'grain' (dot-size) approaching and even exceeding the higher film-chemistry standards.
 
reck242358 said:
Okay.

Still, it seems there is an interesting point in the quote trying
to get out. For instance, what IS the maximum effective file size
in ppi for a particular printer, beyond which is superfluous or
impossible for the printer to handle? And does the extremely high
resolution of the Epson make a smaller ppi file as effective as a
larger one on a lesser printer? Obviously, like software
interpolation, a printer's 'high-dpi interpolation' would only be
good up to a point. But what would be the rules in thinking about
this?
Click to expand...
When I got my 1280, I did quite a bit of testing--using a photo disk test target (I got mine from David Chien (silverace.com))

I saw continuous improvement up to 350-400 PPI file size. between 350 and 400, the difference was not something that would justify dealing with such large files.

More importantly, if I test with actual pictures, I seldom benefit from going higher than 300---many prints on matte paper are made at 240.

The newer printers with drop sizes around 2pL will give different results, but less so for real-world pictures.
 
mherring02 said:
reck said:
Vincent Oliver at photo-i.co says in his review of the Epson 925:

"Print quality in both modes is excellent... At 5760 dpi, the 925
is able to generate maximum quality from 240dpi files. A 10 x 8
requires a file of around 13MB. To achieve similar results on other
printers would require a 1200 x 1200dpi file which would be over
300MB."

This comment stretches my understanding of image resolution and
suggests I don't quite understand how to make my first choice of
high-end photo printer. I would appreciate anyone's clarifying
comments.
Click to expand...
Stretch????---To me, the quote BREAKS all credibility.

There is no inkjet printer on earth that would know what to do with
a 600PPI file, much less 1200. Note the PPI, not DPI

File sizes are in Pixels per inch (PPI). Printers are often
spec'ed in DPI, which does NOT relate to file size, and ---for
typical high-end photo printers--can be effectively ignored.

Printers have differences in resolving power that are inherent to
the printer. The paramater of interest is droplet size and the dpi
rating. If a high-resolution printer gives a certain quality at a
240PPI file size, a lesser printer CANNOT match that performance by
simply sending a larger file.

Without any context--based only on the quote--I give vincent oliver
a credibility score of 0
Click to expand...
Files sizes are in bytes. Stored in that file is a tag to tell a printer how many dpi to print at. A file 3000x2000 pixel digital image can be tagged with 72dpi oe 600 dpi and the file is still the same size and looks identical on screen. Only when sent to the printer does this dpi tag have meaning. Printers need high ratings eg 4800 dpi to place several coloured dots in a region to simulate the colour of a pixel. They don't print at 4800dpi. 300-360 dpi is a maximal print resolution you should aim for. This is set in image resize. Turn off resample and just enter dpi you want to print at upto 300-360dpi. Arguments rage on about > 360dpi for prints, but any improvements are miniscule and you are actually making it harder for the printer; they have to fit several dots into a smaller area now and if the inks droplets are too large it can't improve resolution.
 
Vincent probably make a mistake in word but his statement is true.

All are in the printer driver.

An epson printer driver can send an CMYK at 5760x720 dpi to an Epson printer with an RGB at 240 ppi input.

But with other manufacturer, the driver expect a 1200 ppi RGB input in order to archieve the maximun printer resolution (4800x1200 dpi for example). With HP, if you want to print a 300ppi at 4800x1200dpi, the drivers create a 1200 ppi file before send it to the printer.

This is because Epson printer have less intelligence than other printer but this is a long story (read the gimp-print developper guide if you want more information).

Don't confuse between Dpi and ppi : a printer use dot (dpi) and a computer display use pixel (ppi). Now a printer dot can have only 8 distint color (for 3/4 inks printer) altought an image pixel can have over 16Millions (24bits image).

The printer driver job is to place the dot on paper in order that create the same RGB color (if possible of course) so a pixel is recreate by few dot.

In fact, you never need to use the best printer resolution but in Vincent review of the i950 we can see that the Epson 925 with less thechnical advantage (4pl dot with 5760x720 dpi) can reveal with the i950 (2pl dot with 4800x2400 dpi) in image quality.

And the maximum resolution of an image for printing is 600ppi. More ppi can't offer better result.

The good range is 240-360ppi. 300ppi is the common use.
reck said:
Vincent Oliver at photo-i.co says in his review of the Epson 925:

"Print quality in both modes is excellent... At 5760 dpi, the 925
is able to generate maximum quality from 240dpi files. A 10 x 8
requires a file of around 13MB. To achieve similar results on other
printers would require a 1200 x 1200dpi file which would be over
300MB."

This comment stretches my understanding of image resolution and
suggests I don't quite understand how to make my first choice of
high-end photo printer. I would appreciate anyone's clarifying
comments.

He is making the point that the printer performs a sort of
'interpolation' by supplying so many dpi for every ppi of the image
that the actual image dpi can be fairly small. This is intriguing
and I have never come across the point elsewhere. But for me it
asks more questions than it answers. To wit:

--What constitutes "maximum quality" (how is he defining this)?

--What "other printers" are we talking about? Is it clear to anyone
from his math what printers he is thinking of? 4800dpi? 2400dpi?
1440dpi?

--I don't get the '1200dpi x 1200dpi = 300MB' equation. Does anybody?

Or does anyone have some general comment on the relationship of
printer dpi to the recommended size in ppi of the digital image?

Thanks
Click to expand...
 
I'm not sure what Vincent is really getting at here. He usually is making a good point, though it eludes me here.

First, all the printers that I have been using lately (that being the top end Canon letter and 13x19 photo printers and the Epson R800 and 2200) are capable of resolving more detail on the photo paper than can be seen with the naked eye. They do this with the manufacturers supplied driver.

It is my understanding that printers have a hardware resolution that their driver interpolates the file resolution to and then dithers the individual ppi into many dots (in dpi) to actually create the raster that is printed. It is the last part that really determines the need for very high-resolution dpi printers. The hardware resolution of the Epson printers is 720. Of course the human eye cannot even remotely “see” this type of detail in a continuous tone “real world” environment. I used to be extremely dubious of the advantages of using file resolutions any higher than the mid to high 200’s ppi and I would print everything at 300 ppi just to keep the numbers even. That allowed me to interpret the image to a fixed number and apply a final USM just before pint time.

In a series of dialogs and testing my old Epson 900 and my new R800 it became clear I was living in the proverbial fool’s paradise. It turns out that with small delicate detail the algorithms in Photoshop do a better job of rescaling the image to 360 ppi (an even fraction of the Epson native res) than the Epson driver does in scaling odd-fractional sizes to its native res. What this means is that with finely detailed images, a file will print better at 360 ppi than 300 ppi and presumably even at 400 ppi. What I really found intriguing is that very high contrast repeating patterns of certain frequencies can create moiré. This appears as a rainbow like sub-pattern in the light spaces between the darker shapes of the primary repeating pattern. This does happen and I was easily able to create it in prints using a photographic test chart as the printed file. Printing at 720 ppi eliminated this completely. Skeptics will rightfully SCREAM that this is hardly the norm in real world printing (what ever that is) but situations like this do crop up. Stairs in the distance, clap board siding, shingles, even combed hair when viewed from the right distance are a few examples. To keep file sizes down I print detailed photos at 360 ppi and reprint if moiré manifests itself. (I’ve only found it once since my “enlightenment”).

Bruce
mherring02 said:
reck said:
Vincent Oliver at photo-i.co says in his review of the Epson 925:

"Print quality in both modes is excellent... At 5760 dpi, the 925
is able to generate maximum quality from 240dpi files. A 10 x 8
requires a file of around 13MB. To achieve similar results on other
printers would require a 1200 x 1200dpi file which would be over
300MB."

This comment stretches my understanding of image resolution and
suggests I don't quite understand how to make my first choice of
high-end photo printer. I would appreciate anyone's clarifying
comments.
Click to expand...
Stretch????---To me, the quote BREAKS all credibility.

There is no inkjet printer on earth that would know what to do with
a 600PPI file, much less 1200. Note the PPI, not DPI

yada yada yada.....
Click to expand...
 
Patrick Martin said:
is just use Qimage and let it worry about all of that stuff for
you. It's far better at it than any of us will ever be.
Click to expand...
Unfortunately for me (in this one case!) I use a Mac. I have looked at the Qimage site carefully and then looked for software that seemed to be doing the same thing. I concluded that Qimage may be unique in approaching the problem from the printer end instead of the image-software end. In other words I got the impression Qimage may be dealing with just this 'printer-native' issue we've been discussing. Does anyone else agree?

The closest approximations I could find for the Mac are Fred Miranda's Stair-Interpolation Action and Genuine Fractal software. Both these are bringing new types of interpolation to the problem and both think they do a better job than PS bicubic, but the thrust of this thread so far is that the crux of the matter is: What the heck is the 'native' or 'hardware' resolution of a particular printer, so that unnecessary interpolation can be avoided? Or at least intelligently invoked (to avoid artifacts)?

I have come across this problem in a different setting. A commercial imagesetter like an Agfa lays down a stochastic (random-dot) image at 2400 or 4800 dpi and is executed by a RIP. The operator of such a machine has recommended to me that I work in a file size of 300ppi, then resize to 600ppi before converting to a Diffusion Dither Bitmap in PS at 1200dpi, so it can be output on the Agfa at 2400dpi. In other words, all even multiples, just as this thread seems to be supporting.

The basic ideas here are not so impossibly technical but it seems the printer manufacturers rather intentionally muddy the waters.
 
Sorry if I confused people with the Agfa imagesetter requirements; I neglected to mention that those were for B&W images, which is why bitmaps come into it. Still, I think the issues may be the same.
 
You must log in or register to reply here.

Keyboard shortcuts

f
Forum
m
My threads
Mobile site
About
Editorial content
Cameras & Lenses
Community
All content, design, and layout are Copyright © 1998–2025 Digital Photography Review All Rights Reserved.
Reproduction in whole or part in any form or medium without specific written permission is prohibited.
When you use DPReview links to buy products, the site may earn a commission.
©GPS Media - Guides, Products, Services.
Back
Top