• Please use real names.

    Greetings to all who have registered to OPF and those guests taking a look around. Please use real names. Registrations with fictitious names will not be processed. REAL NAMES ONLY will be processed

    Firstname Lastname

    Register

    We are a courteous and supportive community. No need to hide behind an alia. If you have a genuine need for privacy/secrecy then let me know!
  • Welcome to the new site. Here's a thread about the update where you can post your feedback, ask questions or spot those nasty bugs!

What is your workflow for upressing, sharpening and printing?

Asher Kelman

OPF Owner/Editor-in-Chief
Hi, Asher,



Yes, these are all subtle-sounding matters than can be of great significance. It is a very complex matter, overall.



Doug,


Mike has informed me that Qimage has to hold everything in memory when saving to "print to file". That is limiting for massive files. By contrast, when QImage outputs directly to a printer, it does so in chunks and so there's no RAM anemia issue!

64 Bit Windows allows 32 BIT programs to use up to a 2GB in the Windows milieux, allowing one to work with "massive 1GB images and sometimes a bit more.

Anyway, he considers that 600 dpi is the max real detail that the Canon printer can deliver, except that at 1200 dpi it does tis with more dots. For Epson, it is likely 720 dpi printing that is the max useful setting too.

Of course, we'll have to confirm these limitations. Still, if confirmed, it may absolve us off the burden of massive 1200 dpi files for printing large prints optimally on the Canon printer! :)


Asher
 

Doug Kerr

Well-known member
Hi, Asher,

Mike has informed me that Qimage has to hold everything in memory when saving to "print to file". That is limiting for massive files. By contrast, when QImage outputs directly to a printer, it does so in chunks and so there's no RAM anemia issue!

64 Bit Windows allows 32 BIT programs to use up to a 2GB in the Windows milieux, allowing one to work with "massive 1GB images and sometimes a bit more.Anyway, he considers that 600 dpi is the max real detail that the Canon printer can deliver, except that at 1200 dpi it does tis with more dots. For Epson, it is likely 720 dpi printing that is the max useful setting too.

What does the red passage mean? Will that Canon printer do 1200 px/inch? Does it do that with fewer dots per pixel?

What would be a typical Canon printer to which this pertains? I'd like to install its driver and see what Qimage Ultimate offers to do with it.

Of course, we'll have to confirm these limitations. Still, if confirmed, it may absolve us off the burden of massive 1200 dpi files for printing large prints optimally on the Canon printer!

Indeed.

Well, the only color printer I have is my Epson SP R1900, and as near as Qimage Ultimate can tell, its highest native interface resolution is 720 px/in.

Thanks for the scoop.

Best regards,

Doug
 

Asher Kelman

OPF Owner/Editor-in-Chief
Hi, Asher,



What does the red passage mean? Will that Canon printer do 1200 px/inch? Does it do that with fewer dots per pixel?

What would be a typical Canon printer to which this pertains? I'd like to install its driver and see what Qimage Ultimate offers to do with it.



Indeed.

Well, the only color printer I have is my Epson SP R1900, and as near as Qimage Ultimate can tell, its highest native interface resolution is 720 px/in.

Thanks for the scoop.

Best regards,

Doug

One can set the printer and QImage to 1200 dpi, however, Mike asserts that apart from printing more dots to get what one gets in print, there will be no more detail.

Asher
 

Doug Kerr

Well-known member
Hi, Asher,

One can set the printer and QImage to 1200 dpi, however...

On what model of printer for example? I would like to "install it" and see Qimage willing to deal with it at 1200 px/in.

... Mike asserts that apart from printing more dots to get what one gets in print, there will be no more detail.

I doubt that it uses more dots that way. Probably fewer dots per pixel, more pixels, same number of dots. But what do I know.

Best regards,

Doug
 

Asher Kelman

OPF Owner/Editor-in-Chief
Mike says more dots! ?


For testing at 1200 dpi, try the Canon imagePROGRAF iPF8400 44in Printer - 6565B002AA. That's the one The print atelier uses and the one I'd consider buying.

Thanks,

Asher
 

Doug Kerr

Well-known member
Hi, Asher,

Mike says more dots!

If we consider a Canon iPF8400, according to the Canon site, its dot resolution is 2400 dots/in × 1200 dots/in.

If we ran this at 1200 px/in, then each pixel would be rendered with 2 dots (I guess that is per color).

That doesn't sound too nifty.

I think there is still some misunderstanding here.

Best regards,

Doug
 

Sam Hames

New member
To act as a devil's advocate: Does the whole question really matter?

All of this technical discussion is ignoring something much more important - the role of the human observer. I'd hazard a guess that the idiosyncracies and limitations of the human visual system dominate over any of the fine technical details under discussion here. And furthermore, that most people won't notice the difference between 'fancy' sharpening and moderate unsharp masking. *

Stated another way - given a finite number of hours to spend on photography, I'd rather be doing anything but sharpening :)


* Think of a controlled trial - get a random selection of people and ask them to choose between the same photo sharpened through two different methods (or even no sharpening!). I'd hypothesise that you wouldn't find any difference in preference between the two methods of sharpening.

Even if there were a difference, you might then consider whether the effect size would be larger than say, altering global contrast, or colour balance or something similar.
 

Asher Kelman

OPF Owner/Editor-in-Chief
To act as a devil's advocate: Does the whole question really matter?

All of this technical discussion is ignoring something much more important - the role of the human observer. I'd hazard a guess that the idiosyncracies and limitations of the human visual system dominate over any of the fine technical details under discussion here.



Sam,

Thanks for adding your thoughts. My needs happen to be to obtain the finest appearing print to my eyes that withstands scrutiny at 10-15" as at normal casual viewing distances. This is because I have observed how buyers in the Art exhibitions I frequent might examine and view prints they buy.

Also, some prints look better printed large, say at 72 to 80 inches high or 10 ft wide. For advertising, even 100 dpi might be fine. For artwork however, the standards are much higher. So it does matter to me how the file is upressed and printed. Given the best practice, one can make significantly improved prints that are obviously well crafted. That's all I seek.

However, when one is not in direct control of the printer, at least the file delivered should be the best possible for that printer.

I am eager to learn what others do in sending their precious work to be printed 1-3meters in length at a giclee print house.

Asher
 

Doug Kerr

Well-known member
Hi, Asher,

What if you did 300 dpi?

We must first assume that the printer could adopt a 300 px/in mode.

Then (based on my interpretation of the specifications) the printer engine could potentially use up to 32 dots (per color) to make each pixel.

Whether it actually would or not I don't know.

If the printer has a highest native resolution of 600 px/inch, and we send its driver an image whose pixel dimensions, at the print size "ordered", amounted to 300 px/in, I assume that:

• The driver wouild uprez the image, the best way it could, to pixel dimensions that would, at the print size ordered, constitute a resolution of 600 px/in.

• The printer system could then use up to 8 dots (per color) to make each pixel.

But I may misunderstand all this in some important way.

Maybe that spec actually should have said "the maximum printer resolution is 2400 px/in × 1200 px/in."

What it says exactly is:

Print Resolution (Up to) 2,400 x 1,200 dpi (Max)​

In any case, the Canon iPF8400 printer has 2560 nozzles per color! And it is said that the nozzle pitch is "1200 dpi".

But we are used to people tossing about "dpi" and "ppi" (better stated as "px/in") interchangeably.

"So why are you so fussy about whether we call the animal a 'buffalo' or a 'bison' ?"

" You'll find out when the one you ordered from 'Exotics R Us' arrives."​

I need to know way more about this than I do.

We "communicate" with people who know many of the answers, but what we get does not answer our questions.

It often makes one wonder whether they understand what they "know".

Best regards,

Doug
 

Doug Kerr

Well-known member
As a matter of interest, here is an excerpt from the Specifications part of the Epson site page for the Epsom Stylus Photo R1900 printer (the one I have):

Nozzle Configuration: 180 nozzles x 8 [It has 7 inks plus a sort of "lacquer"]

Maximum Print Resolution: 5760 x 1440 optimized dpi​

What does "optimized dpi" mean? My guess is that it means the effective dot pitch given micro-movement of the head, perhaps in both directions.

In any case, the native resolution this machine reports to Qimage Ultimate is 720 px/in, unless we select the "Draft" quality mode and toilet paper, in which case it reports 180 px/in.

Best regards,

Doug
 

Asher Kelman

OPF Owner/Editor-in-Chief
Doug,

Yes, it would be helpful if the print MFR would explain how they specify things and what difference each change in settings means in practical terms of observation and recognition of printable detail.

For sure, I think it must already be written up somewhere, but to this point, I have not found anything sufficiently comprehensive and lucid!

Manwhile, I discovered this hand Optimal Workflow Guide. It's a good place to start before sending pictures of to a printer and risking a lot of money on miscommunication! The needs there are for media photographs not fine art and the print resolutions recommended are in the order of 360 to 480 dpi for Epson printers and 400 dpi for continuous tone laser-exposed photographic paper.

Asher
 

Doug Kerr

Well-known member
Hi, Asher,

Doug,

Yes, it would be helpful if the print MFR would explain how they specify things and what difference each change in settings means in practical terms of observation and recognition of printable detail.

For sure, I think it must already be written up somewhere, but to this point, I have not found anything sufficiently comprehensive and lucid!

Indeed.

Manwhile, I discovered this hand Optimal Workflow Guide. It's a good place to start before sending pictures of to a printer and risking a lot of money on miscommunication! The needs there are for media photographs not fine art and the print resolutions recommended are in the order of 360 to 480 dpi for Epson printers and 400 dpi for continuous tone laser-exposed photographic paper.

I'll look at it. Thanks.

In the meantime, I can't imagine why a resolution of 360 to 480 px/in would be optimal for my Epson R1900 with a native resolution (except in Draft mode with the medium set to toilet paper) of 720 px/in.

But maybe that article will explain it.

Best regards,

Doug
 

Doug Kerr

Well-known member
Hi, Asher,

Manwhile, I discovered this hand Optimal Workflow Guide. It's a good place to start before sending pictures of to a printer and risking a lot of money on miscommunication! The needs there are for media photographs not fine art and the print resolutions recommended are in the order of 360 to 480 dpi for Epson printers and 400 dpi for continuous tone laser-exposed photographic paper.

An interesting article.

I don't seem to find any of the thoughts highlighted in red in your comment above in it. I don't even find the word "Epson" in the piece. Where am I going wrong?

Thanks.

Best regards,

Doug
 

Asher Kelman

OPF Owner/Editor-in-Chief
Hi, Asher,



An interesting article.

I don't seem to find any of the thoughts highlighted in red in your comment above in it. I don't even find the word "Epson" in the piece. Where am I going wrong?

Thanks.

Best regards,

Doug

He used an Epson Printer and 360 is one of their settings.
 

Doug Kerr

Well-known member
Hi, Asher,

He used an Epson Printer and 360 is one of their settings.
Ah! My mind reader is working poorly these days.

"Their"? Does that mean "Epson's". Or does that mean, "of Epson printers"?

I use an Epson printer and 360 px/in is not one of its "settings" (at least that I can make show up).

For almost all configuration options of interest, it reports its resolution to Qimage Ultimate as 720 px/in.

But with it in place, Qimage Ultimate can be set to treat it as if its resolution were 720, 480, 360, or 180 px/in.

Best regards,

Doug
 

Asher Kelman

OPF Owner/Editor-in-Chief
Hi, Asher,


Ah! My mind reader is working poorly these days.

"Their"? Does that mean "Epson's". Or does that mean, "of Epson printers"?

I use an Epson printer and 360 px/in is not one of its "settings" (at least that I can make show up).

For almost all configuration options of interest, it reports its resolution to Qimage Ultimate as 720 px/in.

But with it in place, Qimage Ultimate can be set to treat it as if its resolution were 720, 480, 360, or 180 px/in.

Best regards,

Doug

Interesting. I earned from Epson salesman to figure 360 dpi as their recommended "standard" to prepare for the printer or simple multiples of that.

A most useful and practical source for understanding viewing distance and print resolution is by Keith Cooper writing from "Northern Light" in the U.K.


"The -maximum- resolving power of the eye is about 60 lines per degree. That means that in bright light you should be able to resolve 60 black/white line pairs over an angular distance of a degree.

The resolution of the Eye - Discussion in more detail, concerning how close to sit to the TV (ordinary and HDTV) - original site lost - link goes to archive.org copy
MTF and resolution - Detailed information from Norman Koren about resolution and lens performance
Note - If you are not familiar with thinking of things in degrees, remember that the diameter of the full moon is about half a degree

The table below (originally from the DigitalWorkflow Yahoo group): gives an idea of the maximum resolution at various distances. Remember that this is for very high contrast at optimal lighting - not your print of a photo hanging on the wall. These values could easily be halved for 'average' or poor viewing conditions.

Viewing Distance (inches)
Resolution ppi​
61
145​
10
687​
24
286​
36
191​
60
115​
120
57​
600
11​


The formula is ppi = 1/((distance x 0.000291) / 2)

The sensitivity of our visual systems to differing angular resolutions also varies with colour. Our eyes are more sensitive to larger scale variations in colour than in brightness. For black and white images, the eye performs a degree of spatial filtering which peaks at around 7 cycles per degree" Source.

In practice, Keith Cooper reports that he uses a rule of thumb of 1.5 to twice the print diagonal as the viewing distance and that folk who look with a magnifying glass at 6" don't buy his prints!

From all of the above, for large prints, I think that 300 dpi is likely to be excellent and 600 dpi will allow even a most fastidious buyer to be happy if the print is otherwise compelling!

Given that I have a totally color managed system until the point of printing, and the price of ink and media runs to about $25 per print versus $135 to $250, purchasing one's own printer is really worth considering. With my last order, I could have had a new 44" Canon printer and several rolls of Baryta paper, LOL!

Asher
 

Doug Kerr

Well-known member
Hi, Asher,
Interesting. I learned from Epson salesman to figure 360 dpi as their recommended "standard" to prepare for the printer or simple multiples of that.

For example, 720 px/in, the native resolution of my Epson R1900.

"The -maximum- resolving power of the eye is about 60 lines per degree. That means that in bright light you should be able to resolve 60 black/white line pairs over an angular distance of a degree."

Indeed. Some authorities cite 100 cy/degree ("line pairs/degree") under ideal conditions.

As a matter of interest, "20/20" vision corresponds (crudely) to the ability to resolve 30 cycles (30 line pairs) per degree. (The Snellen chart is based on this premise.)

From all of the above, for large prints, I think that 300 dpi is likely to be excellent and 600 dpi will allow even a most fastidious buyer to be happy if the print is otherwise compelling!

Oh, you will need much more than that except in the case of line drawings. We need many dots for each pixel for continuous tone and color work.

Perhaps you mean to speak in terms of px/in (often abbreviated "ppi").

Given that I have a totally color managed system until the point of printing, and the price of ink and media runs to about $25 per print versus $135 to $250, purchasing one's own printer is really worth considering. With my last order, I could have had a new 44" Canon printer and several rolls of Baryta paper, LOL!

Well, it is always lovely to be self-contained!

Thanks.

Best regards,

Doug
 

Asher Kelman

OPF Owner/Editor-in-Chief
Oh, you will need much more than that except in the case of line drawings. We need many dots for each pixel for continuous tone and color work.

Perhaps you mean to speak in terms of px/in (often abbreviated "ppi").

Actually, doug, I was indeed talking about the "setting on the printer software" as there are no actual dots produced, to the best of my knowledge, that correspond numerically to dpi. Rather, the intensity of the effort, error's in positioning allowed, constitute a print "quality" of the aforementioned setting of "dpi". Based on that, the printer's thousands of jets, will be positioned and squirt to an accuracy sufficient to correspond to that resolution. But I cannot swear to that. That actual regions of the "dot" consists of numbers colors placed down in a particular order to achiever the desired cold and without color balance failure.

Asher
 

Doug Kerr

Well-known member
Hi, Asher,

Actually, doug, I was indeed talking about the "setting on the printer software"

Do you mean settings on the printer driver or on the "printing software" (e.g., Qimage Ultimate)?

I'm not used to settings either place being denominated in "dpi". Were you just using that as a metaphor? How would I, per the Epson salesman's recommendation, arrange my setup for "360 dpi"?

I hate to keep talking about units, but if we say "gallons" when we mean" pounds" it does not help comprehension.

. . . as there are no actual dots produced, to the best of my knowledge, that correspond numerically to dpi.

Yes, indeed. The dots are invariably on a finer grid than the pixels.

Rather, the intensity of the effort, error's in positioning allowed, constitute a print "quality" of the aforementioned setting of "dpi".

Where is there a setting that works in "dpi"?

Based on that, the printer's thousands of jets, will be positioned and squirt to an accuracy sufficient to correspond to that resolution.

That would make sense for a resolution set in px/in (ppi).

But I cannot swear to that. That actual regions of the "dot" consists of numbers colors placed down in a particular order to achiever the desired cold and without color balance failure.

Sure.

Thanks.

Best regards,

Doug
 

Asher Kelman

OPF Owner/Editor-in-Chief
So here's the dot pattern of an epson print at 600 dpi. Source.




Matte_5x_CloseUp.jpg


Print on Epson Luster Paper. The colored dots are to white balance the paper.
The black area is a black slanted edge used in Imatest charts.





Now read Norman Koren of Imatest. He thinks that there's a lot of hype about so-called printer resolution and informs us that the measured capability of print output with an inkjet does not rise above ~150! Read this.So the dpi claims of 600, 1200 and 2400 are hard to understand!


Asher
 

Doug Kerr

Well-known member
Hi, Asher,

So here's the dot pattern of an epson print at 600 dpi. Source.

Done on an HP DesignJet Z3100, actually.

Print on Epson Luster Paper. The colored dots are to white balance the paper.
Very interesting!

Hard to know just where the dots are. But one can sort of guess.
The black area is a black slanted edge used in Imatest charts.

Ah, yes. Nice.

Now read Norman Koren of Imatest. He thinks that there's a lot of hype about so-called printer resolution and informs us that the measured capability of print output with an inkjet does not rise above ~150! Read this.So the dpi claims of 600, 1200 and 2400 are hard to understand!

I don't find anything in that citation about printers. (My whole day has gone like that!).

Are you speaking of his speaking about dpi claims or ppi (px/in) claims?

No, I will not give up on this, not until my bison arrives.​

Thanks for this good scoop.

Best regards,

Doug
 

Doug Kerr

Well-known member
I haven't found the Normal Korem report you cite, but on his big tutorial site I find this:

Based on the original target printed at 4x magnification, the Epson Photo Stylus 1270 has a resolution limit of about 120 lp/in at the 50% contrast level and 190 lp/in at the 10% level.​

That would correspond to pixel resolutions of 240 px/in and 380 px/in for that machine.

Best regards,

Doug
 

Asher Kelman

OPF Owner/Editor-in-Chief
I haven't found the Normal Korem report you cite, but on his big tutorial site I find this:

Based on the original target printed at 4x magnification, the Epson Photo Stylus 1270 has a resolution limit of about 120 lp/in at the 50% contrast level and 190 lp/in at the 10% level.​

That would correspond to pixel resolutions of 240 px/in and 380 px/in for that machine.

Exactly! So what we are dealing with is a hell of a scam by the marketing departments of Epson, Canon and HP. My estimation is that 1200-200 dpi on the Canon or Epson is equivalent to lest than 300 dpi of actual data. We have to test everything ourselves. I need to read a MFR manual to see if they are more candid than the sales brochures!

I think the key thing is to separate the real world of "pixels per inch" of the image file from the term "dpi' 300, 600, 1200 and 2400 which, until anyone proves otherwise be just considered as basis, good, hopefully better, hopefully best.

The setting of 600 with the Canon printer might be the best setting for printing a 300 pixel per inch file. There is no escaping the need to test a print sample of one's image at 300, 600 and 1200 dpi using 300 and 600 pixel per inch image files - or else one needs to actual determine the MTF, actual the CTF of a test target printed on that printer

Asher
 
Hi folks,

Sorry I'm late to the party here. Lot's of things to address so I won't comment on each and every post sofar, but I'll try to condense things into the most relevant issues one needs to consider. Also note that quality improvements may be small (but they are there nevertheless) when pushing the printer to its limits.

First of all. Current inkjet printers work with drivers that use a few fixed (due to hardware) 'native' resolutions. Canon and HP printers typically offer 300 or 600 PPI resolutions, and Epson printers use 360 or 720 PPI.

The Epsons need to have the 'finest detail' option switched on in the driver dialog, otherwise all input will be resampled to 360 PPI, where as the Canon drivers usually are more clear in what they are going to do.

There may be (on some printer models) some driver options suggesting e.g. 1200 or 1440 but those are typically DPI settings, IOW adjustments of the dot or droplet dithering algorithms, used to dither and weave the droplet patterns in order to create intermediate/mixed colors from only a few ink colors and a print medium's background reflection.

Any output resolution (PPI = number of pixels divided by the output size in inches) that is not an exact match to those 300/360 PPI or 600/720 PPI, will be resampled by the printer driver. On Mac computers resampling routines from the OS are used, on Windows systems the driver resamples itself. However, the resampling algorithms used are built for speed, usually bi-linear resampling is used (and maybe/unlikely bicubic at best), not for quality.

That's why it's better to resample before sending the data to the printer, which allows a) to use better algorithms, and b) to use output sharpening after resampling which potentially makes a huge difference in perceived quality. Especially when using the highest resolution settings, i.e. 600/720 PPI, we have lots of pixels which can be pushed really far without creating visible artfacts at normal viewing distances. Human visual acuity is on average close to the 300/360 PPI settings when detail is viewed at reading distance and at good lighting levels. However, Vernier resolution offers much higher acuity clues which human vision will appreciate. Therefore, the highest resolution settings will make a difference if the output after resampling is properly sharpened.

There are also resampling applications that actually add resolution beyond what the original data has to offer. This typically results in edge detail that is narrower/sharper than the resampled data would allow with regular upsampling. "PhotoZoom Pro" and "Perfect resize" are two of such stand-alone applications that also manifest themselves as (automation) plugins when Photoshop is used as a host program.

I've done a lot of research and analysis of these printer peculiarities, and discussed matters on another website for those who want more background detail on the underlying principles. Here, and here, and here, and here, and here are some links.

The final conclusions are as I laid out above; PPI settings of 600/720 PPI do produce better quality output, if the image has good detail, if the resampling quality is better than what the printer driver does natively, and if one uses good quality output sharpening after resampling the source images to exactly those 600 or 720 PPI in output.

As for software to do the resampling and sharpening, I've found the following options do offer the best workflow/image quality:
1. Lightroom offers reasonably good resampling quality, but is limited (to a few fixed choices) in it's control of what sharpening is used after resampling.
2. Qimage Ultimate offers a superior workflow for printing, and uses very good (proprietary) resampling algorithms. It also offers a very effective automatic 'Smart' output sharpening method which uses another proprietary halo free sharpening method called Deep Focus sharpening (DFS).

3. For those who seek even higher output quality, there are dedicated resampling programs that add resolution (as can be demonstrated by inspecting the resampled data in the frequency domain, after conversion to Fourier space representation). PhotoZoom Pro by Benvista, and Perfect Resize by OnOne, are the two more capable ones. One does have to watch out and not exaggerate the amount of resolution that is added, otherwise there will be what I call a visual disconnect between edge detail and lower contrast features.

4. As an overall recommendation to improve image detail, regardless of which output pipeline one is going to use, but also as a general tool for output sharpening, I cannot recommend Topaz Labs Detail enough. It is one of my 'must use' plugins (usable in virtually all Photoshop plugin aware applications). This plugin does marvels to any image that could benefit from good detail. It can also help to tackle some of the perceptual challenges that stem from viewing image detail at different viewing distances. Images that have been properly processed with this tool, will look good at any distance.

Cheers,
Bart
 

Doug Kerr

Well-known member
Hi, Bart,

So good to hear from you on this.

The Epsons need to have the 'finest detail' option switched on in the driver dialog, otherwise all input will be resampled to 360 PPI, where as the Canon drivers usually are more clear in what they are going to do.

In the driver control panel for the Epson Stylus Photo R1900 (the printer I use here), there is no setting labeled "finest detail".

There are several "modes" : Draft, Text, Text & Image, Photo, and Best Photo. Is it perhaps that we have "finest detail ON" only in the Best Photo mode?

Or might that come from unchecking "High Speed"?

I note that the R1900 driver seemingly reports to Qimage Ultimate a "native" resultion of 720 px/in for all setups other than Draft mode with toilet paper selected ("High Speed" checked in all cases).

Thanks so much.

Best regards,

Doug
 
In the driver control panel for the Epson Stylus Photo R1900 (the printer I use here), there is no setting labeled "finest detail".

There are several "modes" : Draft, Text, Text & Image, Photo, and Best Photo. Is it perhaps that we have "finest detail ON" only in the Best Photo mode?

Hi Doug,

I presume it is the same but under a simplified name. The "Finest detail" is an option in the high end printers, but there it comes with a confusing suggestion to use it for graphical detail, while it also benefits regular continuous tone images.

Or might that come from unchecking "High Speed"?

High speed is often associated with either single, or bi-directional printing. Bi-directional requires extremely good print-head positioning which is hard due to tolerances in mechanics. Therefore uni-directonal printing occasionally does a better job.

I note that the R1900 driver seemingly reports to Qimage Ultimate a "native" resultion of 720 px/in for all setups other than Draft mode with toilet paper selected ("High Speed" checked in all cases).

Qimage is clever enough to ask the printer driver what amount of detail it expects to receive. The result of the interrogation is determined by what the various printer driver options are set to, like choice of paper, and quality settings. Qimage remembers all of the last used printer driver settings (and color profile settings for specific papers), or allows to save specific settings for different kinds of jobs.

Cheers,
Bart
 

Doug Kerr

Well-known member
Hi, Bart,

Hi Doug,

I presume it is the same but under a simplified name. The "Finest detail" is an option in the high end printers, but there it comes with a confusing suggestion to use it for graphical detail, while it also benefits regular continuous tone images.

Very likely.

High speed is often associated with either single, or bi-directional printing. Bi-directional requires extremely good print-head positioning which is hard due to tolerances in mechanics. Therefore uni-directonal printing occasionally does a better job.

Yes, I suspect that is what that is for. (I knew that once, but . . .)

Qimage is clever enough to ask the printer driver what amount of detail it expects to receive. The result of the interrogation is determined by what the various printer driver options are set to, like choice of paper, and quality settings. Qimage remembers all of the last used printer driver settings (and color profile settings for specific papers), or allows to save specific settings for different kinds of jobs.

Indeed.

Thanks.

Best regards,

Doug
 

Asher Kelman

OPF Owner/Editor-in-Chief
Hi folks,

Sorry I'm late to the party here. Lot's of things to address so I won't comment on each and every post sofar, but I'll try to condense things into the most relevant issues one needs to consider. Also note that quality improvements may be small (but they are there nevertheless) when pushing the printer to its limits.


Bart,

At last!!! We needed you all along! But tou arrived at just the right time. I was forced me to learn on my own just how much lack of clarity and complete understanding permeates our communities! Even reputable sources, for example, Michael Reichman, a really meticulous fine art printer, (I collect some of his work), only gets it partly right! He does confirm that 1440 dpi is in fact to do with spaying droplets not another way of expressing pixels one to one with "dots".

"Inkjets
Most photographers do their printing these days with a desktop inkjet printer and the Epson Photo printers are the most popular so I'll use them by way of example. These printers, such as the models 870/1270/2000P are (somewhat misleadingly) listed as 1440 dpi printers. This means that they are capable of laying down that many dots per inch. But, to create a colour image they need to use 6 different inks, so any particular pixel reproduced on a print will be composed of some dithered composite of coloured dots using some or all of these inks. That's why you need more dots from your printer than you have pixels in your image.

If you divide 1440 by 6 you end up with 240. This is the true minimum resolution needed to get a high quality photo-realistic prints from a 1440 dpi Epson printer. Many user, myself included, believe that a 360 ppi output file can produce a somewhat better print."
[/quote]


First of all. Current inkjet printers work with drivers that use a few fixed (due to hardware) 'native' resolutions. Canon and HP printers typically offer 300 or 600 PPI resolutions, and Epson printers use 360 or 720 PPI.


However, Michael Reichman goes on to use just 240 dpi, below and not matching the Epson printers own standard settings. This asks the Mac or Windows OS to come in and make an on the fly resampling, simplified for speed.


"If my original scan is big enough to allow this I'll do so but I don't bother ressing up a file to more than 240 ppi when making large prints." Source.


Unfortunately, this is not isolated. I discovered that a very prestigious Giclée print house, sent my files to a Canon Prograf printer at 200, 183 and 283 dpi, none matching the Canon set of desired file pixel/inch needs. Also they files were printed at 200, 183 and 283 "dpi" respectively!

Furthermore, Jeff Schewe's work indicates that 720 "dpi", (600 "dpi") for Canon Prograf printers, does demonstrate finer detail. Still, he does not reveal whether using a file size in DPI greater than the base levels or 360 and 600, (for Epson and Canon respectively) might do even better. So Bart, your advice helps to confirm my suspicions that we need to resize well, sharpen and send to the printer in Pixels/inch that correspond to the printers native set of "dpi" resolutions, even though the actual resolution or MTF's will, of course be native to that printers capability and not based on the so called print resolution" options of the printer.


There may be (on some printer models) some driver options suggesting e.g. 1200 or 1440 but those are typically DPI settings, IOW adjustments of the dot or droplet dithering algorithms, used to dither and weave the droplet patterns in order to create intermediate/mixed colors from only a few ink colors and a print medium's background reflection.

A great description as it deals with the purpose of the fine drops and irt relates to creating an impression rather than exactly reproducing 1:1 the resolution of the input file in pixels per inch.

Any output resolution (PPI = number of pixels divided by the output size in inches) that is not an exact match to those 300/360 PPI or 600/720 PPI, will be resampled by the printer driver. On Mac computers resampling routines from the OS are used, on Windows systems the driver resamples itself. However, the resampling algorithms used are built for speed, usually bi-linear resampling is used (and maybe/unlikely bicubic at best), not for quality.

Heres' a question: when is the capability of the printer saturated with the resolution of the input file in pixels/inch? Can a printer continue to represent more and more data, or is it fixed, as Norman Koren suggests to about 4-5 cycles per mm or ~ 25.4x5 cycles/inch or ~ 254 marks/inch. That would suggest that it might be pointless to send to a printer more than the lowest natural complete "set" for that printer, of 360 pixels/inch, (Epson) and 300 pixels/inch, (Canon)

Resampling & Output Sharpening:

This next section is both comprehensive and especially helpful.

That's why it's better to resample before sending the data to the printer, which allows a) to use better algorithms, and b) to use output sharpening after resampling which potentially makes a huge difference in perceived quality. Especially when using the highest resolution settings, i.e. 600/720 PPI, we have lots of pixels which can be pushed really far without creating visible artfacts at normal viewing distances. Human visual acuity is on average close to the 300/360 PPI settings when detail is viewed at reading distance and at good lighting levels. However, Vernier resolution offers much higher acuity clues which human vision will appreciate. Therefore, the highest resolution settings will make a difference if the output after resampling is properly sharpened.


There are also resampling applications that actually add resolution beyond what the original data has to offer. This typically results in edge detail that is narrower/sharper than the resampled data would allow with regular upsampling. "PhotoZoom Pro" and "Perfect resize" are two of such stand-alone applications that also manifest themselves as (automation) plugins when Photoshop is used as a host program.

I've done a lot of research and analysis of these printer peculiarities, and discussed matters on another website for those who want more background detail on the underlying principles. Here, and here, and here, and here, and here are some links.

The final conclusions are as I laid out above; PPI settings of 600/720 PPI do produce better quality output, if the image has good detail, if the resampling quality is better than what the printer driver does natively, and if one uses good quality output sharpening after resampling the source images to exactly those 600 or 720 PPI in output.

As for software to do the resampling and sharpening, I've found the following options do offer the best workflow/image quality:

1. Lightroom offers reasonably good resampling quality, but is limited (to a few fixed choices) in it's control of what sharpening is used after resampling.

2. Qimage Ultimate offers a superior workflow for printing, and uses very good (proprietary) resampling algorithms. It also offers a very effective automatic 'Smart' output sharpening method which uses another proprietary halo free sharpening method called Deep Focus sharpening (DFS).

3. For those who seek even higher output quality, there are dedicated resampling programs that add resolution (as can be demonstrated by inspecting the resampled data in the frequency domain, after conversion to Fourier space representation). PhotoZoom Pro by Benvista, and Perfect Resize by OnOne, are the two more capable ones. One does have to watch out and not exaggerate the amount of resolution that is added, otherwise there will be what I call a visual disconnect between edge detail and lower contrast features.

4. As an overall recommendation to improve image detail, regardless of which output pipeline one is going to use, but also as a genaral tool for output sharpening, I cannot recommend Topaz Labs Detail enough. It is one of my 'must use' plugins (usable in virtually all Photoshop plugin aware applications). This plugin does marvels to any image that could benefit from good detail. It can also help to tackle some of the perceptual challenges that stem from viewing image detail at different viewing distances. Images that have been properly processed with this tool, will look good at any distance.

Bart,

A superb dissertation and one place to look up everything! Thanks.

Some limitations:

1. QImage won't output to file an image of 600 pixels per inch 30"x40" in size, although it can send it in chinks to a waiting printer.

2. At high file sizes, (30,000 pixels width in CS3, for example), Topaz is not available and grayed out in Photoshop. In Ohotoshop CC 2014 the same happens with 26"x38" files at 1200 dpi, but I will look for the lower limit.

Asher
 
Last edited:

Doug Kerr

Well-known member
Hi, Asher,


You quote somebody as follows:

"But, to create a colour image they need to use 6 different inks. . .If you divide 1440 by 6 you end up with 240. This is the true minimum resolution needed to get a high quality photo-realistic prints from a 1440 dpi Epson printer. Many user, myself included, believe that a 360 ppi output file can produce a somewhat better print."

I'm afraid that is misleading.

The reason we need more than one dot per pixel is not (just) because we may use six different kinds of ink, although that of course contributes to the actual dot count..

It is also because the printer cannot control the density of the ink in the dots (of any given color) at all, and can only control the size of the dots to a limited degree, if at all. Thus we must use (at least in part) a scheme in which for each pixel, for each color) we pepper the pixel's real estate with a pattern of dots and no dots, more dots for a "denser" application of that color of ink, fewer dots for a "less dense" application of that ink.

I think the dot density described in the manufacturers' admittedly-cloudy specs is for each color.

Best regards,

Doug
 
Top