[Na Horuk]

Yeah, MF can deliver stunning results, but it is much more difficult to use. It is big, heavy, the controls are more advanced, there are more things to think about. But the optics and image recording are a step better.
This is why most people start with an entry level DSLR, Yes, its cheaper, but its also made for beginners, with functions and automation that beginners need, without the most complicated features that only some pros appreciate.
But if you are willing to power through the steeper learning curve, you can start anywhere. But it might take longer to see good results if you start with 645Z than if you start with K-50.

[normhead]

My take on it is this; the medium format options out there today, even the "light-weight" Z and S 007, are a fairly demanding type of camera to acquire the best results from, even for an experienced photographer.
My old 5D2 was a fairly simple camera to shoot, and the heavily increased tolerances required by the Z required me to almost relearn how to use a camera properly for best results. Settings which you knew in the back of your head were good enough, now led to poor results. While jumping feet first into a high-end system will leave your with plenty room to grow, it's also like getting into an F1 race car to learn how to drive. OP needs a "fun" camera before a serious one.

I wouldn't even suggest the 5DSR for a start, but the new 5D4 that's been announced. 30MP is still a lot of detail, but it's also going to be a very fast and forgiving camera to use for most subject matter, while still having a large selection of lenses to work with, including some truly superlative zooms and primes.
If you really want to print BIG a better option is stitching frames anyway, because you need 15,840 pixels across to make a 44" (1.2m) print at 360PPI, and even the Z can't manage that without at least three shots (accounting for overlap).

By the time OP gets some good experience under their belt, a new P645 camera may be out that's even better than the Z, and there might also be new lenses to pick from as well. In the mean time, something snappy and easy to get results with is a better option, you'll want such a camera anyway for your backup.

You've completely missed the part where he wants to do large scale prints. Not my experience but I was recently talking to one of the pros who helped me get into the craft sale circuit. He recently went from a Canon 3Dmk3 to a Nikon D810 and I had to listen to him rave on about the added detail for about 5 minutes.

For those of us who start with super hi-res on view cameras, you've got us wondering... do you actually understand that to do really high res images you have to give up run and gun shortcuts? Another thing you miss. As format size increases, the need for high quality glass decreases. Canon's stable of lenses means next to nothing to an MF shooter.

If you really want to print BIG a better option is stitching frames anyway, because you need 15,840 pixels across to make a 44" (1.2m) print at 360PPI, and even the Z can't manage that without at least three shots (accounting for overlap).

I understand that you have to print at 360 DPI on some printers to force the printer to use it's best printing algorithms, but I've seen nothing to indicate that the original file has to be 360 DPI. You can take a file that works out to 150 DPI, enlarge it to 360 DPI, and odds are the images will be pretty much indistinguishable. From my perspective it's unfortunate that this notion that the fact that "fine" print mode on the printer is triggered by a 360 DPI file has led some to believe a 360 DPI image is necessary for the best large scale printing. But it seems to be one of those common beliefs.

There was never more disappointment around my house than the day I did my first 300 DPI (Canon "fine printing " kicks in at 300 DPI) native image on my new Canon Pixma Pro 9000 with the 10 ink cartridges and pigment type inks.) I literally thought WTF? ) When they test cameras and lenses they test the lenses and give you a number. Like 2700 lw.ph. When you get a printer, no one tells you how many paired lines per inch the printer will out put. And to me that's next to criminal. It's easy to test info that should be available. My guess is, in the real world, you can't depend on a printer to give you more than 100 lines per inch, even though the literature will tell you you can get better by using larger files. And the fact that there is no definitive info out there published by any printer company that says different, is just more evidence to suggest I'm right. They tell you what size file you need to input, but in terms of what resolution they will output from that file, no information is available, even from sites like Imaging Resources who always release the same numbers for lens and camera combinations.. It's one of the industry conspiracies no one will talk about, or address.

I've seen no proof one way or the other, but what comparisons I have seen would suggest a 120 DPI file expanded to 360 DPI with your favourite post processing software, will be pretty much indistinguishable from a file printed on a 360 DPI native file.

The fact that a printer needs 360 DPI to trigger it's best print mode, doesn't mean it can effectively print 360 dpi. Create a 360 DPI file of alternating one pixel wide black and white lines and print it, look at the mess you get, and you'll see exactly what I'm talking about.

I tend to think detail on a level of 100 distinct lines per inch ( a real world kind of measurement, different cameras achieve different lw/ph with the same DPI, DPI doesn't tell you what you need to know)) is great detail on a photo. A 645z will give you that at 40 inches by 70 inches with no stitching. I'm not at all certain that you're seeing this right. The evidence I've seen would suggest otherwise.

I have seen test that show for large scale printing, 72 lw/ph is enough. that pushes your 645z file to huge sizes. I mean how big are we talking here? With a 645z producing more than 4000 lw/ph of strong detail you could conceivably push it to more than 60 inches by 90 inches and still show very fine detail down to 1/70 of an inch.

I know some people think they need to be able to count threads on an image of a 300 threads per inch bed sheet for the image to be acceptable, but folks, in the real world, that's preposterous. You need magnification to do that in real life.

[Kolor-Pikker]

You've completely missed the part where he wants to do large scale prints. Not my experience but I was recently talking to one of the pros who helped me get into the craft sale circuit. He recently went from a Canon 1Dmk3 to a Nikon D810 and I had to listen to him rave on about the added detail for about 5 minutes.

I didn't miss anything, 30MP is easily enough for large-scale prints. As you go up in resolution, the percentile bonus to resolution progressively becomes less and less, which is what we actually perceive as detail. In case of your friend, going up to the D810 was not just a 250% increase in resolution but the added image clarity due to the wide dynamic range. The difference between that and an extra 14MP with maybe half a stop extra DR is still there of course, but much less impressive.

For those of us who start with super hi-res on view cameras, you've got us wondering... do you actually understand that to do really high res images you have to give up run and gun shortcuts?

Of course, I'm simply not sure whether or not OP understands that, in which case it's better to err on the side of usability than risking a disappointment. I'm going out a limb here and assuming they don't really know what they're getting into, which is all that can be inferred through text on a forum. Who knows, maybe they'd even prefer the use of a tech camera. The best solution is to try these systems out, I've tested every MF camera out there before settling on the Z.

Another thing you miss. As format size increases, the need for high quality glass decreases. Canon's stable of lenses means next to nothing to an MF shooter.

The MF shooting style is also largely incomparable, where you often stick to smaller apertures to maximize sharpness and DoF. You'd be silly to imply that any of Pentax's legacy lenses, and even the 55mm are anywhere near as sharp at their max aperture as some small format lenses are. Yeah, the lp/mm tolerances are much lower, but the lenses merely take advantage of that to avoid being mediocre. Why do you think some people unironically mount small format lenses like the 24mm TS-E II on tech cams? Or why a lot of people are clamoring for Pentax to release new 645 lenses in general? They're pretty good, but not as good as the format should allow for, which is what makes it feel as much of a side-grade to a nice 35mm camera such as the D810 as an upgrade.

And what if you wanted a lens like the 70-200/2.8 IS II? The best we have is the FA 80-160/4.5, which is 1.5 stops slower, has no stabilization, focuses slower, and is hardly as sharp wide open as the Canon lens would be on the 5DSR. Sure, stopped down on a tripod the difference would be in favor of the Pentax again, but then we're back to the stone age.

I need to be able to count threads on an image of a 300 threads per inch bed sheet for the image to be acceptable, but folks, in the real world, that's preposterous. You need magnification to do that in real life.

I was under the impression that's exactly what he wanted. Don't take my hyperbole the wrong way, I print full-width prints on my Epson from a single 645Z shot (often cropped) all the time, and it holds up even to close inspection.

[normhead]

I didn't miss anything, 30MP is easily enough for large-scale prints. As you go up in resolution, the percentile bonus to resolution progressively becomes less and less, which is what we actually perceive as detail. In case of your friend, going up to the D810 was not just a 250% increase in resolution but the added image clarity due to the wide dynamic range. The difference between that and an extra 14MP with maybe half a stop extra DR is still there of course, but much less impressive.


Of course, I'm simply not sure whether or not OP understands that, in which case it's better to err on the side of usability than risking a disappointment. I'm going out a limb here and assuming they don't really know what they're getting into, which is all that can be inferred through text on a forum. Who knows, maybe they'd even prefer the use of a tech camera. The best solution is to try these systems out, I've tested every MF camera out there before settling on the Z.


The MF shooting style is also largely incomparable, where you often stick to smaller apertures to maximize sharpness and DoF. You'd be silly to imply that any of Pentax's legacy lenses, and even the 55mm are anywhere near as sharp at their max aperture as some small format lenses are. Yeah, the lp/mm tolerances are much lower, but the lenses merely take advantage of that to avoid being mediocre. Why do you think some people unironically mount small format lenses like the 24mm TS-E II on tech cams? Or why a lot of people are clamoring for Pentax to release new 645 lenses in general? They're pretty good, but not as good as the format should allow for, which is what makes it feel as much of a side-grade to a nice 35mm camera such as the D810 as an upgrade.

And what if you wanted a lens like the 70-200/2.8 IS II? The best we have is the FA 80-160/4.5, which is 1.5 stops slower, has no stabilization, focuses slower, and is hardly as sharp wide open as the Canon lens would be on the 5DSR. Sure, stopped down on a tripod the difference would be in favor of the Pentax again, but then we're back to the stone age.


I was under the impression that's exactly what he wanted. Don't take my hyperbole the wrong way, I print full-width prints on my Epson from a single 645Z shot (often cropped) all the time, and it holds up even to close inspection.

Thanks for the great replay. I think we are on the same page here.
How wide is "full width".

[texandrews]

You've completely missed the part where he wants to do large scale prints. Not my experience but I was recently talking to one of the pros who helped me get into the craft sale circuit. He recently went from a Canon 3Dmk3 to a Nikon D810 and I had to listen to him rave on about the added detail for about 5 minutes.

For those of us who start with super hi-res on view cameras, you've got us wondering... do you actually understand that to do really high res images you have to give up run and gun shortcuts? Another thing you miss. As format size increases, the need for high quality glass decreases. Canon's stable of lenses means next to nothing to an MF shooter.



I understand that you have to print at 360 DPI on some printers to force the printer to use it's best printing algorithms, but I've seen nothing to indicate that the original file has to be 360 DPI. You can take a file that works out to 150 DPI, enlarge it to 360 DPI, and odds are the images will be pretty much indistinguishable. From my perspective it's unfortunate that this notion that the fact that "fine" print mode on the printer is triggered by a 360 DPI file has led some to believe a 360 DPI image is necessary for the best large scale printing. But it seems to be one of those common beliefs.

There was never more disappointment around my house than the day I did my first 300 DPI (Canon "fine printing " kicks in at 300 DPI) native image on my new Canon Pixma Pro 9000 with the 10 ink cartridges and pigment type inks.) I literally thought WTF? ) When they test cameras and lenses they test the lenses and give you a number. Like 2700 lw.ph. When you get a printer, no one tells you how many paired lines per inch the printer will out put. And to me that's next to criminal. It's easy to test info that should be available. My guess is, in the real world, you can't depend on a printer to give you more than 100 lines per inch, even though the literature will tell you you can get better by using larger files. And the fact that there is no definitive info out there published by any printer company that says different, is just more evidence to suggest I'm right. They tell you what size file you need to input, but in terms of what resolution they will output from that file, no information is available, even from sites like Imaging Resources who always release the same numbers for lens and camera combinations.. It's one of the industry conspiracies no one will talk about, or address.

I've seen no proof one way or the other, but what comparisons I have seen would suggest a 120 DPI file expanded to 360 DPI with your favourite post processing software, will be pretty much indistinguishable from a file printed on a 360 DPI native file.

The fact that a printer needs 360 DPI to trigger it's best print mode, doesn't mean it can effectively print 360 dpi. Create a 360 DPI file of alternating one pixel wide black and white lines and print it, look at the mess you get, and you'll see exactly what I'm talking about.

I tend to think detail on a level of 100 distinct lines per inch ( a real world kind of measurement, different cameras achieve different lw/ph with the same DPI, DPI doesn't tell you what you need to know)) is great detail on a photo. A 645z will give you that at 40 inches by 70 inches with no stitching. I'm not at all certain that you're seeing this right. The evidence I've seen would suggest otherwise.

I have seen test that show for large scale printing, 72 lw/ph is enough. that pushes your 645z file to huge sizes. I mean how big are we talking here? With a 645z producing more than 4000 lw/ph of strong detail you could conceivably push it to more than 60 inches by 90 inches and still show very fine detail down to 1/70 of an inch.

I know some people think they need to be able to count threads on an image of a 300 threads per inch bed sheet for the image to be acceptable, but folks, in the real world, that's preposterous. You need magnification to do that in real life.

While you and I have had serious differences with each other on the forums in the past, this second portion of your post is something that should be read by all who print, with respect to the combination of the resolution of files, the camera's capabilities, and the realities of the printer's capabilities (and the "fine print" thereof...!), and real world viewing of a print. I have commented not so much to add needless affirmation but to make sure this is "bookmarked" and accessible to me in the near future.

[normhead]

While you and I have had serious differences with each other on the forums in the past, this second portion of your post is something that should be read by all who print, with respect to the combination of the resolution of files, the camera's capabilities, and the realities of the printer's capabilities (and the "fine print" thereof...!), and real world viewing of a print. I have commented not so much to add needless affirmation but to make sure this is "bookmarked" and accessible to me in the near future.

I've always said that the folks I have disagreements with usually know as much or more than I do. For some reason our communication styles just rub each other the wrong way. Glad I finally said something that didn't do that.

[TonyW]

I will not comment on which is the better 50 MP camera to purchase MF or FF 35mm. They are quite different to handle and perhaps one more suited than another to specific photo situations. So I guess the OP will really need to make a decision based on many things not least of which is what is the main use - assuming pocket depth is not an issue.

The matter of printing is the thing that interested me and I hope the information provided is of help and does not muddy the waters.

Taking the Pentax 645Z as an example the native resolution is 8256 x 6192 pixels, therefore printing at this native resolution will yield different sizes depending on printer requirements.

Canon/HP drivers report 300/600 ppi to the OS

300 ppi will yield a print 27.5" x 20.6"
or using the best resolution setting
600 ppi will yield a print 13.8" x 10.3"

Epson print drivers report 360/720 ppi to the OS therefore:

360 ppi will yield a print of 22.9" x 17.2"
720 ppi will yield a print 11.47" x 8.6"

It is generally considered that for best IQ upscaling should be performed in PS or LR (or other printing application) rather than leave the print driver to perform the task

It should also be a given that the quality of the original capture plays a huge role in IQ and the ability to resolve fine detail (should this be required)

Any of the above will stand very close scrutiny. i.e. photographers viewing distance which is limited by one of two factors
a. The length of the photographers nose
b. How close can you get to the print wall before the belly prevents closer inspection

You should also really take into account expected print viewing distance vs the required number of pixels to present a continuous tone impression. A quick way is to use the formula 3438 ÷ Viewing Distance.
So for instance if you anticipate viewing at no closer than 24" you would use 3438/24 = 143 ppi required. Upscale therefore to 300 ppi to Canon print driver.

Normal calculation for a 'correct' print viewing distance is 1.5x - 2.0x the diagonal of the print.

Photoshop and Lightroom are both very capable of upscaling images to at least 200 -300% without any issues and if 'reasonable ' print viewing distances anticipated then IQ will be excellent.
So in the case of 300% upscale we end up with an image size of 24768 x 18576 pixels giving a print size of 82.56" x 61.92" sending the printer 300 ppi image to Canon or HP printers.

While it is often suggested that line pairs per inch equate to print ppi /2 i.e. 300 ppi = 150 lppi I do not think that this is a reliable measure. Printers are often quoted in terms of DPI but this cannot really mean Dots per Inch as printer DPI is really a measure of ink volume (expressed as picoliters : A picoliter is a trillionth of a litre) i.e. Droplets per inch and does not bear comparison with dots per inch due to ink spread on different papers. Further many inkjet have variable sized droplets making LPI calculations difficult

[normhead]

Tony, if Hadn't sold three copies of a 10 MP Opito 80W images,
the file size is 3500x 2300 and it's printed at 23x33 for a canvas wrap... so 100 pixels per inch....

Has sold 3 times at $350 a pop, earning us more than $700 after printing costs.

Until some one explains tome how that is possible.... posts like the above just leave me shaking my head.

Your chart shows 300 DPI yielding a 27.5 x 20.6 print.

My actual print uses 100 DPI to make a 23x33 print that is good enough to sell at craft shows. So what are you actually getting for printing the 645z image?

Here's my point

Canon/HP drivers report 300/600 ppi to the OS

300 ppi will yield a print 27.5" x 20.6"
or using the best resolution setting
600 ppi will yield a print 13.8" x 10.3"

What do you get for using 300 dpi or 600 ppl.

What is the difference to the output?

Printer companies want to talk about input. They don't say anything about output.
Personally I'm tired of the nonsense.
Printer companies should be forced to advertise the output in lw/ph graphed against input MP.
That's really the only information we care about.
There should be a standard print file like the MTF charts for printer so their output performance can be measured. This slight of hand by the printer companies to obscure the nature of their output curves is simply false advertising,

When you print, you don't care what the preferred input sizes should be, you care what the preferred input size gets you in terms of output. That's all that's relevant. We do it for sensors, we do it for lenses, it should be done for printers.

[TonyW]

Hi normhead,
I was merely trying to add some objective and verifiable facts about printing to the thread, not suggesting that anyone's view is particularly wrong. So if I may answer your comments to the best of my knowledge and based on accepted wisdom related to human visual acuity:

Tony, if Hadn't sold three copies of a 10 MP Opito 80W images,
the file size is 3500x 2300 and it's printed at 23x33 for a canvas wrap... so 100 pixels per inch....

Has sold 3 times at $350 a pop, earning us more than $700 after printing costs.

I have no doubt that you have had excellent quality at 100 ppi but there are a few things to bear in mind in this particular instance:

a. Most canvas printing does not benefit from anything over 100 ppi due to generally not being able to resolve fine detail due to the rough surface - at least compared to smooth art papers

b. 100 ppi is not really enough resolution to view close without seeing pixels for those with 20/20 vision (potential for dot pattern to be revealed or at least a coarseness to be perceived)

c. It is reasonably easy to see why 100 ppi may be acceptable even for the highest quality photo paper (i.e. that which is capable of resolving the highest detail).

Taking a 20" x 30" print on smooth art paper. The diagonal of the paper is 36.06" suggesting a viewing distance of 36x1.5 = 4.5 ft approx. The actual resolution required is only around 64 ppi to present an even toned image.
Should you wish to print smaller and view the same image even closer then the ppi requirements will be larger e.g. 4 x 6 print viewing distance 11" = 313 ppi.

d. While you may think that you are printing at a particular ppi it is likely that you are not! If you just send your image to the printer driver, lets assume to reach your required print size this happens to be 100 ppi the printer driver must do some interpolation using its own algorithms to send the printer data it can work with correctly. These algorithms are not necessarily seen to be optimum compared to PS or LR. If you use a specialist printing package such as the excellent Qimage you will find that maximum print quality setting is 600 ppi and high quality is 300.

e. It has to be said that image content/quality plays a role here. If you have a subject with loads of fine detail, architectural or birds feathers whatever then taking notice of recommendations and test printing will yield benefits

Until some one explains tome how that is possible.... posts like the above just leave me shaking my head.

Your chart shows 300 DPI yielding a 27.5 x 20.6 print.

My actual print uses 100 DPI to make a 23x33 print that is good enough to sell at craft shows. So what are you actually getting for printing the 645z image?

The camera/sensor size is irrelevant if you print at 100 ppi image data - the net result is going to be close for the same image and equivalent size whatever the format. I assume that DPI was a typo referring to image output

Here's my point

What do you get for using 300 dpi or 600 ppl.

What is the difference to the output?

You are talking about Canon/HP printers here and again ppi not dpi? The difference between 300 and 600 ppi will be observable on certain images i.e. those that contain fine detail and particularly lines running at angle other than truly horizontal or vertical or circular objects and of course potentially show better resolution with the correct data. It may only be slight gain going from 300 to 600 but nevertheless a gain.

Printer companies want to talk about input. They don't say anything about output.
Personally I'm tired of the nonsense.

On the contrary they do talk about both input 300/600 ppi and output is mentioned usually in multiples of print head nozzles e.g. 1200/2400/4800/9600 dpi.

Printer companies should be forced to advertise the output in lw/ph graphed against input MP.
That's really the only information we care about.
There should be a standard print file like the MTF charts for printer so their output performance can be measured. This slight of hand by the printer companies to obscure the nature of their output curves is simply false advertising,

How can they honestly advertise line pairs or other objective measures when they have so many variables to consider e.g. Variable Droplet Size from 3 - 10 Picoliter, many different paper types that will cause ink spread and therefore size to be very different depending on surface type and absorption rates?

When you print, you don't care what the preferred input sizes should be, you care what the preferred input size gets you in terms of output. That's all that's relevant. We do it for sensors, we do it for lenses, it should be done for printers.

On the contrary when we print we should be aware of the input requirements for our printers to enable the sending of optimum data to ensure that we are getting the best IQ out of our file data

[normhead]

How can they honestly advertise line pairs or other objective measures when they have so many variables to consider e.g. Variable Droplet Size from 3 - 10 Picoliter, many different paper types that will cause ink spread and therefore size to be very different depending on surface type and absorption rates?

By taking a simple test chart in a file, and interpreting the output of the file on glossy photpaper.....

I don't want the long answer, I want the short answer.

You are talking about Canon/HP printers here and again ppi not dpi? The difference between 300 and 600 ppi will be observable on certain images

And exactly who has observed these images? Are these practical observations or theoretical observations based on known printer characteristics?

I have seen or heard of very little data, the one set of independent data I have seen suggest people couldn't tell the difference between a 72 DPI file, a 100 DPI file and a 150 DPI file, that was the only independent test I've ever seen. Yet you say people can see the difference between 300 and 600 ppi. In such situations, I always go with the empiracle test. My answer is we have very primitive research that suggests the input file size of has very little to do with the quality of the print. As for zaggies on horizontal lines etc..that hasn't been true since dot matrix printers and very small files. A while ago when testing various lenses and TCs (in which I demonstrated that FL whether or not a TC was used determined resolution, one poster was seriously confused by the images I posted. That was because the software I used to upsize my wider angle files to make them the same size, clearly used some vector algorithms to smooth out the zaggies etc, The upsized file looked cleaner than the native file. But non vector random variations in the print chart were reproduced in the longer FL shots, Random details that were the result of the $900 printer not being able to accurately reproduce a test chart without bleeding ink into places it should not have been.

Contrary to the zaggies theory, smaller files that have been upscaled by good software can look better than higher resolution files. Not as much detail but better looking detail You could almost make the argument that upscaling will make your prints cleaner looking, by suppressing non vector clutter and by magnifying captured detail to the point where it actually visible to the naked eye. . So again, my own empirical testing is contrary to what you are proposing. All the printer companies are telling us is what size the file needs to be to make maximum use of the printer. that is completely a separate topic from what DPI the original file needs to be, before upscaling, for the best results.

To my mind, I don't care about the smokescreen printer company nonsense. All I want to know is "what kind of output can I expect from an arbitrary print file." The nonsense about droplet size, 300 or 600 dpi... it's for the most part irrelevant. That is not output, that's input. And I don't care about print nozzles and their sizes, I want to see test charts. I want to know how accurately the printer reproduces the file it's sent. Just as I want to know how accurately my camera reproduces the image I see.

You solve variables buy adopting common standards. Agree on the type of paper to be used. Or let the printer company pick the paper they think will give their printer the best opportunity to perform. No one cares, but, every time I have devised a test to confirm the conventional theories posted above, the results have not matched conventional wisdom.

So, I'm loathe to recite printer company marketing information and or excuses. Every single time I have tested them in my real world photography, they have not proved to be true. I'm looking right now a an image that I printed a 300 DPI to trigger Canon's fine print option. The same image printed twice the size at half the resolution looks better, This is way more complicated than printer company information is going to lead you to believe.

I have looked for even one example online of an two images that will clearly show the advantage of using file with more than 100 DPI for the original file before upscaling. As I've yet to see proof that any image exists that in a blind test will be noticeably better printed at 600 DPI rather than 300 DPI, or 100 DPI upscaled to 600 or 300 DPI.

The simple fact is we don't know what we're getting. We submit our the highest res work and hope for the best. to me, that's not acceptable. I won't be buying another printer, until I can understand the difference in the output of the various models, and I can have some kind of empiracle test to confirm my belief that what I am doing is actually working as expected. When i was burning $220 every few months for cartridges and doing a lot of my own printer, simple fact, most of the time I could not see a difference in quality between my HP 4 colour printer/scanner/fax machine and my Canon Pixma Pro 9500 10 colour printer. Owning my own printer was one of the more disappointing aspects of my career.

[noelpolar]

Whilst managing the stamp program of Australia Post (which included many millions of dollars of print work per year) I eventually got around to looking at all the pre press we did to ensure we were not overspending on quality that wasn't relevant to our customer base.

Well did that almost do my head in! So many opinions and so much practicle experience..... but trying to bring it all back to some kind of fact base..... from both a technology and human perception perspective almost made it to hard to make any sort of agreed measured progress. A lot of the time one just ended up "trialing" some perceived "reduction"..... and whilst the world was forcast to fall in on us it never really even got noticed by the vast majority of the market.

[photoptimist]

For text printing, 300 dpi is distinguishable from 600 dpi. For images, just look at the difference between normal computer monitors (typically about 100 dpi) and the "retina" models (higher than 200 dpi). But the biggest issue is that you don't want the visual resolution of the final result to be limited by the printer. Just as one needs to oversample a fine-grain scene to catch all the detail, one needs to over-sample in printing.

Printer specs like droplet size are like ISO specs on film -- you know smaller is better but there's no gaurantee that all printers of the same small droplet size or films of the same low ISO have the same fine grain resolution.

Finally, at some level the whole printer performance issue is out of the control of the printer company -- it depends on the paper, temperature, humidity, and maintenance of the print head. And there's a lot the image creator can do such as over-shaprening to compensate for printer issues such as dot gain.

[TonyW]

By taking a simple test chart in a file, and interpreting the output of the file on glossy photpaper.....

So you are really looking for the printer manufacturers to quote the very best case scenario, using the best quality paper and images of the highest standard for resolution?
This would be fine only IF all manufacturers used the same standard testing criteria.

I don't want the long answer, I want the short answer.

Really, and yet your answer in this thread is rather long?
I am sorry to say that the short answer is that on this subject you are mistaken or misinformed.

To the quote that a difference between 300 and 600 ppi will be observed you reply

And exactly who has observed these images? Are these practical observations or theoretical observations based on known printer characteristics?

Apart from my own printing, such people as Jeff Schewe, the late Bruce Fraser, Mike Chaney (Qimage), Bart Van Der Wolf, Jim Kasson, Ernst Dinkla, the late Uwe Steinmueller to name just a few proffessional users and their practical observations
As it seems that you have seen very little data relating to the phenomenon a very short list below:
Schewe
The Right Resolution - Digital Photo Pro
Chaney
The Qimage Print Quality Challenge
This one had to think about because really it comes across as someone who has actually rediscovered the wheel. Nevertheless this does not mean the findings incorrect
Introducing the Ultraprint ? Ming Thein | Photographer

I have seen or heard of very little data, the one set of independent data I have seen suggest people couldn't tell the difference between a 72 DPI file, a 100 DPI file and a 150 DPI file, that was the only independent test I've ever seen.

72, 100 and 150 PPI testing hardly surprising although questionable as no mention of image content or just as important print viewing distance. 10 ppi can look good and appear to show detail viewed at the correct distance i.e. a poster viewed from across the road!
Would you care to share a link to your set of independent data?

Yet you say people can see the difference between 300 and 600 ppi. In such situations, I always go with the empiracle test. My answer is we have very primitive research that suggests the input file size of has very little to do with the quality of the print. As for zaggies on horizontal lines etc..that hasn't been true since dot matrix printers and very small files.

One individuals empirical testing lead many to believe the earth was flat. Another writing in the early 1800's wrote "What can be more palpably absurd than the prospect held out of locomotives traveling twice as fast as stagecoaches". Empirical testing by an individual needs proving and usually by ones peers. Otherwise there is the all too human reaction to having a theory and rejecting all evidence that does not support that theory.

. . So again, my own empirical testing is contrary to what you are proposing. All the printer companies are telling us is what size the file needs to be to make maximum use of the printer. that is completely a separate topic from what DPI the original file needs to be, before upscaling, for the best results.

See above for empirical testing.
As to the statement that DPI is a seperate topic to file size seems somewhat confused. First it is very important in this type of discussion to get our terminology correct to enable the reader to understand.
DPI has absolutely nothing to do with file size, PS, LR and any worthwhile imaging application do not understand DPI they work with image data as PPI or even PPC (pixels per inch or pixels per centimetre). DPI as stated earlier is not a measure of size (printing related) - it cannot be as it is a measure of volume. And further this volume will spread or be absorbed differently depending on the print settings and the paper type this liquid is sprayed onto.

To my mind, I don't care about the smokescreen printer company nonsense. All I want to know is "what kind of output can I expect from an arbitrary print file."

You have mentioned previously printer company smokescreens / nonsense or similar. Would you care to provide examples that can be corroborated?

The nonsense about droplet size, 300 or 600 dpi... it's for the most part irrelevant. That is not output, that's input

???Droplet size is output based garnered from the input that is your image data

You solve variables buy adopting common standards.

The common standards have already been touched upon in my post here

... No one cares, but, every time I have devised a test to confirm the conventional theories posted above, the results have not matched conventional wisdom

With respect the testing methods that deny the the standards suggest flaws in either the methods or the data tested.

So, I'm loathe to recite printer company marketing information and or excuses. Every single time I have tested them in my real world photography, they have not proved to be true. I'm looking right now a an image that I printed a 300 DPI to trigger Canon's fine print option. The same image printed twice the size at half the resolution looks better, This is way more complicated than printer company information is going to lead you to believe.

Canon's fine print option is not 300 PPI but 600 PPI are you talking about printer laying down ink (dpi) or your file size in PPI?

If you are disappointed with your printers output then there can only be one of two conclusions.
a. The data presented to the printer is not of optimum quality in the first instance or the print setting are incorrect

b. There is a mechanical fault in the printer or printing pipeline

FWIW. I am somewhat loathe to mention this and I will not be drawn into further discussion on this particular subject but.. My own practical experience is based on over 40 years working as a professional (Photographer and Consultant) in the image and information field including setting up analogue and digital printing services and educating photographers in best practice.

[normhead]

So you are really looking for the printer manufacturers to quote the very best case scenario, using the best quality paper and images of the highest standard for resolution?
This would be fine only IF all manufacturers used the same standard testing criteria.

You mean like Imatest for printers... sorry, I'm not seeing the problem.

72, 100 and 150 PPI testing hardly surprising although questionable as no mention of image content or just as important print viewing distance. 10 ppi can look good and appear to show detail viewed at the correct distance i.e. a poster viewed from across the road!
Would you care to share a link to your set of independent data?

Apart from my own printing, such people as Jeff Schewe, the late Bruce Fraser, Mike Chaney (Qimage), Bart Van Der Wolf, Jim Kasson, Ernst Dinkla, the late Uwe Steinmueller to name just a few proffessional users and their practical observations
As it seems that you have seen very little data relating to the phenomenon a very short list below:
Schewe
The Right Resolution - Digital Photo Pro
Chaney
The Qimage Print Quality Challenge
This one had to think about because really it comes across as someone who has actually rediscovered the wheel. Nevertheless this does not mean the findings incorrect
Introducing the Ultraprint ? Ming Thein | Photographer

I did quite a bit of research on this.... and I found not one blind test, where it was established that higher ppi prints produced images viewers found to be better.

72, 100 and 150 PPI testing hardly surprising although questionable as no mention of image content or just as important print viewing distance. 10 ppi can look good and appear to show detail viewed at the correct distance i.e. a poster viewed from across the road!
Would you care to share a link to your set of independent data?

A straw poll done in a photography class, 20 students, 8x10 images students could hold in their hands. Any data is better than no data.

Long story short, regardless of the printer company misdirection, my advice to people , as long as your original file size means you are going to have 100 pixels per inch, give it a try, you might like it.

Since I've been thinking for a while I have my own conjecture about why many photographer think what they think, I noticed a while ago comparing a friends D810 and my K-5 images at UWA waterfall thoughts. The photographers looked at the slight distortion of the 8mm Sigma and said "this is the K-5 file, meaning they preferred the D810 file. They were completely incapable of evaluating the artistic value of the prints because they were caught up in the technical. Ordinary people didn't care, they selected the image they thought was most pleasing to look at. They didn't care what format it was taken with and they didn't look for the fact that one image had twice the resolution of the other. . They looked at the image. and some preferred the APS-c image. This is the kind of thing I think technical people get caught up in. You tell me to look at the difference between my display and my grandson's retina display. Well, I have , and while some images look better on the retina, some are fine on either, and in that case, the retina display costs a lot more money to see the same thing.

I believe this comment

FWIW. I am somewhat loathe to mention this and I will not be drawn into further discussion on this particular subject but.. My own practical experience is based on over 40 years working as a professional (Photographer and Consultant) in the image and information field including setting up analogue and digital printing services and educating photographers in best practice.

Probably fits into that category.

If it's any consolation, my current evaluation is that I'm going to use a printing service for further printing. The ability to do top notch prints is just as much a skill as photography, and just because you're good at one doesn't mean you're good at the other. And I'm starting to think the same about my final PP work. The final PP needs to be done by someone who understands the characteristics of the printer, and what the intended result is.

I see a lot of use for print professionals. But, I'd hate to see folks held back by the ridiculous input files size suggestions published in various places around the web. To me saying you have to have native 300 dpi or 600ppi to have the most enjoyable to look at print is next to criminal. It's industry hype encouraged by those encouraging the MP race.

An image can be a great image, without those kinds of file sizes.

Canon's fine print option is not 300 PPI but 600 PPI are you talking about printer laying down ink (dpi) or your file size in PPI?

Actually I looked far and wide for this info and even wrote to Canon, so yo might be right.. I never found a definitive answer, but, some of the Services I used asked for my files going to their Canon printers to be 300 ppi. I guess they didn't know about the 600ppi thing either.

I'm quite happy to accept that I can't put the time and energy and expense into finding out exactly how my Canon printer worked. But, in high school, if finished in aptitude above 95% percentile in english comprehension and math, so, if I can't figure it out, how many people can? TO me this is middle ages guild type stuff. You don't know a thing if you don't know the secret handshake.

My conclusion after all this, is if you really want top notch printing, if you think it's critical to your print, and in most prints it's not, but the flip side of that is, in some prints it is, go to a pro, pay him to do some PP to clean up your work for the printer as well as work out the best way to print it.

But don't sell your lo-res files short. Modern upscaling software will get you amazing prints with little to work with, at much lower input resolutions than printer companies would suggest. And remember your best image printed from a low res file, is always better than a crap image printed at 600 ppi. Print standards have little to do with the enjoyment people get out of your images in most cases.

And don't be a criminal, don't tell folks that the most enjoyable prints will be 600 ppi without upscaling. That's just nonsense.

I'm sure a print professional will look at those 600 dpi prints and feel a certain degree of satisfaction, but for the average Joe to put his favourite picture from his vacation on his wall, for the most part, a lot less is just as good. Better, because he gets a nice big print that's great to look at, and didn't have to take a 645z and stitch four images together to get it.

[Canada_Rockies]

An amateur user's input:

My wife and I have only one colour printer in our house. It is a little Canon Selphy C910. It prints using die sublimation on 4x6 paper using a one use film. It prints very nice 4x6 prints, once you discover the correct settings in LR. My wife and I both use LR for our editing; I also use DxO and Franzis plugins.

If I want 4x5 inch prints up to 8x10, I send full res jpegs to Cranbrook Photo. If I want bigger than that, I go to print shops. My 20x30 metallic mentioned earlier in the thread was printed by Posterjack here in Canada from a 10 mpx file from my K10D. I sold it to a very happy customer, who wanted to be able to show prospective home buyers the neighbourhoods in an almost helicopter viewpoint. Every house is clear, but you cannot read the house numbers. I have an 11x14 print from Velvia 50 that is nothing short of gorgeous. The slide development, print specifications and the printing were done by a pro shop for a quite large sum of money - around $300 I remember - but the results are nothing short of spectacular. That's where I'm with Norm - let the pros do their professional thing. I'll take the image. If I want a huge print from my K-3, I will send the raw file, exported as a full resolution TIFF to a pro shop and let them have at it and my wallet.

I seem to be siding with Norm Head on this issue, and I am. I got my new-to-me K-3 (used) because my K10D was failing, not to get more resolution in the sensor. For scenery, I still prefer the colours I got without effort out of the K10's CCD sensor at ISO 100. I have had to set up special treatment to approach that in LR form the CMOS sensor in the K-3. On the other side, I can crop more from the K-3 without losing detail. That's a nice bonus.

For me, printing is someone else's forte. My forte is taking prictues. I hated the darkroom, and I find my Lightroom a pain, too, but I don't get migraine headaches from the fumes.