[rednax]

Been trying to find out how the printing is effected by number of megapixels and format. Is it worth it going up to full frame ? The real difference doesn´t seem so big when you read about it.
So, what should you look for ?

[MPrince]

How big do you plan on printing? I have had images shot with a K-7 and DA 18-135 printed as large as 40"x60" that look fantastic.

[amoringello]

Of course the bigger you print, the further back someone is likely to be when viewing the image, so the less resolution you need. (i.e. billboards have only a few "pixels" per inch) This is a very difficult and subjective question to relate directly to camera resolution.

Some would say it is hardly worth going above 10 MP, as you generally won't see much difference under normal viewing conditions regardless of the size printed.

Of course, FF also does not define megapixels. You can have a much lower megapixel FF camera than with crop.
Don't look at FF as having higher resolution.
I'd have to do the math, but I believe I've read that if the K1 has a 36Mp FF sensor, that would be slightly lower resolution than the 24Mp crop sensor in the K3, for example. (I think you'd have to go to 42+Mp to get similar resolution). Assuming your *only* criteria was Mp, the FF sensor might give your "worse" images. :-)

Unless you can define a need for FF v.s. Crop, there is no real need to change.
There may be reasons to go to FF. Plenty of reasons to stick with crop sensor, as well.

But printing size is not one of them.


You want sharp detailed images? Use good lenses.
Do not use a camera with an AA filter unless the need exists to reduce moire, etc...
If you don't need flash and subjects are perfectly still, use "Pixel Shift" to increase color resolution and sharpness.
Ultimately, get a medium format camera. The sharpness for the resolution is noticeable. But again it isn't the Mp that makes the difference.

[dtmateojr]

I've been in front of a huge 85" 4K TV and at a distance of about 1.5m I honestly could not see the pixels. My conclusion is that 8Mp is big enough for anything.

[amoringello]

I've been in front of a huge 85" 4K TV and at a distance of about 1.5m I honestly could not see the pixels. My conclusion is that 8Mp is big enough for anything.

I believe the suggested viewing distance for a screen that size should be more like 3-4m. You're going to burn your eyes out.
Although honestly, who wants to sit that far back when their entire field of view can be filled with video! Nice!

[dsmithhfx]

the bigger you print, the further back someone is likely to be when viewing the image, so the less resolution you need.

This is not necessarily so. Some medium (& especially large) format images retain tremendous detail even when printed large, such that you can examine the printed image from a fairly close distance to appreciate the detail. I have no idea how that translates to megapixels, but I imagine there is some correlation.

[amoringello]

This is not necessarily so. Some medium (& especially large) format images retain tremendous detail even when printed large, such that you can examine the printed image from a fairly close distance to appreciate the detail. I have no idea how that translates to megapixels, but I imagine there is some correlation.

One can nit-pick anything. I guess I should have (and thought I did) qualify the statement with "... in general..."
My apologies.

In general what I said can still be true. Even with higher resolution source images, the printing process may (or may not) actually retain the detail. It depends on the decision made about the resulting print and need.

Regardless, back to the OP's issue; a crop factor v.s. full frame sensor is not likely to be the sole factor on how that final images looks.

Like I said, it was a difficult and subjective question to answer.

---------- Post added 02-17-16 at 11:22 AM ----------

FYI, here is some good starting places for determining resolution for printing.
I could not find the page I wanted, but this gives a rather decent start for what to think about. You can certainly search the Internet and keep yourself up for the next few days reading technical detail about what works and why (which again boils down to being somewhat subjective because no two eyes are exactly the same).

Resolution for print viewing distance

Another thing to add, is the medium you're printing on... glossy prints more detail than matte, since the ink tends to bleed more on matte paper. So printing higher resolution can be a waste. Even more so on canvas with rough texture. So instead of starting with 300dpi, you might be just as good with 150dpi.

Still, none of the relates directly to crop vs full frame.

[normhead]

The only thing for certain is that some people will pay huge dollars for pictures that are very detailed, and some will pay huge dollars for images with less detail but that are superior compositions by well known artists. The thing is, the idea is not to try and do what you don't like to do, the object is to find out what you like to do. Whatever it is, there's almost sure to be a market out there if you find the right people.

But if you're looking for something that will satisfy everyone, there would seem to be enough people that don't require high res images to make it moot, unless that's what turns your crank. Shoot what you love, find people who like what you shoot. Don't worry about what every one else is doing.

With my Gray Jays, I was going to see them anyway, might as well take along a camera, and I'm glad there's a few folks that like the pictures. The fact that 99% of the forum don't follow, don't care, don't pay attention at all is irrelevant. I shoot them and post them for folks who enjoy them.



You can ask us all you want about resolution for prints. What you need to do here is decide what makes you happy. I've found, taking an image I like, and printing it at the size I want to see it, almost always produces a satisfactory result. Once you are over 10 MP, it's very unlikely you will ever produce a print you don't like from a picture you like, no matter how large you go, if for no other reason than because you are unlikely to print so large you'll ever go beyond the point where the composition of the image is detracted from, by the resolution of the print.

There are a lot of folks on the forum blowing smoke about they have to have this or that resolution, ignore them, odds are, you aren't them. And if you are, you'll find out real quick.

[Just1MoreDave]

I think all current DSLR technology is fine at 20"x30". Sure, you can get caught up in pixel density calculations and print tech, showing one to be "better" (more) than another. But if someone is standing six inches away from your large print looking for pixels, the print is a failure for them.

[vonBaloney]

What is always glossed over in these discussions is that there is skill involved in making big prints, or any prints. Make prints with whatever you've got now. Learn how to make them better (this could just mean "learn how to prepare better the file you send to lab to be printed", learn how to upsample effectively, learn what different paper mediums look like, etc). Now make bigger ones. At the point you've exhausted your skill and you're dissatisfied with the results you are getting, then with the knowledge you've gained in getting that far you will know exactly what you need to do to get the results you desire. It may be upgraded gear with more MP, or full-frame, better lenses, or simply "I really need to start using a tripod". (Or all of the above.)

If you're thinking "but it should be great out of the camera, and I shouldn't do have to do anything but hit the print button or send the file to the lab unaltered", then your results will never be great (for the gear you have, as you'll need a 24MP camera to get results you could have gotten with a 10MP camera with some extra care taken), and the answer of how to make bigger better prints will always be "better cameras, sharper lenses, more MP, bigger sensors", etc because you relying solely on the gear and nothing beyond it.

[Zephos]

Been trying to find out how the printing is effected by number of megapixels and format. Is it worth it going up to full frame ? The real difference doesn´t seem so big when you read about it.
So, what should you look for ?

Just learned about this in my Physics class:

Optical resolution is limited due to the diffraction of light. The formula is like this:

Minimum separation between two printed object (pixels) = (1.22 X Distance from observer to objects X wavelength of light) / diameter of your pupil.

If we choose a wavelength of red 750 nano-meters, it will give us the minimum separation for visible light in general. The human pupil varies between 1.5 and 8 mm, so we can choose the median of 4.75. The formula simplifies down to:

min separation in meters = 1.926 e^(-4) X Distance from observer in meters

This is actually how Apple's Retina displays work. They pack the pixels so close together that your eye is incapable of resolving them. So yes, if your separation between pixels becomes large enough, and your observers are close enough, a full frame would help. Only playing with the numbers will tell you the answer though.

[normhead]

Just learned about this in my Physics class:

Optical resolution is limited due to the diffraction of light. The formula is like this:

Minimum separation between two printed object (pixels) = (1.22 X Distance from observer to objects X wavelength of light) / diameter of your pupil.

If we choose a wavelength of red 750 nano-meters, it will give us the minimum separation for visible light in general. The human pupil varies between 1.5 and 8 mm, so we can choose the median of 4.75. The formula simplifies down to:

min separation in meters = 1.926 e^(-4) X Distance from observer in meters

This is actually how Apple's Retina displays work. They pack the pixels so close together that your eye is incapable of resolving them. So yes, if your separation between pixels becomes large enough, and your observers are close enough, a full frame would help. Only playing with the numbers will tell you the answer though.

There are a couple of criticisms of this. Red is one end of the spectrum for example other use yellow/green as a reference. You get many times more resolution out of the blue spectrum, but using red would be as bad as using blue, and claiming APS-c can give you 100 MP imAGES which it will on APS-c.

The human pupil varies between 1.5 and 8 mm, so we can choose the median of 4.75.

Not really, you have to decide what you have and what looks good to you. What your pupil sees at 4.75mm is not really applicable to what it sees at 1.5 or 8mm.

So yes, if your separation between pixels becomes large enough, and your observers are close enough, a full frame would help.

You haven't said anything. We know that. The only thing that matters is the point at which a full frame will help. This tells us nothing. The magical point at which a full frame will help has never been defined, and for good reason. It's different for every images. And on top of that, science has never established the physical requirements for a person, to perceive an image to be "sharp enough". IN photography we deal with not "sharp focus" but "acceptable focus." A completely subjective term that could be established only by multiple large scale studies of multiple images and subjects.

Only playing with the numbers will tell you the answer though.

Actually playing with the numbers doesn't tell you anything, getting out and printing and looking at some images will help. Physics is pretty much useful until variables such the eye sight of the viewer, the viewers likely minimum distance from the print, how much the resolution in blue makes up for the lack of resolution in the red spectrum, this has to be done image by image, because sometimes the combination of blues and reds can lead to the impression of resolution in the reds, even if there is very little red resolution. I have never seen anyone in physics take an image and actually apply all the variables that have to be accounted for to come up with any kind of answer supported by numbers. I've never even seen a post by anyone using physics that has even acknowledged mathematically all the variables that exist, forget about applied that knowledge in an empircle fashion to see if they are right. What happens when you apply physics is you apply the worst case scenario to everything, like assuming the predominance of red light, and coming up with an overkill situation that would not be necessary for a vast number of prints. If you go with this type of science, you will always predict way more than your actual requirements that are necessary.

Amateur physicists who attack this kind of problem are always way to lazy, and simply ignore the more complex variables to come up with a number, which almost always means next to nothing in the real world.

Your instructor is teaching you physics, and doing a fine job, these are fine instructional tools that are showing you how to use physics to attack certain problems. But it falls well short of answering the question asked, in anything but a strictly theoretical sense. In a practical sense, it's not even good theory.

For that to happen, their has to be an empirical proof, using actual people and actual prints to show the correlation between the physics being applied and the perception of acceptable focus. That, as far as I know, no one has ever provided.

Retina Display is a marketing term developed by Apple to refer to devices and monitors that have a resolution and pixel density so high – roughly 300 or more pixels per inch – that a person is unable to discern the individual pixels at a normal viewing distance.

The fact that 300 pixels per inch obliterates individual pixels when holding a phone in your hand actually means very little about your perception of a 100 DPI print that has been upscaled to 300 dpi and then printed. The individual pixels are still invisible , and detail that goes to a resolution of 100 lines per inch is still quite sharp. The mistake is in assuming that the 100 DPI image will be viewed, printed at that resolution. With digital, there is no reason to do that. You can expand your "negative", unlike film where you are stuck with the size film you shot the image on.

So the real question here is, how much can I upscale an image before it starts to look soft from pixelation etc? I've seen images upscaled to 300 dpi from 72 DPI images that look quite good, at 20x30 inches, and has sold multiple times.

Until physics comes up with a formula that accounts for that, it's all academic.

[Zephos]

There are a couple of criticisms of this. Red is one end of the spectrum for example other use yellow/green as a reference. You get many times more resolution out of the blue spectrum, but using red would be as bad as using blue, and claiming APS-c can give you 100 MP imAGES which it will on APS-c.



Not really, you have to decide what you have and what looks good to you. What your pupil sees at 4.75mm is not really applicable to what it sees at 1.5 or 8mm.



You haven't said anything. We know that. The only thing that matters is the point at which a full frame will help. This tells us nothing. The magical point at which a full frame will help has never been defined, and for good reason. It's different for every images. And on top of that, science has never established the physical requirements for a person, to perceive an image to be "sharp enough". IN photography we deal with not "sharp focus" but "acceptable focus." A completely subjective term that could be established only by multiple large scale studies of multiple images and subjects.



Actually playing with the numbers doesn't tell you anything, getting out and printing and looking at some images will help. Physics is pretty much useful until variables such the eye sight of the viewer, the viewers likely minimum distance from the print, how much the resolution in blue makes up for the lack of resolution in the red spectrum, this has to be done image by image, because sometimes the combination of blues and reds can lead to the impression of resolution in the reds, even if there is very little red resolution. I have never seen anyone in physics take an image and actually apply all the variables that have to be accounted for to come up with any kind of answer supported by numbers. I've never even seen a post by anyone using physics that has even acknowledged mathematically all the variables that exist, forget about applied that knowledge in an empircle fashion to see if they are right. What happens when you apply physics is you apply the worst case scenario to everything, like assuming the predominance of red light, and coming up with an overkill situation that would not be necessary for a vast number of prints. If you go with this type of science, you will always predict way more than your actual requirements that are necessary.

Amateur physicists who attack this kind of problem are always way to lazy, and simply ignore the more complex variables to come up with a number, which almost always means next to nothing in the real world.

Your instructor is teaching you physics, and doing a fine job, these are fine instructional tools that are showing you how to use physics to attack certain problems. But it falls well short of answering the question asked, in anything but a strictly theoretical sense. In a practical sense, it's not even good theory.

For that to happen, their has to be an empirical proof, using actual people and actual prints to show the correlation between the physics being applied and the perception of acceptable focus. That, as far as I know, no one has ever provided.



The fact that 300 pixels per inch obliterates individual pixels when holding a phone in your hand actually means very little about your perception of a 100 DPI print that has been upscaled to 300 dpi and then printed. The individual pixels are still invisible , and detail that goes to a resolution of 100 lines per inch is still quite sharp. The mistake is in assuming that the 100 DPI image will be viewed, printed at that resolution. With digital, there is no reason to do that. You can expand your "negative", unlike film where you are stuck with the size film you shot the image on.

So the real question here is, how much can I upscale an image before it starts to look soft from pixelation etc? I've seen images upscaled to 300 dpi from 72 DPI images that look quite good, at 20x30 inches, and has sold multiple times.

Until physics comes up with a formula that accounts for that, it's all academic.

Uh.... wow you need to get a hobby or something if you are going to rip apart my post like that. There IS a real physical limit to how much detail your eye can perceive, and using the optical resolution formula might help to estimate that point (the same point you mention of pixelation). Anyway, I was only trying to be helpful and I was excited that I actually had learned something that could be applied to the question at hand. I'm sorry that I offended you with my "lazy science" because I must have to warrant such a response.

My choice of 750 nm wavelength is based on the formula. To get the largest minimum separation, so that it would work for all visible light, I chose the largest wavelength.

The diameter of the human pupil varies between 1.5 and 8mm, and because I was doing a quick estimate, I chose the midrange. Obviously if I knew more about the lighting conditions I would be able to make a better choice, but seriously who is going to go measure the observers pupils? I never said this was an exact calculation and made it abundantly clear how I chose those numbers.

I understand that there is a lot more that goes into it, and simply stating so and mentioning some examples would have been enough for me to realize that my calculations were too simplistic. Picking it apart like you did stinks of snobbishness and reminds me of the elitism of other forums such as StackOverflow.

Sincerely,
Lazy amateur physicist

[normhead]

Uh.... wow you need to get a hobby or something if you are going to rip apart my post like that. There IS a real physical limit to how much detail your eye can perceive, and using the optical resolution formula might help to determine certain things. I never claimed to be an expert or that this was the only way or that I knew everything involve, I was only trying to be helpful and I was excited that I actually had learned something that could be applied to the question at hand. My choice of 750 nm wavelength is based on the formula. To get the largest minimum separation, so that it would work for all visible light, I chose the largest wavelength. Sorry for trying to be helpful, but next time you might try posting your own answer instead of attacking mine.

One person's "ripping apart" is another person's "correction." Learn the difference.

There IS a real physical limit to how much detail your eye can perceive, and using the optical resolution formula might help to determine certain things.

Science with a "might" in it often needs years more work to get rid of the "might". So far, I'm not even sure you know what the limits of what you are suggesting are. So I'll just say what you are suggesting "might" be somewhat useful at some point, or it "might not." It's all in how the parameters are defined and on what research they are established. Right now there is no research to indicate that resolution at the limits of human perception are necessary for an image to be judged "acceptably sharp." while there are endless examples of images that fall well below the thresholds of human perception that are considered to be great art. To say that resolution will ever determine print quality, based on the way human's experience of it, is a bit of a stretch. What people perceive as enjoyable is different from "what is the highest resolution" people can perceive. No relationship has been established between the two.

[Zephos]

One person's "ripping apart" is another person's "correction." Learn the difference.



Science with a "might" in it often needs years more work to get rid of the "might". So far, I'm not even sure you know what the limits of what you are suggesting are. So I'll just say what you are suggesting "might" be somewhat useful at some point, or it "might not." It's all in how the parameters are defined and on what research they are established. Right now there is no research to indicate that resolution at the limits of human perception are necessary for an image to be judged "acceptably sharp." while there are endless examples of images that fall well below the thresholds of human perception that are considered to be great art. To say that resolution will ever determine print quality, based on the way human's experience of it, is a bit of a stretch. What people perceive as enjoyable is different from "what is the highest resolution" people can perceive. No relationship has been established between the two.

Okay okay okay. Obviously you take this very seriously. Thank you for correcting my blatant and intentional lies about science. If it weren't for good people like you we would all be fooled because as you know, we all believe everything we read on the internet.

All snarkiness aside, I have no ill feelings about your posts and hope you have a good day.