Is 5d3 pixel count enough to print large size prints?

[TrumpetPower!]

Just for clarification, I was always taught that 100% will show you the level of detail a printer will print full size at it's native resolution... So when you're enlarging, are you looking at 200%-400% to see what a print at 100 DPI would print? Just curious how other people soft proof their enlargements before printing as big prints, especially rolls and inks, aren't cheap.

That's technically true in a sense that's quite misleading.

If your display is 100 ppi (and most displays are in that ballpark) and you print, say, a 2400 px x 3600 px print at 100 ppi = 24" x 36", then you can hold the print next to the display and they'll be the same size; the display (unless you've got a 43" monster) will show an actual-size crop of the final image, and both will effectively have the same resolution.

A big part of where the misleading part comes from is that most printers these days have pixel pitches well over 2000 ppi, maybe even approaching 5000 ppi. However, each of those pixels would be a single dot of a single ink color. The display, on the other hand, can show any of a billion colors (give or take, depending on gamut, YMMV, etc.) for any of its pixels. The end result is that the effective visual resolution for either isn't all that terribly far off. The printer will be sharper, yes, but only significantly so closer than arm's length. For most human eyes and a modern printer, a 300 ppi print is about as good as it gets, though there are barely-noticeable marginal improvements to be had up to about 600 ppi. But you can make absolutely stunning prints as low as 100 ppi -- which makes sense: if you can have a stunning image on a 100 ppi monitor, there's no reason you shouldn't have an equally stunning image on a 100 ppi print.

So, really, for the most part, basically nobody should be giving a second thought to resolution these days. There's more than enough. It's your vision and skill that'll set you apart from the competition, not the megapickles of your camera. And if you're working in one of the very few settings where a modern DSLR -- any modern DSLR -- doesn't have enough megapickles to get the job done, then you're charging your clients enough that the medium format (or bigger) gear you need can slip under the petty cash radar.

Cheers,

b&

 

[dukeofprunes]

It should be noted that the maximum print size can often not be directly correlated to the number of available pixels - the perceived sharpness of a print is determined by the motif and the size of the details in it. As such, a print will seem very sharp if it contains a lot of small details that are clearly rendered by the pixels. On the other hand, a print will not seem impressively sharp if the motif contains details that are too small to be tack sharp (imagine a landscape photo with a tree where the leaves are on the edge of the resolving power of the camera and just turn to mush). In this respect, I think large size prints from the D800 and 5D3 will be very similar in most cases. In other words, the difference in resolving power of the D800 and the 5D3 is relatively small compared to the large variation in the size of details in a photo.

Looked at differently, if you walk around a photo gallery, the sharpest looking prints will not consistently be the ones taken with the highest MP cameras, it will be the prints where the motifs contain a lot of clear details.

Great thread btw! I love high resolution and the idea of being able to make big-screen-TV sized prints, but as already mentioned by others, the difference between D800 and 5D3 is not that big - they both have plenty of pixels.

Btw, how does a 22MP D800 crop compare to an uncropped 5D3 image? One of the reasons for why I like high resolution is the ability to crop if I dont have enough reach or if I in PP realize the possibility for a better composition.. I would expect the 5D3 to be sharper/less noise. Anyway, the D800 will make room for more cropping.

 

[rmblack]

Just wanted to add that if you shoot a panorama or tilt, the noise from a high ISO will be compressed, and in some situations you'll be able to shoot ISO 10,000 and it will look like ISO 800.

 

[1982chris911]

< sarcasm starts here > No the 5D MKIII is not sufficient for any prints larger than postcards ... as was none of all the other Canon bodies with lower resolution ever produced before it ... That is why no serious PRO ever used a canon to shot other things than website sized images ... < sarcasm ends here >

In reality the 5D MKII which has a lower resolution and the 1Ds III are very good cameras for large prints and if 300dpi is your goal neither the 5D MkIII nor the D800 can deliver that at 36x24 ... In all those cases a program like Genuine Fractals is used to resize the pictures to up to 10x their actual size... This is what all agencies use when they need files of this size ...

 

[Radiating]

Hello,

I am considering to to buy 5d3 but there is one thing that I am not sure of. Is 5d3 pixel count good enough to print large size prints? When I say large size i mean something along the size of 4 by 2 feet. As you probably guessed i am between d800 and 5d3.
I am not sure what else comes to play when printing large size prints but I believe ISO has to do with it. Can somebody give me some guidelines when it comes to that.

Thanks!

Thank you!

The D800 has absolutely no resolution advantage over the 5D Mark III when using zooms and low end primes this has been proven by many sources ranging from DXO to lensrentals to many minor reviews. The D800 only has a resolution when shooting with the best primes and only between f/4.0-f/8.0. That advantage is also at absolute best only an 11% increase in linear resolution, despite having more than 60% more megapixels.

The reason why both cameras will show such similar resolution is that most lenses cannot outresolve the 5D Mark III and when they do it's not by much. Canon has actually stated that it did not push the megapixels of the 5D Mark III because the lenses cannot handle it.

If you want more resolution 20 something megapixels you need to shoot with a medium format camera and with a medium format lens, the d800 is not a replacement for a medium format camera.

If you don't beleive me here's a comparison between the 5D Mark III resized to 36 megapixels and the D800 at 3350 LP/PH, this is shot with the same Sigma 70mm f/2.8 Macro, and heavily sharpened to show any difference, there is none.

Many lenses acheived more than 3400 lp/ph in the center but will show no difference slightly off center across over 85% of the image area such as the famed Zeiss 21mm. That's why to see any real difference you need a prime that acheives over 3400 lp/ph near the mid-frame and edges, which means you need a truely stelar lens - most primes do not acheive those numbers, you'll want lenses like the Nikon 85mm G, 24mm PC-E, 60mm Macro and the super telephotos on your D800 if you want more resolution than a 5D Mark III.

If you don't plan on shooting under carefully controlled circumstances with carefully selected lenses it will make no difference to you if you chose the 5D3 or D800. Either way a medium format will blow both cameras out othe water with nearly twice the resolution.

A 4 fott wide print with the 5D Mark III should be fine though.

If you're curious here's what the difference will look like if you use a high quality prime from the D800 to the 5D Mark III, the images are randomized, 5D Mark III images are resized to 36 megapixels versus D800 images.

http://www3.picturepush.com/photo/a/7801636/img/Picture-Box/crops5d3vsd800.jpg

Hope that helps.

 

[1982chris911]

The D800 has absolutely no resolution advantage over the 5D Mark III when using zooms and low end primes this has been proven by many sources ranging from DXO to lensrentals to many minor reviews. The D800 only has a resolution when shooting with the best primes and only between f/4.0-f/8.0. That advantage is also at absolute best only an 11% increase in linear resolution, despite having more than 60% more megapixels.

That is also something that many people don't realize when looking at the D800 vs 5D MK III comparison. The better results especially resolution wise and added details come from using the Nikon lenses (14-24, 16-35 II and new 24-70) ... if canon would finally be able to make a UWA that competes with the 14-24 in sharpness and overall optical quality (even the 14mm f2.8 II L Prime can't) many people would not see differences at all. BTW also the very very good 14-24 of Nikon is not able to use the full resolution of the 36MP sensor across most of the frame ... This said it is still much better than any Canon lens in this area

 

[dukeofprunes]

Thanks Radiating and 1982chris911, very informative posts.

I think it is important for pixel peepers to realize the limited resolution advantage the D800 will have over 5D3 in most cases when it is even difficult to measure under controlled test conditions!

It seems the D800 resolution advantage can only be "significant" with excellent glass under ideal conditions (e.g. studio work, good lighting, tripod), and even then it will be far lower than what is expected from the 60% more pixels! Otherwise, for many situations/shooting styles, it will not be possible to take full advantage of the resolving power of a D800 system due to lack of

-excellent optics which are sharp across the whole frame
-available light to use low iso and shutter speeds fast enough for handholdability or subject-stopping
-apertures small enough to achieve sharp focus across the image without causing too much diffraction

I find myself using 1:1 pixel mapping quite a lot during post-production for selecting the sharpest exposure and applying sharpness, NR, etc., and I realize I would probably be more content working with consistently pixel sharp 5D3 images rather than constantly being reminded that my gear, shooting conditions (or technique) was not up to par with the resolving power of the D800 sensor (across the whole image).

 

[1982chris911]

Thanks Radiating and 1982chris911, very informative posts.

I think it is important for pixel peepers to realize the limited resolution advantage the D800 will have over 5D3 in most cases when it is even difficult to measure under controlled test conditions!

It seems the D800 resolution advantage can only be "significant" with excellent glass under ideal conditions (e.g. studio work, good lighting, tripod), and even then it will be far lower than what is expected from the 60% more pixels! Otherwise, for many situations/shooting styles, it will not be possible to take full advantage of the resolving power of a D800 system due to lack of

-excellent optics which are sharp across the whole frame
-available light to use low iso and shutter speeds fast enough for handholdability or subject-stopping
-apertures small enough to achieve sharp focus across the image without causing too much diffraction

I find myself using 1:1 pixel mapping quite a lot during post-production for selecting the sharpest exposure and applying sharpness, NR, etc., and I realize I would probably be more content working with consistently pixel sharp 5D3 images rather than constantly being reminded that my gear, shooting conditions (or technique) was not up to par with the resolving power of the D800 sensor (across the whole image).

Well regarding the D800's additional resolving power I think as long as you use the full image it is not really important at all. This said the D800 does have an advantage if you intent to crop (NOT Resize !!!) the images to match the 22MP - 18MP of current Canon bodies and use only the center of your image as some ppl. here suggest e.g. for sports. It is somehow like having an APS-H 22MP body - some ppl. need this as it is about the same difference as using a 300mm lens compared to a 400mm lens which you would need on 5D MKIII to achieve the same result... However this also requires that the picture is taken with top level glass and is really 100% in focus without any blur to be significant ...

 

[dukeofprunes]

Good point. To me, cropping center image (where the glass more often will be sharp enough) seems to be the best advantage of having more MP than the 5D3. It can save you plenty of $$$ from not having to buy expensive tele lenses, and, possibly more important, it gives more freedom from switching between lenses.

Good thing this is not crucial for me. I think I will be very pleased when I eventually upgrade my 5D2!

 

[TrumpetPower!]

The D800 has absolutely no resolution advantage over the 5D Mark III when using zooms and low end primes this has been proven by many sources ranging from DXO to lensrentals to many minor reviews. The D800 only has a resolution when shooting with the best primes and only between f/4.0-f/8.0. That advantage is also at absolute best only an 11% increase in linear resolution, despite having more than 60% more megapixels.

That is also something that many people don't realize when looking at the D800 vs 5D MK III comparison. The better results especially resolution wise and added details come from using the Nikon lenses (14-24, 16-35 II and new 24-70) ... if canon would finally be able to make a UWA that competes with the 14-24 in sharpness and overall optical quality (even the 14mm f2.8 II L Prime can't) many people would not see differences at all. BTW also the very very good 14-24 of Nikon is not able to use the full resolution of the 36MP sensor across most of the frame ... This said it is still much better than any Canon lens in this area

<cough>TS-E 24mm II</cough>

Cheers,

b&

 

[1982chris911]

The D800 has absolutely no resolution advantage over the 5D Mark III when using zooms and low end primes this has been proven by many sources ranging from DXO to lensrentals to many minor reviews. The D800 only has a resolution when shooting with the best primes and only between f/4.0-f/8.0. That advantage is also at absolute best only an 11% increase in linear resolution, despite having more than 60% more megapixels.

That is also something that many people don't realize when looking at the D800 vs 5D MK III comparison. The better results especially resolution wise and added details come from using the Nikon lenses (14-24, 16-35 II and new 24-70) ... if canon would finally be able to make a UWA that competes with the 14-24 in sharpness and overall optical quality (even the 14mm f2.8 II L Prime can't) many people would not see differences at all. BTW also the very very good 14-24 of Nikon is not able to use the full resolution of the 36MP sensor across most of the frame ... This said it is still much better than any Canon lens in this area

<cough>TS-E 24mm II</cough>

Cheers,

b&

True for both L TS-E's, their resolving power is beyond anything you can have in wide angle on a Canon Camera, but they are not what I would call comparable to the 14-24 or 16-35 of Nikon ... and that is the problem. They are also not as wide - I am wondering if the optical elements of both would also make great conventional UWA primes and if yes which focal length they would be when using the very wide TS full image circle.

 

[TrumpetPower!]

The D800 has absolutely no resolution advantage over the 5D Mark III when using zooms and low end primes this has been proven by many sources ranging from DXO to lensrentals to many minor reviews. The D800 only has a resolution when shooting with the best primes and only between f/4.0-f/8.0. That advantage is also at absolute best only an 11% increase in linear resolution, despite having more than 60% more megapixels.

That is also something that many people don't realize when looking at the D800 vs 5D MK III comparison. The better results especially resolution wise and added details come from using the Nikon lenses (14-24, 16-35 II and new 24-70) ... if canon would finally be able to make a UWA that competes with the 14-24 in sharpness and overall optical quality (even the 14mm f2.8 II L Prime can't) many people would not see differences at all. BTW also the very very good 14-24 of Nikon is not able to use the full resolution of the 36MP sensor across most of the frame ... This said it is still much better than any Canon lens in this area

<cough>TS-E 24mm II</cough>

Cheers,

b&

True for both L TS-E's, their resolving power is beyond anything you can have in wide angle on a Canon Camera, but they are not what I would call comparable to the 14-24 or 16-35 of Nikon ... and that is the problem. They are also not as wide - I am wondering if the optical elements of both would also make great conventional UWA primes and if yes which focal length they would be when using the very wide TS full image circle.

When I got my 24, I did a quick-and-dirty backyard comparison between it and the 16-35 using the proverbial brick wall. And there's no comparison. Even in the far corners at maximum shift, the 24 is a significantly better 16mm FOV equivalent than the 16 (and the FOV is just about the same). (In case it's not obvious, I'm talking about a shift panorama with the 24).

Cheers,

b&

 

[V8Beast]

In all those cases a program like Genuine Fractals is used to resize the pictures to up to 10x their actual size... This is what all agencies use when they need files of this size ...

Dumb question: If you re-size an image with Genuine Fractals, how good is the end result? How much better are the results than, say, using CS5 to interpolate an image? Most my work is for editorial clients, in which case my 5DIII's resolution is plenty for two-page spreads at 300 dpi. Even my 5DC was adequate in this regard.

Genuine Fractals claims that you can enlarge an image by 1000% without sacrificing quality, which just sounds insane to me. Is this legit claim, or just marketing hype? I've got some 30x20 prints taken with a 5DC, which is about 2x it's resolution at 300 dpi, but they look pretty good to me. I'm just curious how much better the results might be with Genuine Fractals, and heck, the software is pretty cheap.

 

[TrumpetPower!]

In all those cases a program like Genuine Fractals is used to resize the pictures to up to 10x their actual size... This is what all agencies use when they need files of this size ...

Dumb question: If you re-size an image with Genuine Fractals, how good is the end result? How much better are the results than, say, using CS5 to interpolate an image? Most my work is for editorial clients, in which case my 5DIII's resolution is plenty for two-page spreads at 300 dpi. Even my 5DC was adequate in this regard.

Genuine Fractals claims that you can enlarge an image by 1000% without sacrificing quality, which just sounds insane to me. Is this legit claim, or just marketing hype? I've got some 30x20 prints taken with a 5DC, which is about 2x it's resolution at 300 dpi, but they look pretty good to me. I'm just curious how much better the results might be with Genuine Fractals, and heck, the software is pretty cheap.

You can find reviews of the various resampling methods without too much trouble, but it basically boils down to this:

Regardless of what method you use, view the image at actual size (hold a ruler to the screen and adjust the magnification until the onscreen ruler matches) and do your final sharpening that way before sending it to the printer.

If the final resolution is at least 100 ppi, just print it. Chances are excellent your print driver will do a better job at resampling it than anything else you have.

If the final resolution is at least 85 ppi, make a test print. Chances are still good that the print driver is your best bet. If you don't like the results, resample in Photoshop to somewhere in the area of 150 ppi. Try each of the bicubic options and go with whichever looks best. Be sure to sharpen again after resampling -- and, of course, revert to the unsharpened version before doing the upsampling.

If the final resolution is even less than that, you're hopefully making a billboard or a mural. Hand the best, NOT resampled image you've got to the printing agency and let them worry about it -- that's what you're paying them for.

If you're the printing agency, buy every tool there is on the market, learn which is best suited for which types of images, and how that fits into your workflow.

Note that, with modern DSLRs, you either have to crop insanely or be printing something the size of a door before you start to run into these resolution guidelines. If you're shooting with something other than a modern DSLR, then high-resolution printing obviously isn't a concern.

Have fun!

b&