Tony Northrup - D810 vs. 5D Mk3

[LOALTD]

If i had to start over today, i'd go with d810 over the older 5d3. But i don't want to go through the hassel of selling all my lenses so i'll wait to see how Canon responds. Competition is great for all us consumers and Canon could use a hard kick in the butt like this to start stepping up.

YES!

This is exactly where I am at. I talk a lot of smack about Canon bodies and generally steer people towards Nikon...

And occasionally they'll be all "wait...don't you shoot Canon?" Like I'm trying to hide my secret from them!

It's just...it's too late for me.

I'm not into losing hundreds of dollars to sell and then re-buy lenses.

I'll just sit tight and hope the next 5D gives us some more dynamic range and some 4k.

 

[dtaylor]

Are you saying that the 'naturally crisper, sharper images' from the 36 MP D8x0 can be readily and easily distinguished from images taken with a 24, 22, 20, or 18 MP FF sensor when downsampled for small prints or web-sized images?

I don't know about him, but I will go on record as saying that anyone who walked into a gallery of 24x36" prints made using both cameras, but by the same photographer, would never know the prints came from two different cameras.

Even at larger sizes I don't think anyone would know unless you showed them identical prints (i.e. same scene and all factors but sensor made equal) and asked them to look for differences. Then a critical reviewer observing very closely (i.e. 6-12" from the print) might pick up a difference here or there.

Northrup said the 5DIII was ok for Facebook, but you're saying the D810 would be better for Facebook?

The claims just get more and more ridiculous, don't they?

 

[docsmith]

Sorry, but there is a relatively large difference in detail between the two. That has nothing to do with dynamic range, simply the pixel size. The D800 resolves quite a lot more detail than the 5D III. That either translates into the ability to enlarge more with the same level of detail as smaller enlargements with the 5D III, or it translates into naturally crisper, sharper images when downsampling for smaller prints or online publication.

Maybe with the very best of Nikon's lenses. But for even their very good lenses, there really is not much of a difference. Which may explain the observations made by the photo club.

There are many comparisons, but, a quick one, the new Nikkor 58 mm f/1.4 on the D800 vs the old Canon 50 f/1.4 on the 5DIII result in 18 p-Mpix and 17 p-Mpix, respectively (using DXO data). I doubt anyone would be able to discern a difference between those two systems.

But yes, put the very best of lenses on the D8xx and you have more resolving power.

 

[dtaylor]

Not to mention the 85mm f/1.2's enormous CA on the sides such that it doesn't sharpen up until somewhere between f/2.8 and f/4:

Oh boy...

http://www.the-digital-picture.com/Reviews/ISO-12233-Sample-Crops.aspx?Lens=397&Camera=453&Sample=0&FLI=0&API=0&LensComp=732&CameraComp=614&SampleComp=0&FLIComp=0&APIComp=0

The Nikon is sharper across the frame, but that is some of the most disgusting CA I've ever seen. I've seen Chinese c mount lenses with less CA!

I'm not going to wade into the AF mud fest because without reproducible tests...or a really calm, rational, ego free group of people who also have hundreds of hours of stick time on both...it's pointless.

 

[Sporgon]

Sorry, but there is a relatively large difference in detail between the two. That has nothing to do with dynamic range, simply the pixel size. The D800 resolves quite a lot more detail than the 5D III. That either translates into the ability to enlarge more with the same level of detail as smaller enlargements with the 5D III, or it translates into naturally crisper, sharper images when downsampling for smaller prints or online publication.

"Large difference in detail" #1

http://s11.postimg.org/kyh7iwp8j/5d3_d810_1.jpg

"Large difference in detail" #2

http://s16.postimg.org/i7ha1whwl/5d3_d810_2.jpg

"Large difference in detail" #2

http://s24.postimg.org/swj8ybket/5d3_d810_3.jpg

All images: left 5D mark III, right Nikon D810, DPReview studio scene RAW files at ISO 100 / NR0 converted using ACR standard settings. 5D3 file was resized to match (Bicubic Sharper) and then had light sharpening applied. I encourage everyone to download test files such as these and try them for yourself.

At best one can say tiny high contrast details are rendered a little better at pixel peeping sizes, and the D810 will sometimes pick up light, small patterns or textures missed on the 5D3. Shown at 50% on a 96 ppi monitor (i.e. 25x40" print) this stuff would be invisible. In print it would be invisible without a loupe.

We are well into diminishing returns for small format. People talk endlessly about AA filters and 22/24 vs. 36 MP yet they hardly matter at all even when pixel peeping tripod mounted shots made with primes at optimum apertures. Subject lighting, lens, technique, post processing, each 1,000x more important to fine detail and sharpness. I hope when Canon releases a high MP body they A) don't charge 1D prices, and B) shoot past 50 MP even if high ISO suffers. Because at this point I'm wondering if even >50 MP would yield truly significant differences in print without also jumping to a MF sized sensor.

This, btw, is why I'm going with the Sony A7 over the A7R. I see the exact same thing there. I would rather have electronic first curtain shutter, the phase detect points, and the extra cash in my pocket for lenses.

I think this is why we haven't seen a 'very high' mp FF sensor from Canon yet. Once you go beyond a certain level of potential resolution the format becomes the restriction in the sense of the magnification (focal length of lens) that you have to use pro rata, the size of image capture and then the magnification required to view the image. So smaller format equals shorter focal length (less magnification) equals smaller capture equals greater magnification to view.

This is why the 'large FF mp' camera doesn't excite me very much.

Incidentally this is why ( excluding dof, low light etc etc) you just can't tell the difference in a (moderately enlarged) image from APS against FF when the subject is close to you and filling the frame, but you can when the subject detail is much further away, ie landscapes.

 

[dtaylor]

But yes, put the very best of lenses on the D8xx and you have more resolving power.

Put the same lens on the 5D3 via an adapter and...no, not really. Not what people imagine. There is a difference, it's just incredibly difficult to reveal.

 

[unfocused]

The D800 resolves quite a lot more detail than the 5D III. That either translates into the ability to enlarge more with the same level of detail as smaller enlargements with the 5D III, or it translates into naturally crisper, sharper images when downsampling for smaller prints or online publication.

Sorry, but this is just crazy talk. Ink is not infinitely scalable. It is a liquid that has to be absorbed into the paper. Which means it must bleed or it wouldn't adhere to the paper. It isn't like pixels, where you can keep getting smaller and smaller and with sufficient magnification still see individual pixels.

And, as far as online publications go, the limiting factor is the resolution of the monitor, which is always way, way less than the resolution of any camera.

The D810 might – emphasis on might – offer some advantage for those making very large prints. But, ordinarily, large prints are viewed at a greater distance, so once viewing distance is taken into account, the difference in resolution is generally negligible.

Tony Northrup made a transparent infomercial. If he isn't yet on Nikon's payroll, it isn't for lack of trying on his part.

 

[dtaylor]

I think this is why we haven't seen a 'very high' mp FF sensor from Canon yet. Once you go beyond a certain level of potential resolution the format becomes the restriction in the sense of the magnification (focal length of lens) that you have to use pro rata, the size of image capture and then the magnification required to view the image. So smaller format equals shorter focal length (less magnification) equals smaller capture equals greater magnification to view.

Yep. Saddens me to say this, but I don't think we are going to get those wall sized detailed prints you see from LF film / MF digital shooters in 35mm format (except for stitching of course). A somewhat affordable Canon MF system with brand new lenses designed to the best of their abilities would excite me much more then a 50 MP Canon DSLR. (But even "somewhat affordable" would probably land in the $20k range for a minimum kit.)

Incidentally this is why ( excluding dof, low light etc etc) you just can't tell the difference in a (moderately enlarged) image from APS against FF when the subject is close to you and filling the frame, but you can when the subject detail is much further away, ie landscapes.

I would go a step further and say it can be hard to tell even in many landscape shots. But basically we're in agreement here.

We're in a golden age of photography. A relatively poor guy can pick up a Rebel and a decent lens and produce IQ nearly as good as a D810 / A7R with the latest, most expensive lenses. It should be even more about the image content now, but instead we become even more fixated on tiny differences and exaggerate their significance to silly proportions. (Isn't this a known phenomenon in psychology? Where people presented with many very good choices become more stressed and analytical then when they have one good and a few worse choices?)

When I step back and think about how far technology has come, I am completely blown away by what a $300 EOS M + EF-M 18-55 cheapo kit lens + ACR/PS can produce. Jump in a DeLorean and take that thing back to the 1990's and it literally would have been able to hang with medium format film up to 16x24".

 

[dtaylor]

Tony Northrup made a transparent infomercial. If he isn't yet on Nikon's payroll, it isn't for lack of trying on his part.

Having him on your payroll would backfire. His work is so transparent, late night infomercial that anyone watching it would immediately assume your product was actually the weaker one and you were trying to bash the competition ;D

 

[Loswr]

"Large difference in detail" #1

http://s11.postimg.org/kyh7iwp8j/5d3_d810_1.jpg

"Large difference in detail" #2

http://s16.postimg.org/i7ha1whwl/5d3_d810_2.jpg

"Large difference in detail" #2

http://s24.postimg.org/swj8ybket/5d3_d810_3.jpg

All images: left 5D mark III, right Nikon D810, DPReview studio scene RAW files at ISO 100 / NR0 converted using ACR standard settings. 5D3 file was resized to match (Bicubic Sharper) and then had light sharpening applied. I encourage everyone to download test files such as these and try them for yourself.

I don't really see significant differences. Since 36 MP >> 22 MP, I must conclude that your test is flawed.

Phhhtththhhththththt

</sardonic trolling>

 

[dtaylor]

I don't really see significant differences. Since 36 MP >> 22 MP, I must conclude that your test is flawed.

Well...in fairness DPReview did not underexpose a black cat in a coal mine and push 5 stops with all noise reduction turned off.

If they had done that...game over man.

Update: I added this to the original post, but these were from Imaging Resource, not DPReview. I had them on my HD already from a prior comparison and just typed the wrong site.

 

[9VIII]

This reminds me of discussions on the maximum resolution that people can see.
People get really confused about that for various reasons, but the crux of the argument always comes down to diminishing returns, or "price to performance".
1080p is great, but 4K is better, how much better? Depends on who you ask, how much it cost, how it affects the surrounding ecosystem, etc...
20MP is great, but 40MP is better. How much better? ...

 

[3kramd5]

1080p is great, but 4K is better, how much better?

It depends on size and viewing distance. Weird.

 

[Loswr]

Double your resolution from 20mp to 40mp, you could downsample the 40mp image to the 20mp sizes, and have considerably better data that looks sharper, crisper, clearer without any additional sharpening.

What happens when you downsample both to 2 MP for web display, or make an 8x10" print? Do you stand by your earlier claim that a 36 MP image would yield a "naturally crisper, sharper" result than an 18-24 MP image?

At some point in the future, you guys are going to be the new generation of "film forever" guys.

No, we're just honest about the relevance and impact of the comparatively minor technological differences of which you're apparently so enamored.

 

[V8Beast]

Why is someone whose portfolio is full of glorified snapshots so concerned about his equipment?

 

[3kramd5]

Are you saying you can slap on a fast prime, focus dead center, and then rotate ("recompose" generally involves much more rotation than translation, obviously) the camera about its own axis so the initial subject as seen by the AF unit goes from one side of the frame to the other (or at least from the borders of the PDAF sensors) and significantly out of the DOF and it will track it?

If so, that's cool and I do not believe my 5D3 could do such.

LOL, thank you for writing this And, yes, that's exactly what I'm saying. Again, just watch this video:
https://www.youtube.com/watch?v=C5J7zALyHic

It doesn't look like the camera is rotating much (maybe 5 degrees with a rather wide DOF).

I'll give it a shot this evening with a 5D3 just to satisfy the curiosity.

Okay, it's hard to replicate since my camera doesn't illuminate the point it's using during tracking, so I took some snaps at extremes and, looking at them in LR, it did fine.

However, putting a pronounced object like that on a continuous white background isn't really indicative of real world performance; that's about the easiest scenario imaginable. With a complicated background, I don't know how well it would keep up.

 

[Orangutan]

Why is someone whose portfolio is full of glorified snapshots so concerned about his equipment?

I've lost context -- to whom are you referring?

 

[lucasjld]

Isn't comparing a march/2012 camera with a mid/2014 one like comparing the same gap between two macbooks (cpu upgrade/gpu/higher bech scores/etc?)
Two more freaking years, one is an upgrade of the former "rival".

I just think it would just be stupid to sell something two years "younger" with nothing better.

 

[V8Beast]

Why is someone whose portfolio is full of glorified snapshots so concerned about his equipment?

Did you just call all my photography glorified snapshots? ??? I like to think I'm at least a step above that. I don't consider my work to be the best, but neither do I consider it to be the worst...

Sheesh...I really don't understand this community anymore.

No way, dude I was referring to Tony Northrup, the guy whose video this thread was originally about before the discussion got derailed ;D I can see how the tangents that have followed could lead to confusion, though ;D

I don't consider my images to be the best either, far from it in fact, which is why I don't feel the need to create videos stating how useless my current gear is.

 

[DominoDude]

Since I see mentions of video and watching distances...
I had a great chart showing suitable viewing distance for various resolutions and screen sizes, but it's embedded on a site that is highly linked to very X-rated content, so I think we should skip that URL. However, it showed the farthest distance at which an eye with perfect vision could resolve all the detail. In short, and as an example, it boils down to a 60 inch screen best being viewed from less than 10 feet if you want the eye to resolve all detail in a 1080p movie.

Let's toss in some Wikipedia that is less prone to being X-rated -> http://en.wikipedia.org/wiki/Optimum_HDTV_viewing_distance#Human_visual_system_limitation

I'd say that for a family of 4-5 to watch and enjoy every minute detail of a high-res video snippet they better huddle together really close. Thankfully the human brain is supposed to watch and enjoy the content, and it won't throw itself on the floor in a temper tantrum just because every single pixel isn't distinguishable.