Nikon D800 vs Canon 5d3 sample images.

[Daniel Flather]

I'm assuming you are comparing sample images from canon website and nikon website.

D800 images are 20mb jpeg and 5D3 images are 5~10mb jpeg.

I agree that 5D3 images seem too soft, but still I won't change to Nikon just after comparing sample images...

+1 on both points, the d800 looks sharper at 2 to 3x file size. The 5D3 pics looks worse than my 50D, so I'll assume the samples are just plain bad.

If the 5D3 can't best my 50D then Canon is f*cked.

 

[ashe]

Video samples: http://www.petapixel.com/2012/03/05/comparing-video-from-the-canon-5d-mark-ii-and-5d-mark-iii-at-iso-12800/

 

[Grum]

I have to agree that the Canon samples are fairly poor - difficult to believe on such a major release but it does look like they are not quite in focus on at least a couple of the shots.

These look pretty good though IMO

http://translate.google.com/translate?sl=auto&tl=en&js=n&prev=_t&hl=en&ie=UTF-8&layout=2&eotf=1&u=http%3A%2F%2Fwww.dcfever.com%2Fnews%2Freadnews.php%3Fid%3D6365

Though I'm definitely holding judgement until we've seen proper RAW file conversions and reviews/tests.

 

[Martin]

Hope that AA filter theory is wrong, on the other side the theory is convincing. I can belive I waited so long and the camera will produce soft images due to AA filter. After I bought canon gear and worked with it for several months I knew that I had made bad decision but I hoped that 5d3 will change a lot. I don't understand guys who say that presented official samples are sharp and ok. Samples are out of detail and very soft (not mentioning the price) We need to wait.

 

[Kernuak]

Just seeing the sample images (and I'm not giving conclusion) I admit.

All the sample I've seen from the 5D3 looks way too soft. It can be 2 things : First is post prod NR. Second is 5D3 NR. If It's the 5D3 NR they use it wayyyy too much.
Ok nice, you see no noise at 1600 ISO or even 6400 but it ruin the quality.

Regards,
David

It could also be due to JPEG compression. I would expect at least some of those images to be closer to 15-18 MB if uncompressed. Certainly most of them should be over 10 MB.

 

[altenae]

Maybe due to jpeg compression.
More likely the file size is smaller due to heavy NR.

I am afraid so.

 

[Martin]

If this is true there will be a huge disappointment on the pro/semi pro market waiting for the camera.

 

[Kernuak]

Maybe due to jpeg compression.
More likely the file size is smaller due to heavy NR.

I am afraid so.

It does looks like heavy noise reduction to me (although it could also be compression on top). From what I saw yesterday when I switched the noise reduction off (only on the back screen though), the results can be much better and don't need all that noise reduction. I seem to remember that previous sample images from new cameras also had heavy noise reduction.

 

[ashe]

Here you go... some additional samples:

Canon
http://photography-on-the.net/forum/showpost.php?p=14022610&postcount=2459

D800

http://www.fotosidan.se/forum/showpost.php?p=1792662&postcount=1457

After seeing these it looks like it is heavy NR.

 

[studio1972]

Hope that AA filter theory is wrong, on the other side the theory is convincing. I can belive I waited so long and the camera will produce soft images due to AA filter. After I bought canon gear and worked with it for several months I knew that I had made bad decision but I hoped that 5d3 will change a lot. I don't understand guys who say that presented official samples are sharp and ok. Samples are out of detail and very soft (not mentioning the price) We need to wait.

Apparently the AA filter is weaker than the one in the 5D2, so no. Anybody who buys a camera on the basis of the example images on the manufacturers website is a prize numpty in my book anyway. Wait until some respected pros (not paid by Canon) have had a go with it first, then make your mind up.

 

[hyles]

Well, if we are talking about the pictures on canon site and the one on the nikon site, i may say they are quite different. Most of the nikon are taken f8 to gain dept of field and increse detail. Nikon maximum ISO setting is 640.
Most of the canon portrait are wider than 2.8 with iso reaching 3200.
Nikon pictures are taken in studio, the canon are close to the pictures anyone can take. Don't think there are any pictures really comparable.
The nikon bride is 640 ISO f4.5, the ones on the canon sites are wider and/or higer ISO.
Actually the canon bride @ 7.1 is quite good to be shot at 3200 ISO.
And last but not least, tiger are jpg generated by the on camera raw converter.
Think that most of the people complaining for the lack of sharpness are better wait for more pictures...
Diego

 

[SomeGuyInNewJersey]

The D800 samples specifically say that they are raw files that have been processed in Nik software and are well though out photos taken to show the camera at it's best. Appropriate lenses are used for the shots too, this also makes the camera look good... As sample photos should be.

The 5d3 samples are all let's face it... S___. I really really hope they are off the camera jpg's and not processed raw's.

The one of the fall leaves looks to have been taken using the on camera hdr tool. The fact they show this image as sample with the abysmal dark fringed artifact border between the trees and sky is mind boggling. Did they give the camera to a bunch of Nikon execs and ask them to take the photos?

The 5dr images are virtually all appallingly soft and some are not even in focus. Some of this isn't the camera though, they make some of the lenses look bad as the image gets blurred as they get further from the center, I'm thinking of the 17-40mm shots here. One thing that has shown me is if I do keep my 5d3 on preorder and no get a d800 instead is that I will need to upgrade my 17-40mm ... Although Canon do not have an ultra wide that has anything like the corner to corner sharpness of the Nikon 14-24mm.

Not making a proper effort to get properly thought out, well taken and well processed sample shots seems a pretty stupid move by Canon. I know you can't read too much into sample images but it does take something away any initial buzz when they are this crap.

For these two cameras it also perhaps shows the target use. D800... Highly detailed images in good or highly controlled conditions. 5d3... Good high iso images in badly lit conditions and action shots.

It's certainly not helping me decide to stay with Canon already tempted by the "buy a d800e and 14-24mm, it's closer to the medium format cameras you can't afford" gremlin that keeps whispering in my ear.

 

[Brad Trent]

I just read through this thread (and most of the other New 5D-related posts) and I can't believe how many guys actually saying it's not important to see the same scene, photographed by both cameras side by side, before they make a decision about which camera is gonna make the better final image! My first impression was, "You can't be serious?!!", but then I really shouldn't be too surprised, based on the mountains of fanboy praise piling up about the new 5D that I had to climb over to find any comments that might have a kernel of level-headed reason to use in making an informed opinion. My God...the lengths some people will go to prove their undying allegiance to a camera company! Unless there are a few 'Explorers of Light' in attendance, it's not like Canon is paying any of you to publicly kiss their ass.

I think I'm gonna give up on trolling the boards for any further news on either the 5Dmkiii or the D800. Until these things hit the streets, it's pointless to go on when 99% of what's getting tossed about is speculation. I now know all the specs and once I can get my hands on them to do a real test, all this back-n-forth from people who just seem to wanna hear their own voices rise above everybody else (and the backbiting that ensues) is serving no purpose. Or maybe dpreview will get a full review out on both bodies that I know I can trust...

BT

 

[te4o]

I think I'm gonna give up on trolling the boards for any further news on either the 5Dmkiii or the D800. Until these things hit the streets, it's pointless to go on when 99% of what's getting tossed about is speculation. I now know all the specs and once I can get my hands on them to do a real test, all this back-n-forth from people who just seem to wanna hear their own voices rise above everybody else (and the backbiting that ensues) is serving no purpose. Or maybe dpreview will get a full review out on both bodies that I know I can trust...

BT

Not only this, Brad, you and I will have to take OUR OWN lenses to a cam-shop and give it a good test. It's not that simple with a FF above 12 MP: the lens-sensor alignment, optical micro-adjustments and I don't know what have to fit so that I can decide for or against a purchase. Seeing these samples makes me a fan of the 5D2 again, for my usage this will be enough. But I guess that the lenses used for the samples and the narrow apertures on the landscapes are spoiling the sharpness too. No way anyone can decide for or against the two new big DSLRs without seeing them not only in the hands of "independent" reviewers but YOUR hands too!

 

[JoeDavid]

I concentrated on the landscape shots because you would think that they were included to show off the resolution capabilities of the 5DM3. I saved the images locally and used DPP to see the full Canon information.

The shot of the village on the hillside was taken with the 50mm f1.2L at f8 and 1/400 sec, Image Quality setting was Fine, and a Picture Style of Landscape. To me it looks like a soft image over-sharpened a little when viewed at 100%. It worries me if that is the best it can do at ISO 100. Metering was set to Evaluative with a -2/3 Exposure Compensation setting (so much for Canon's perfect metering...).

The fall foliage shot across the water was taken with the 70-200/2.8L IS II at 85mm using f8 and 1/320 sec. Image quality was again set to Fine with a Picture Style of User Defined 1 at ISO 200. To me this shot also looks like a soft shot that has been slightly over-sharpened when viewed at 100%. This one was also metered with Evaluative metering with a -2/3 Exposure Compensation.

The last one (the one of the lake with the hills in the background) was taken at 17mm with the 17-40mm f4L; NOT the 17mm TSE. The shot was taken at f11 and 1/160 sec. Again, Image Quality was set to Fine and used the Landscape Picture Style at ISO 200. Knowing that the shot is focused fairly close judging by the foreground sharpness, I can live with the soft foliage in the distance. Evaluative metering as the others but this time no Exposure Compensation was used so the metering hit 1 out of 3.

All in all, it looks like a pretty weak showing for landscape photography. One other interesting thing to note is that they were all taken in October of last year with three different versions of the firmware...

 

[sanj]

For me its not just the sharpness but a PERCEPTION of depth and DR which hit me hard. I just loved the Nikon images.
I find it bit weird when we feel that we should wait for 'real' photographers to post their photos and not look at sample images!! I would think that a huge company like Canon would post the best pictures using the best photographers for marketing..!!! Really..
If I did not own such a huge collection of Canon glass I would switch, but as of now I have to do the best with what Canon provides - which is good enough, I should not have looked at the neighbour's plate...

 

[justsomedude]

G-dammit... this is really starting to chap my ass, and it's already starting to catch fire...

http://5dmkiii.wordpress.com/

Did I just make a bad decision with my 5D3 preorder? Some of these samples remind me of the 7D melted hay disaster....

WTF is going on here?!?!??!?!

I want to cry. :[ I hope this is just demo model results and not final shipping versions. If I had to be 100% honest... most of these "official" sample files seem to have less detail than my 5D2 images. I can't read about this anymore.

:'(

 

[Mark D5 TEAM II]

That Wordpress site looks dodgy, that's the only article on that site, heck, anyone could have put up that article (read: Nikon marketers or fanbois with an agenda) to bash the camera before production-ready units are released.

 

[justsomedude]

That Wordpress site looks dodgy, that's the only article on that site, heck, anyone could have put up that article (read: Nikon marketers or fanbois with an agenda) to bash the camera before production-ready units are released.

That's a good point about that domain - and I'm sure it's just a Canon basher who made that site. But now there's a huge thread on DPreview.com with Canon users up in arms about the same sample photos.

This post in particular is pretty interesting...

I agree that, after looking at all the samples on both the Canon Japan site and the Canon Professional CPN site, almost all the 5DIII sample Jpegs are unacceptably soft-looking at 100% viiewing, and substantially more soft than would be shots taken with my 1DsIII. Why this is, is not easily explained, but it is a very bad sign for the fate of the camera in the marketplace. I can't comment on the Nikon D800 images because I have neither looked at them, nor have I much interest in them at the moment.

Aside from the possible bad judgment and/or low talent of the people Canon chose to shoot their samples, I've come to a preliminary but worrisome conclusion, but I would like to see images from raw, shot by a good photographer and processed in DPP by a good craftsman before I could be sure about it. If, under those conditions, the results from raws look great, then I'd have to say that Canon has rolled up a big fat fail with its 5DIII in-camera Jpg processing.

But, there is another possiblity that is something I have deduced from the available specs and samples, plus statements by those associated by Canon with the 5DIII release. It's possible that Canon, in order to satisfy the potential video users of this camera who had previously complained about the bad aliasing and moire on 5DIII video, has used a much stronger AA filter on this camera to help cure it, and that the softness you see in the released Jpegs is mostly from this. There are two other points that can support this. One is that Canon is offering the unprecedented function of supposedly deconstructing the effect of the AA filter in the latest version of its raw software converter, DPP, to be released with the camera. Why do this now, unless it is especially relevant to the 5DIII? Although this might be their software answer to the Nikon D800 without AA filter, I doubt Canon could have developed this software function in the short time after which they likely learned of Nikon's plans. Also, in looking at many of the Canon 5DII samples, I noted an unusually low level of moire in many of the samples, several of which I know would certainly have displayed some level of moire if shot with my 1DsIII. I hope I am wrong about this, but I am afraid that I may well be correct in my deduction.

Regards,
David

http://forums.dpreview.com/forums/read.asp?forum=1032&message=40816151

I'm getting a little worried.

 

[LetTheRightLensIn]

Canon claims in some of their press info that they made the AA filter weaker on the 5D3. The horrendous detail loss and mush whereby even their ISO100 samples sometimes have only like 12MP if information appear due to appallingly heavy NR being applied to almost all of the sample images (and even sample videos, where the 5D2 was much noisier but had FAR more detail and the 5D2 doesn't even have close to 1920x1080 res to begin with). They have really gone out of control with their NR algorithms for in cam processing.
Even at ISO 200 and even 100, areas of modest contrast detail get NR to pure wax.

This shouldn't effect RAW at all but it does mean that you can't look at the samples and be like wow there is no noise at ISO6400 now (when there is like only 5MP of detail left).