Camera Test – Canon 5D Mark III, Mark II, 1Dx, 1Ds, Nikon D800 and Hasselbad H3D
We were messing around today as my studio partner and I both just bought the Canon 5D Mark III and some of the first files looked a little too digital. We got a little nervous and we started are own impromptu mini test. Full disclosure, neither of us ever ever ever tried a a technical test before, so we might have made a few errors. That said, we have a lot of cameras, so we thought we would share what we uncovered. Maybe it helps somebody!
By no means were we exhaustive. We shot each camera (Canon 5D Mark III, Mark II, 1Dx, 1Ds Mark III, 1Ds Mark II, & Nikon D800) on a tripod with an 85mm 1.2 lens set to I think F6.3 at 1/160, ISO 100. Autofocused and took 1 or 2 frames. We shot raw and jpg on the standard setting, K5100. The light was Broncolor Scoro with a Para88. One light, pretty head on, no fill or anything against a white seamless, no background lights, so it went grey. The only exception was for the Canon 1Dx. We don’t own this camera, so my assistant ran down to Fotocare, where they generously took a couple shots against the wall. Hence the shadow! We shot every image head to toe, and below is a screen capture of the 100% zoom in. For the Nikon D800, I am not sure if it is 100% or slightly less. Below are only the raw versions. I will say more below, but yes, I blew the opportunity to shoot the Hasselblad H3D 39MP side by side. I luckily own one and just didn’t think to put it in the mix. I will give some opinions on it in the final thoughts.
The first shot I took was with the 5D Mark III. I shot tethered and opened it up in Digital Photo Professional 3.11.26.0 (DPP). I have read some problems with the software with images coming out too soft, but I am not sure whether those have been resolved. WE DID NOT have that experience. Set to picture style Standard, with no adjustment to contrast or sharpness. The image came out too sharp. It looked digital, bitty and fake. Below, it is the middle image (4th from left or right). The skin looks spotty, and the eyes look too fake. It looks like we went crazy with unsharp mask.
We were freaking out, and I was thinking I am going to return the camera. Maybe I would get the 1Dx or maybe just stick with the Mark II. Next camera I tried was my 5D MarkII, same conditions. Below it is the 3rd to the right. I thought the skin looked much better, smoother, and natural. Though it look noticeably less sharp. I figure I could always sharpen up the eyes in post, and I would be good like always. I have been using the Mark II as my everyday camera for a couple years with few complaints.
Next, I grabbed my 1Ds Mark III. Hasn’t seen too much action recently. Though I loaned it to an assistant who normally shoots 7D a couple weeks back and he was really impressed with the 1Ds files. Below (2nd from right), the 1Ds looks great. A little warmer, but very similar to the 5D Mark II. Maybe a touch sharper. It got me thinking about whether these cameras are actually improving or if they are just adding video. That said and NOT pictured below, we hooked up a 1Ds Mark II and we could not get a frame that was acceptably sharp. Not sure if it was the camera or the focusing, but we tried a few frames and got less then we were willing to show.
Already freaking out, thinking the baseline sharpness on the 5D Mark III is too much, my assistant mentioned we could just turn it down. Like an idiot, I am like, “Can we?” YES! So you see below (3rd from left) we moved the sharpness down from the factory setting to considerably lower in Digital Photo Professional. Maybe it was Sharpness 1? Looking at it now, it seems like we probably went a little too far, but the image is way more acceptable and looks like an actual photo. I was pleased enough that I think I will keep the camera. That is a file, that with the right retouching, will look great.
Next my assistant ran down to Fotocare and they grabbed the rental copy of the Canon 1Dx. I believe they shot with Broncolor still, but I wasn’t there. Notice (2nd from left) the same TOO sharp quality too the image on the standard settings. Again, we tried lowering the sharpness in DPP and I think the file looks great. You see it is a little warmer, but that could be just from having a different light source/modifier. I cannot be sure. You can also see it is a touch smaller at 18mp.
We didn’t stop here. My other assistant just purchased the Nikon D800. So we put it on the tripod for whirl. Still an 85mm lens, but he has the Sigma F1.4. We still shot the same settings and the light had not moved or been altered. My first thought was that I really liked the image (Far right). I thought it did much better with the highlights, so that the skin tone looked more even. The image was taken after the run to Fotocare, so maybe my assistant’s complexion had a bit of a change and the highlight between his eye brows might have been naturally less. Not sure. The overall sharpness looks like a good amount; not too much or too little. You can really make out the separation in the lashes. I think better than all the other cameras, even the over sharpened versions. The only thing that really scares me about the D800 is the odd Magenta cast on the hair. You can see a distinct line edging the hair that definitely does not exist and is hard to fix in Photoshop. For a professional with the need for accurate colors, that is pretty unacceptable. Overall the color seems more purple, which I don’t believe was present or accurate. As much as I would love a 36mp every day SLR, I don’t think my fashion clients would like their clothes coming out random colors. I don’t know if this is actually a big problem, but over and over the reviews are mentioning it. I think without the color concerns, I might get the D800 over the 5D Mark III, and just buy a few new lens/accessories.
Click to enlarge
If there are any difficulties enlarging this, you can download this exact file from my smugmug account. Scroll over the image and you will see a folder with a green arrow. Click that!
http://michaelcreagh.smugmug.com/Other/Camera-Test/24785600_MzsFZk
Final notes:
I am keeping the 5D mark III. I do not think there is an image quality jump from the Mark II. I don’t see it. Frankly, I don’t think the image is any better than the older 1Ds Mark III. I think the files all just have subtle characteristics. My assistant asked my why I would keep the more expensive 5D Mark III if the files are no better than the Mark II and the 1DS. Simply, I think the new focusing system is worth the extra $1000. It is such a welcome improvement. I haven’t really delved into the other features yet, but I am happy to have the extra SD card slot back and a quieter shutter mode that I had on the 1Ds. And yes, the Mark III feels better in the hand. I have the battery grip and the sealing seems better. There are so many new menu options, and I haven’t even explored the video yet.
I do want to point out that the over sharpness issue does not seem to be as apparent outside the studio. I went back and looked at a number of files that I shot of models outside with the 5D mark III and while a little over sharp, were not unacceptable like the ones above.
Jpgs. We also shot jpgs and kind of skipped the comparisons because frankly the files were noticeably worse on all the cameras. I like shooting jpg, as I shoot thousands of photos a week. I am not sure I will trust it. I have to experiment more. We did try converting the raw to jpg in Photoshop and that looked better….but that is not what I want to do. I want to shoot straight jpg and get a great file. I am going to have to try a bunch of real shooting situations and see what I think. I trusted my Mark II and felt that was acceptable.
7D. We did not include it in the test. I am biased and am not a fan of the files. I have had several assistants shoot it for behind the scenes photos and just don’t like the look. Others are free to disagree.
My big regret was not to test the Hasselblad next to the 5D Mark III and D800. I own the Hasselblad H3D 39MP. It was sitting in the studio and I didn’t grab it! When we were testing the cameras, I had no plan to post this for others. We were just trying to figure out what was right for us in a particular set of circumstances. I know my thoughts very clearly on this and after seeing the Canon’s and Nikon’s perform, I had no doubt that my thoughts on the Hasselblad would change. I am now sorry not to be able to compare it here. I personally believe, the Hasselblad is much better for this type of studio shot. Hands down. I have shot head to toe studio shots with the H3D for over 5 years now. The eyes are sharp while the skin is smooth. Period. I saw the video from The Camera Store comparing the D800 to the H4 40MP and maybe the sharpness is comparable. But the Hasselblad owners in the video said 10 times; the skin tones. I totally agree. The look and feel are just so different. Of course, the Hasselblad is not a flexible or fast camera. I don’t trust it at 200 ISO and I just did a major fashion shoot on location mixing strobes with daylight. It looked beautiful, but what a pain in the butt. Canon and Nikon would have been so much easier.
I know this is NOT fair and was unplanned. But out of guilt, here is the same 100% crop from a head to toe image from the Hasselblad H3D 39MP. It is the first image I found on my hard drive that was shot with the same Broncolor Para 88, one light only on a seamless. It has a different amount of diffusion and is closer to the model here than in the test above. Shot with an 80mm lens, ISO 100 at F8 and handheld, no tripod. It is unretouched, but of course this is a model with makeup done. Which will make a pretty big difference. The H3D Hasselblad is a raw only camera and I have it set to sharpen ZERO. That is the image on the left and it meets or exceeds all the Canons or Nikon in sharpness. Like the D800 there are some odd color artifacts in the hair (I don’t see those often, but have seen them from time to time. I think if I got them more often I would be upset and have the same concerns as the color on the D800). The rest looks fantastic. Look at those smooth skin tones while we have great detail to play with. On the image on the right, I barely sharpened the whole thing with Unsharp Mask at 70 with radius .9. Skin still looks great, while the details are even better. The closest file from the Canons and Nikon, seems to be the D800. I would really love to see how the newer Hasselblad 50 or 60 MP compare, or Leaf’s 80Mp!! My H3D has over 5 years of shooting on it. Scary to think the new Nikons and Canons do not have a better look.
Click to enlarge or use smugmug link under first photo.
Thanks!
Michael Creagh
http://michaelcreagh.net (iPad viewers + archive)
Buy My Book!!!
http://www.blurb.com/bookstore/detail/1106047
http://highartweddingphotography.com
http://michaelcreagh.tumblr.com/
i think the 1ds mk3 looks the best, and i agree the 5d mk3 looks over sharpened and synthetic. I dont like the look of the d800 image- it looks like camera shake
but seeing as you used flash that is obviously impossible, very strange.
I agree with you about the H3D images, their great. I have had the pleasure to use them regularly whilst in uni. The only problem i find with the blads is the phocus software.
Not only did we use Broncolor Scoro flash, but also had the D800 on a tripod. The H3D shot was on a different day and was actually handheld (but with the Scoro Flash).
I personally like the Phocus software. Maybe I am not that picky as I usually use the Canon software. I never really got into Capture One. I used it a few times. Hasselblad’s old Flexcolor software was pretty bad in comparison. One good thing about Phocus…..it is free.
as i already own a D800 for months i can assure you that the picture comparison of the D800 is not perfectly in focus. (or you use a pretty bad lens what i don’t think so).
this might be slight but shooting often at 1.4 with the 85mm 1.4G this camera forgives nothing, and if it appears quite okay, just try it out in LV mode. the difference will surprise you. A big amount of D800 and D4 are suffering from back focusing issue proportionnally to the distance with the subject. this affects only the autofocus mode not the live view. This is the painful way i discovered the way my D800 needed AF calibration. as i discovered that i tested all AF focus points and all left part was damaged.
even more painful is the way the camera came back from repair with the left part issue fixed but a fucking huge front focus in the center hardly recoverable even with micro adjustment at maximum.
as i tried also the 5D3 which is delightful to use, i’m still not gonna change because of the potential of the d800 files i find way more interesting. i will just hope they move their fucking ass at nikon’s factory and fix that for real. as some reviews showed the canon even at 100 iso can hardly recovered very dark areas showing banding and colored noise. so i do use this non legend great dynamic range. BUt if nikon continues this way releasing defecting products such as D800 focus, D600 dust stuff, D7000 something else, greenish screens.. i’m not gonna last yellow very long and will wish canon improves the low iso dynamic range on next products
Or… maybe secretly dream that fuji does its come back on reflex market! for skin tones the X-pro one APS-C is already digging deep the graveyard of canikon full frames and doing as good in iso. everything for a cheaper price… If people are having fantasies on future foveon sensors, the X-trans fuji has nothing to be ashamed of.
btw, the magenta thing on the hair with the D800 probably comes with the sigma not in focus. same problem with the nikon but less significant. it actually shows your subject is front focused making the high amount of CA’s of the sigma (all wide opened 85mm anyway) obvious. this is not normal, the sigma is razor sharp on the D800 even at 1.4. i owned one for a short time but found it a bit too heavy and the AF not reliant enough unfortunately, while the optical qualities have nothing to envy the nikon i now have.
Hi Thomas
One of my Nikon friends thought that magenta cast at the edge of the hair that we got on the D800 was from the lens. He is not a fan of the Sigma 50mm 1.4 and claimed that was it. Interesting since you think so as well.
He has been pushing me to do a more extensive test and include the D4 and my new Hassleblad 50mp.
Thanks for your insight. I think they are very helpful for people to read.
Thanks
Michael
you’re welcome 🙂
i actually looking forward to see the next sigma coming under their new line line Art, Sport and Contemporary (?) if they can do new 50/85 mm 1.4 with the quality/price + AF tuning of the 35mm I’m in! same with an universal zoom lens.24-105/24-120.
i don’t think making the hassy 50 in contest is very relevant… the dynamic range of MF is different and oriented to recover highlights and not shadows. makes me wondering why the fuck nikon did the opposite on the d800 supposed to be a studio/MF killer? we can see on tests that definition of a sensor is not what will kill MF. anyway for the hassy the skin tones 16bits colors and the lack of AA filter will definitely kill the contest.. if i had 2000 bucks at the moment i’d sell all my nikon gear and go to a used H3/4D31.
that said, what about a canikon fullframe AA filter free, xtrans or foveon sensor with 16 bits encoding? 😀
this is actually already technically possible… and this time sounds more dangerous for MF industry.
Thanks Anto for your contribution. Since this post, I also bought the H4D-50mp. I like it a lot, though I find it very similar to the H3D 39mp. I do find the Hasselblad at a much higher level than my Canons.
Good to hear about the D800E. I was curious about it and have heard some other great things. I have been hesitant to change from Canon to Nikon, as I have lens, flashes, batteries, etc, etc. I wasn’t impressed enough to switch to the D800. I played with it several times with one of my assistants who owned it. But maybe to D800E would be more my speed.
Michael I have the h4d-50 and a D800E,i also had a D4 and 1Dx for a few months,the Hassy is still the best (at base ISO) although the D800E is a very impressive camera,try a portrait with a Zeiss lens if you get a chance (the new 135mm is awesome) the colors will be different,I find Nikon files look good when processed with Phase one’s C1P7..
The D4 is really only for speed or extreme condition’s.the I.Q is lacking when compared to the D800E.
IN the end Hassy is still king for me.. it’s skin tones and overall acuity is excellent…
P.S I don’t think the 1DX is much good as it has the typical pattern noise and banding problem when the files are pushed but the A.F is top notch and build and ergonomic’s too..I am looking forward to 2014 for canon’s new sensor.. hopefully it will resolve these issue’s with banding and noise..I also believe Hassy will introduce a CMOS sensor to the H line in the next couple of years..
NIce analysis from an actual owner, i wish i could put my hand on the zeiss 135 on my D800!!! I’d like to see some of your pictures if possible. but i guess i will have to wait what does nikon and sigma (135 1.8 VR?) to get something hopefully more affordable.. 😦
Hi Thomas
Thanks! You are absolutely correct that it would be great to see a different lens on the D800. One of my favorite photographer friends has the D800 and says that color line comes from the lens.
Michael