Description image

5D Mark III/D800 Hands-On Part 2: Initial Impressions (D800)

Monday we talked in-depth about the 5D Mark III, and today we’ve got the D800. Nikon definitely surprised a lot of people with this one, and it’s interesting that Canon didn’t really see them coming – or they are afraid to hurt their higher end sales (which could include a possible 4K camera priced below the C300). Either way, you can’t go wrong with clean 4:2:2 HDMI out of the Nikon D800, and still photographs which rival medium-format backs costing $20,000 or more. So let’s get down to it.

Nikon D800 – The Design

The D800 continues on the traditional Nikon form, just as the 5D Mark III did with the 5D Mark II. I have to say, I always enjoyed holding Nikon cameras, but button placement isn’t nearly as friendly for my thumb as the Mark III. I can’t seem to find a good spot without resting it on some buttons, which I never like to do. The big control pad on the back could ideally be a bit lower – this might not bother someone with small hands but I like to really hold on to my camera tight with one hand and I can’t really do that with my right hand. Aside from that it is well-balanced.

Nikon hasn’t done anything too drastic compared to the D700 in terms of button placement. It’s all pretty much the same except the AF selector has been replaced with a photo/video selector and live view button. This is obviously a necessity on a camera that is taking video seriously – it’s a lot more problematic trying to shoot video without one of these selectors – and it also gives you piece of mind that you’re in the right mode when you need to be.

For all of those Canon shooters who are upset that the magnify/reduce has moved (my problem is the combination into one button), Nikon still has theirs on the left side, just like on the D700. So it’s really more a matter of habit, not of functionality. I have no problem with these functions being on the left side of the camera, because you have to operate the camera with two hands most of the time anyway. If you’re handheld – you’re other hand has to go somewhere – and if it’s underneath the lens, the magnify and reduce buttons are in a great place for your thumb. ISO is still on the top left, in contrast to the 5D Mark III which is on the top right. Again, most of these problems that people have are related to habit – which shouldn’t take too long to break if you’re shooting enough.

One hugely positive difference over the 5D Mark III are the side ports. Nikon’s port cover is much better designed, and it opens in a more sensible direction for keeping cables attached. It’s also quite a bit easier to open than Canon’s – I have not struggled once trying to get it open. Nikon also upgraded the camera with a headphone jack, and they are looking towards the future and have placed a USB 3.0 port in the camera. While not really an issue for video shooters – it’s a welcome addition, and proof that Nikon is pushing the technology a bit less conservatively than Canon.

Nikon D800 – Operation

Having never used the menu on the D700, this just feels like a typical menu that you’d find on a Nikon DSLR – or any DSLR for that matter. Canon’s menu is much better – and it’s also prettier to look at (even though that doesn’t mean anything). It’s really as good as it needs to be – but you’ve got to be aware of a couple settings that could affect you greatly.

One huge difference between the 5D Mark III and the Nikon D800, which hasn’t gotten much attention, is the fact that the D800 in live view does not record the full sensor cropped to 16:9, like the Mark III does. Nikon has an additional crop in live view, rather than just cutting off the top and bottom of the frame to get to 16:9. The Nikon camera actually just takes a more centered crop, at 32.8 x 18.4 and then uses that to record video. Effectively this means that the field of view is slightly reduced compared to the Mark III. In practice it’s not really a big deal, and it’s effectively a 1.1 crop – this has been confirmed by placing both the Nikon and Camera cameras at the same spot and comparing the focal length with the same lens attached. It’s just something to keep in mind if you are ever mixing the Nikon with any of Canon’s full frame cameras.

Shooting video actually feels a bit snappier than the 5D Mark III, as the mirror seems to be much faster in snapping up into place. Overlays can be removed on the HDMI out, and Nikon has been able to keep the LCD on even when a monitor is attached. This is essential for DSLRs in 2012! Canon needed to figure out a way to keep the LCD working even when a monitor is attached – but thankfully Nikon has come through – and we can get a reference picture on the LCD with the full output coming from the HDMI to an external monitor. Speaking of HDMI, this is where things get a little tricky.

Nikon’s only HD options for HDMI out in the camera are 720p and 1080i – which seems rather odd for a company that is touting it’s ability to record the clean output. Nowhere in Nikon’s manual does it describe actually recording that output (another glaring omission). When the camera is set to 1080i, it really is 1080i – the Ki Pro I had attached saw the output of the camera as 1080i 29.97. This absolutely was baffling to me. Setting the HDMI out to 720p actually gave 1080p 23.98 – but this is not the correct setting. Buried deep in the internet there is a thread with a response from Jon Thorn of AJA:

The Nikon D4, with no QXD or CompactFlash card inserted, can be configured to output clean HDMI 1080p 23.98. The AJA Ki Pro Mini, beginning with version 3.0 firmware, which was recently released, will correctly interpret the D4′s output and will report and record 1080p 23.98. A few things to keep in mind: the camera must be set to Auto for the HDMI output format and Live View must be activated for the “handshake” to happen properly. (It is doubtful that all external video recorders will behave in the same way as the “handshake” requires some work to achieve.) Also, it is good to be aware that when entering Menu on the camera, be aware that 1080i will be output. Another thing to keep in mind, noted in some very brief initial testing (so not completely quantified), is that it appears that Live View audio may not be delayed to match Live View video so some compensation in post may be needed for a/v sync when using an external video recorder. (If double-system sound is being recorded, likely this isn’t an issue since there are software tools that provide compensation and “auto” a/v sync that are available.) Additionally, this has only been very briefly tested with the Nikon D4, not the Nikon D800.

Also from that thread is confirmation that the PIX recorders will do 1080p when the cards are removed. The Atomos Ninja should also be able to do the same, but this little bit of info seems to throw a wrench in the mix:

Ninja’s 2.1 firmware update includes the following:

Nikon D4/D800 support – added custom handling for Nikon 1080p24 pulldown mode.

NOTE: When connected to a Ninja, the Nikon D4/D800 HDMI output must be set to 1080i when the movie setting is 1080p30, 1080p25 or 1080p24. The Nikon D4/D800 HDMI output must be set to 720p when the movie setting is 720p60 or 720p50 (See page 281 in the Nikon manual: “HDMI Options”. Also note on manual page 74: “Movie Settings” – the *actual* frame rates used by the camera, for example, the actual frame rate of the camera when set to 1080p24 is 1080p23.976.)

Shouldn’t the Ninja be able to record 1080p with the cards removed by setting to Auto? I would think so, but it’s quite possible that Nikon is doing something funny that must be recognized by the recorder when the camera is set to Auto. This is going to take a bit more investigation to confirm, but it’s unfortunate that Nikon has made this so complicated – and failed to mention anything about recording the HDMI in the manual.

You can use the shutter release to start movie recordings, but it must be set in the custom settings menu, under g4: Assign Shutter Button. The big difference here that certainly gives the Mark III an advantage, is that the D800 manual audio settings cannot be changed once you press record. Nikon has historically been difficult about manual settings while recording video, but thankfully you can change all of them while recording except for the audio.

Regardless, the Nikon has been a pleasure shooting video – and from what I’ve seen so far, I think the D800 is actually outputting more resolution than the 5D Mark III (D800 was recorded through the AJA and the Canon recorded ALL-I). How much more is hard to say, but I saw a clear difference between the two. I’m not posting anything yet because I want to make sure that the test is rock-solid and there are no possible reasons for the discrepancy – which will probably come early next week when I have done the full test. It’s definitely true that the D800 has moire – much more than the 5D Mark III, but how will that play out in reality, again, will take more testing.

Things are certainly getting interesting…

Link: Part 1: 5D Mark III and D800 Hands-On Review: Initial Impressions (Mark III)


We’re all here for the same reason: to better ourselves as writers, directors, cinematographers, producers, photographers... whatever our creative pursuit. Criticism is valuable as long as it is constructive, but personal attacks are grounds for deletion; you don't have to agree with us to learn something. We’re all here to help each other, so thank you for adding to the conversation!

Description image 33 COMMENTS

  • Considering the processing power needed to output uncompressed video, IN THEORY a firmware update could fix this. Not that Nikon will bother mind you, but I’m positive some savy hackers could turn this into a monster of a camera.

  • My hunch is: both the D800 and MKIII are lineskipping/binning but the D800 has a less powerful LP filter over the sensor, similar to the MKII. However, Nikkon improved the skipping better than the MKII, therefore it has slightly more resolution. However, the MKIII with skipping and a more powerful filter over the sensor produces reduced moire but softer images with less resolution.

    Remember, moire tricked our eyes into thinking more resolution was there on the 5dmkII. We liked it until it reared it’s head when we didn’t want to. But honestly, moire was creating our images anyway, just sometimes creating a pattern we didn’t like.

    Full sensor read out is the only way to go. Let’s see: Gh2, C300, RED and most pro video camera’s do, but I know of no company having a 35mm sensor (photography sized) doing a full read out. DO you?

    • This has a lot to do with processing power I would think (and maybe some heat), keep in mind how much smaller the GH2 sensor is compared to full frame 35mm. Panasonic also smartly developed the sensor to be able to do a type of A/D conversion right on the sensor. This allows downscaling to happen a lot easier. The GH2 is obviously the anomaly here, but it’s due to sensor development that it’s able to get so much resolution (don’t forget that Panasonic also crippled the internal recording which hurt resolution – it took some hackers to be able to get the full potential out of it). RED does not actually downscale their sensor for any of the recording modes, only for the HDMI and HD-SDI out (heat and processing reasons). Obviously the RED and C300 are in a league of their own in terms of price, so again I think it’s hard to compare.

      • That is my point Joe: the full sensor has over 4x the area as the Gh2. And when we look at RED doing a full sensor read out of the C300 from smaller sensors, and the processing required, I personally don’t believe the power is here yet with heat and power consumption issues, to do a full sensor read out and scale down on the large format DSLR’s.

        Therefore, resolution wise, I don’t see 720P let alone 1080P.

        Thus, moire is our friend to a degree. When you take it completely out, the image gets soft.

        It is really a catch-22 for us right now.

      • Also, what does the company think? Do you think Canon or Nikon would really spend the time to develop the processing power to do so from such a large sensor? Sure the c300 does full resolution, but that is a smaller sensor.

        Then we get into what we see with Canon: fear of cannibalizing the pro sales.

        Honestly, what needs to happen is people push for false advertising on these camera’s. They are not 720P or 1080P, they are SD with 720 and 1080 frame sizes.

        All they would have to do to differentiate the price points is say: SD 720P, SD 1080P for their DSLR’s, HD 720P HD 1080P for pro stuff and then 4K resolution for their new forthcoming cameras.

        I am fairly confident a court could bring false advertising to Canon. We see with the tests they are not HD cameras.

        • If you really want to be picky all of these companies have been doing stuff like this forever. The 5D Mark II is around 720p – actual resolution – so if you just downscaled in post you’ll get the equivalent of every other 720p camera. Think about this – the Panasonic HVX-200 had a 960 x 540 sensor, and then upscaled to 720 – and in the later models upscaled to 1080! JVC did similar things, they were 720p for a long time, and with the introduction of the HM700 they used the same sensor but upscaled to 1080. Canon’s XH-A1 was the closest to 1080, with a 1440 x 1080 sensor.

          Those are just stated resolutions, that doesn’t take into account actual resolved detail.

          It’s honestly not worth getting that worked up about, if you know what you’re doing (as it seems you do) then you’ll know the limitations.

          But to answer the question, Nikon might be the only one to consider spending the money because they don’t have any other higher-end cameras to cannibalize.

          • Very true. Well, I love my 5dmkII but need another body for weddings and definitely want to also use it for film since that is my real passion but I’m simply an amateur.

            With a bunch of Canon glass, what would you buy? Gh2 or D800? Is auto focus disabled on both when converting to Canon glass? Oh, and if I am getting off topic, sorry :)

            I’d say that after waiting 3 years in curious anticipation for the mkIII (while using my mkII a ton of course), I’m at a spot of considering a new brand. Never been hear before. It is new. It is scary. I don’t want to buy another set of lenses!

            • The D800 will not be able to use any Canon lenses no matter what, but the GH2 can at least manually adapt them. No autofocus adapters have been made that I know of yet that can work with the GH2, I know Birger was working on one for the AF100 that could autofocus but it hasn’t been released yet. I’m not sure if it would work on the GH2 – from the looks of it I wouldn’t think so because it seems like they are creating a whole new mount.

              I think if you’ve got a lot of Canon glass, the options are limited – I mean for low-light with weddings, the Mark III would be really fantastic – and the autofocus on low-light is supposedly improved (haven’t tested this too much yet). But I can probably make a much better assessment after I throw the gauntlet at both cameras. Then you’d also be able to see what are the positives and negatives for your particular usage.

          • @Marine
            So, you would say that for one who wants to put a movie to a 2k cinema theather, a 720p camera would be enough, as the image can be gonflated in post to either 1080p or 2k? Interesting point. Then, all this talk about camera resolution would be just marketing.

            • I didn’t exactly say that, I actually think there’s a big difference between a 720p movie or a 4K movie when displayed in a 2K theater. How well it can blow up really depends on how clean the image is, and how good the compression is. One advantage some of the high-end 720p cameras have, like the Varicam, is that they have much cleaner compression options and don’t suffer from image artifacts in the same way as DSLRs. I watched a movie that was shot on an EX1 in 1080p and then a movie shot on RED in the same theater and I could immediately tell the difference even though it wasn’t a 4K projector. I was actually really saying that for most uses – like the web, or TV, it’s not as big of a deal.

          • Thanks Joe for your help and write-up. As far as 720P for movies, well the 5dmkII does scale quite nicely in the theater, we know that, George Lucas knows that, Philip Bloom knows that and Shane Hulburt knows it.

            I saw Act of Valor twice, up close on huge screens. The mkII was soft many times, OOF many times but still looked great. I saw some doctored moire shots. But here is the thing: I grilled to the point of aggravation the two groups of regular movie goers that are my friends (until I bug em with tech talk too much).

            NONE of them commented about soft shots, OOF or moire. They commented on the acting (wasn’t too good from the Seals of course), commented on the story, commented on the first person VOF and a few noticed the shallow depth of field.

            Throw 4k in some scenes, ya I’d notice. Movie goers: nope.

            • Of course, but we also know the average moviegoer will watch highly compressed standard definition Netflix movies on a 40″ HDTV and be perfectly fine with it. You’re right, we as filmmakers tend to be more protective of our work – and want it to look it’s absolute best – but the average public tends to accept many more flaws than we do. I would say it scales well enough for the public on a film print, but on a 4K screen, it’s going to look soft compared to a 4K camera.

              I’ve actually found something interesting, that the more people are trained in filmmaking and understand quality, the more they notice it – so if you give someone basic knowledge about 720, 1080, 4K and actually show them the difference, it’s very possible they are going to notice it in every theater from then on. It’s like opening Pandora’s Box for some people – but I guess that’s the “magic” of the movies for people, that most don’t know any better and can just watch the film.

          • That was quite my point: from a indie-business and movie-goer perspective, a 720p camera is good enough to tell a story. If the story is sound and the actors good enough, then you have a winner that can go around the world and make you a living out of movie making. From this perspective, I would dare to say that all this talk about “buy this new camera, it has 1080, 2k, 4k, 5k, 4:4:4, +5 stops latitude, etc.” is just marketing talk that has only one end: take our hard-won money. There would be no substantial box-office gain if your movie was shot 5k or 720. And no theater or distributor would reject you movie because it was originally shot in 720; the only rejection you get os on story, movie flow and overall tech package. Correct me if I’m wrong.

            • Yes your point is right, the story matters more. But you’d never be directly dealing with a theater anyway. Whomever is distributing your film would probably have paid for a $50,000+ 35mm transfer because half the screens out there are still doing film and it’s also usually the only way to get a trailer seen at those theaters. So going forward there are a few advantages to 4K over 720p or 1080p even. A 4K digital cinema camera is going to give you a much better 35mm transfer – there’s just more color information and resolution to work with. 4K (to a much lesser extent real 1080p) is going to give you a lot of flexibility in post – for example David Fincher on Girl With the Dragon Tattoo used 4.5K and 5K RED One and Epic respectively and used a center crop 4K so that he could reframe and stabilize in post.

              So yes, you are partially right, and it’s something I try to say often, that a camera will never make a good film on its own. But as artists we always want the most flexibility we can have, and higher resolution, more latitude, and better color space all allow us to have that. But we absolutely shouldn’t kid ourselves that a distributor wouldn’t prefer we already had the least compressed, higher resolution image possible – because they are going to spend hundreds of thousands getting it to look perfect.

              On a side note, I think a camera like the Panasonic Varicam at 720p and 3CCDs has produced some of the cleanest and most beautiful images I’ve ever seen compared to some lesser “1080p” cameras. Granted it’s not a DSLR in lowlight and it’s very, very expensive (dirt cheap used now without a lens), but there is something to be said for having extra color information and not having to debayer.

          • yes I agree again Joe, great conversations. I love being technical and want sharp images, tons of color information, ect. The reality from the Zacuto tests show, resolution wise, latitude wise – film is still the winner. Most of the population knows what a movie “should look life.”

            Surprisingly, I’ve had a lot of ‘regular’ non tech friends ask me why their new HD tv is playing blu-rays wrong. When I ask, I find out they have the hz rate up high and auto frame rates, so the blu-ray and TV are playing in an upscaled, if that is the term, frame rate. So many people comment that it “looks like made for TV now instead of theatrical.”

            That surprised me because I didn’t think frame rate would be recognized more than say resolution, but I think it is. That to say, I get than at 24P and everything is good to go! lol.

            A close second is that people notice “green-screen” very well. I’ve been perplexed because me in all my “tech nerdy glory” can’t see the dirty key. I say, “it’s keyed superb.” I finally found out people notice wrong lighting, like wrong shadows or mis directed shadows, even if they don’t have the terms, and so the only term they can say is “green-screen!”

            Goes to show: light correctly, keep the source fps correct across mediums, grade cinema like, and a great story, people will be satisfied.

            Now, I am confident that on their personal HD TV’s, people are going to critique a soft shot more. Depending on how close they are to their 1080P TV, because of the theory of proximity, they are able to see more resolution than sitting further back in a theater with 4k.

  • John Jeffreys on 03.28.12 @ 1:52PM

    36.3 megapixels on the D800 would be interesting for billboard applications, or projects and art installations that require super large images. I’ll probably rent one of these when i get some spare money to fool around with. Also, could you possibly post a video comparison between this and the D800e?

    • I don’t have access to a D800e, but I would say it’s not worth it for video – the D800e is designed to be a stills only camera. Without the low pass filter, moire is going to be much worse, and as we know it’s next to impossible to get rid of in post for video (a little easier for stills). Maybe in practice it won’t be that much worse, but if you need both the regular D800 is the way to go.

  • Page 324 of the D800 manual shows how to set the shutter button to record video

    • Corrected! Thanks for that – it’s in the custom menu (and on page 324) so that’s probably why I didn’t see it.

  • When you record in pal do you happend to know if the hdmi outpit is 1080p at 25fps.
    I ask this because nikon has forgoten of pal user before, I believe they haven´t made that mistake but will like to be double shure

    • I’ll do my best to figure that out – the thought hadn’t occurred to me but obviously enough of you out there are PAL users.

  • I had the opportunity to play with the 5D MK III today and I was surprised to find that Canon has not made much of an improvement, if any, to the in camera preamp. The audio recorded in camera sounds great when run through an external preamp, but there is a notable hiss without. This means you will still need to set the recording level to 1 notch above zero and crank up the volume on your external preamp. I was excited about being able to adjust levels while recording, but this feature is useless if you have to leave the audio locked at a notch above zero.

  • Thanks folks, but I’m dizzy now.

    The specs on the cameras are NOT true 1080p, oh well, as 70% of TV commercials play fine in HD cable band on the LED/IPS set and, are being shot on HDslrs and Call to Duty yadda, yadda.

    The finer point (Joke) is the MP’s from the sensor, being routed into the complexity of 2mp effectively, are making a difference yes or no?

    My big question is about the sensor capture so perhaps that may explain the resolution being seen plus, I’m still hedging for the D4 for stage lit and pre/post dawn landscape takes in Death Valley etc. even though my new D800 is waiting for me in Culver City.

    I’ll get it soon enough but might wait until you post up your bench tests.

    As for Canon glass, a question. Why wouldn’t the adapters that Canon shooters use to attach Nikon lenses work on the D800?

    My DP pal promises to buy the D800 if I don’t like it, so I’ll hold off registration until he gets back from his RED gig in April.

    Thanks for the great stuff so far.


  • Chris Mammarelli on 03.30.12 @ 11:10AM

    I have a Mark III and recently shot with a few sketches with it. While the images seem soft, they are extremely clean and sharpen right up in post. It does not appear that canon is using any line skipping and the megapixel increase was indeed to create a 1920×1080 grid cubed. Rather than line skipping they are aliasing a 3×3 grouping of pixels into a single pixel. As Philip Bloom also notes in his review, the aliasing on this camera is quite amazing. Moire is exaggerated with line skipping, and I think the only way canon could have all but eliminated it (some of the actors wore plaid shirts with no noticeable moire) is to take this new down sampled approach. Canon did listen to their DSLR videographers and took notice of the post workflow that includes grading flat footage and other effects and created a camera that supports said workflow.

    This is not a video camera and should not be used as such. It is the equivalent of a low end digital cinema camera which requires proper adjustment in post.

    • Thank you Chris for the comments, that is really informative. Perhaps I have indeed spoken wrong with their LP filter. I hope I am wrong. Either way, it is a super clean low light camera and if the images are clean enough, a little post unsharpen mask never hurt anyone!

  • I just hope that you do a honest and thorough review. It seems that every reviewer is trying all sort of sharpening to show that camera is sharp. What they forget to show is that these level of sharpness on a human face and everybody would through up. Doing sharpening/saturation/contrast on landscape scene and human being are two different things. That is why most landscape profile from camera manufacturers are the most vivid, contrasty with sharpness at its maximum.

    The second thing is the ISO label between the two. It is either Nikon is underestimated its ISO rating or Canon overestimating its ISO values. They show at least 2/3 stop difference. Photo test show that where they expose the test 2/3 stop more to match the exposure. So it seems that ISO 3200 on the 5d3 should be compared to ISO 2000 on the Nikon. On gizmodo they did a supposed test where the reviewer was claiming victory for the 5d3 while the image was like a sop under, you could mistaken the 1600 D800 with 3200 Canon at same level of luminosity.

    I am no fanboy, but the Canon coolaid is flowing a bit too much lately, so some level headed test would go far to at least put the true facts. There are things one wins on the other like the moire/aliasing of the 5d3 seems much better than the D800.

    • Chris Mammarelli on 03.30.12 @ 6:10PM

      Do you own either camera? I’ll post some before and after face footage mid next week for all to look at. If you have the D800 I would love to set up a head-to-head.

      • No, I hope I will get one for my photography in the next weeks, but I live on the other side of the world. In fact I had given hope on Nikon video for some time for my peace of mind. Got a gh2 now hacked as my guaratee for not being disappointed in another lame implementation. I was getting my D800 for my photos primarily but the video side is getting me a bit exited. Looking at this video my hopes are getting a bit high. Very very good DR, low light and very low rolling shutter. The aliasing seem to be much lower from previous example (the shooter put sharpness down completely) and even at the low setting it is very resolute and detailed. For my use of the Canon 7d, gh2 hacked, this could be the best all round camera with perhaps the mosaic filter if the moire/aliasing is still apparent.

        • I forgot one last thing, the clean 4.2.2 hdmi out. I am used working with high bitrate intra frame from the gh2 and this would be even better because of the 422 colour information for colour grading .