No doubt if your a landscape photographer the D800 looks like a good choice. Since I don't shoot landscapes it wouldn't be my choice. It does look like Nikon is going to have to make some improvements to their lens lineup to handle this beast.
As for the D800 liveview I read about that on a video forum.
Some guy discovered how video is created from the D800. It reads one out of every three lines going across fully and then ends up with something slightly great than 1920x1080p and it then downsamples that in software to 1920x1080p. So it only uses 1/3 of the photosites and thus the video is like 1.5+ stops worse SNR than 5D3 video.
Anyway back to the point, he also looked into liveview and was horrified that instead of using the chips built in read from portion of sensor ability as the D7000 exmor did for liveview that it seems to for some reason hook into the video code instead and instead of sending a perfect zoomed look at the sensor and debayered and processed it gets some sort of messy partially line skipped mess from teh video system and shows that for liveview instead. He said that you need to try to focus on getting detail of vertical orientation looking sharp and trying to ignore worrying about detail going the other direction (actually that sounds reversed but hmmm).
Anyway it seems that one way or another it sending to some partially line skipped mess from teh video subsystem instead of doing the sensible thing they did with the D7000 (or that canon does with all of their cameras).
It is the one knock that landscape people have been giving the D800.
Alek Komarnits wrote:
I'm blown away by the noise difference when boosting the shadows at ISO100. While I'm a Canon shooter, I'm not "religious" about it ... but can't help but think something weird is going on for that much of a difference - is LR itself screwing up by chance(?) BTW, I didn't see any mention in the text of what you used to boost the shadows.
If you look closely at above 5d3 image you can see data has been contaminated before pushing any shadows. This can be illustrated by shooting 5D series and D800 side by side at low iso high DR scenes. So banding and shadow noise are strictly camera DR properties and they has little to do with LR or comparable image editor. Some has blamed on under exposed
Great photos and review, Fred! It confirms much of what has been postulated.
Bluntly speaking, the EXMOR sensor blows away the Canon at most low ISO settings, which is where most landscape photographers live. The usable detail is MUCH higher, and the ability to recover shadows is in a different league entirely.
So, it appears that if you're a landscape photographer and you make large prints the Nikon utterly destroys our beloved 5D3. And it is less expensive. But this only rings true if you can find the right Nikon glass for the job.
And therein lies the rub. Ultimately, when comparing bodies, once must consider available glass, "systems" and kits. And that's where any kind of choice becomes imperfect, and needs to be reflective of individual needs.
The 5D3 is still a VERY GOOD general purpose camera, but Canon would probably not be able to get away with $3500 for this body (even with the great AF) unless they had the great overall "system" (e.g. lenses) to keep loyal users loyal. It will be interesting to see what, if anything, Canon does with its sensor designs in the coming months to answer this most recent challenge from Nikon.
If one could only combine the best of both worlds...
It was shocking how much difference there was when lifting the shadows. I mean I knew there would be a difference, but man, that Exmor is impressive.
I also was very interested to hear you verify that the 24TSEII is better than the 24PC-E. (I don't suppose you did any controlled tests shooting the TSE vs the PC-E?) That, plus plus the fact that there is no 17TSE equivalent from the Nikon, plus the live view issue with the D800...those things definitely reaffirm my choice to stick with the 5DII.
skibum5 wrote:
Anyway back to the point, he also looked into liveview and was horrified that instead of using the chips built in read from portion of sensor ability as the D7000 exmor did for liveview that it seems to for some reason hook into the video code instead and instead of sending a perfect zoomed look at the sensor and debayered and processed it gets some sort of messy partially line skipped mess from teh video system and shows that for liveview instead. He said that you need to try to focus on getting detail of vertical orientation looking sharp and trying to ignore worrying about detail going the other direction (actually that sounds reversed but hmmm).
Anyway it seems that one way or another it sending to some partially line skipped mess from teh video subsystem instead of doing the sensible thing they did with the D7000 (or that canon does with all of their cameras).
It is the one knock that landscape people have been giving the D800....Show more →
It's not just a slight knock. For me, when I photograph landscapes, I heavily rely on live view to ensure proper focusing especially since modern cameras have megapixel sensors. In this regard, the D800 is an absolute no-go for me.
So, thanks to Fred for pointing out this Achille's heel. Maybe this is why the D800 is US$500 cheaper in the US?
thw2 wrote:
It's not just a slight knock. For me, when I photograph landscapes, I heavily rely on live view to ensure proper focusing especially since modern cameras have megapixel sensors. In this regard, the D800 is an absolute no-go for me.
So, thanks to Fred for pointing out this Achille's heel. Maybe this is why the D800 is US$500 cheaper in the US?
Alek Komarnits wrote:
Holy better than Ansel Adams' pictures Batman! ;-)
As others have said, gorgeous pictures.
I'm blown away by the noise difference when boosting the shadows at ISO100. While I'm a Canon shooter, I'm not "religious" about it ... but can't help but think something weird is going on for that much of a difference - is LR itself screwing up by chance(?) BTW, I didn't see any mention in the text of what you used to boost the shadows.
Nice Moire comparison - be interesting to see how the Nikon800E does with this test.
The big "gotcha" on the Nikon was the Live View ... but can't help but think that might be fixable in firmware perhaps. And once they refresh their lens lineup, your writeup suggests that Canon just gets clobbered in the Landscape department.
Yes, Nikon needs to address the LiveView issue they have. If it's a hardware limitation
in their image processing chip we will have to wait until the D900, but if it's a firmware
issue, I expect they will fix it within a few months. Otherwise, it seems like a landscaper's
dream camera. Slap a Zeiss 21/2.8 on it or a 50 MakroPlanar or other great landscape glass
and it should be wonderful.
For many applications the Canon should still be great. Shadow detail appears to be
it's big limitation. If there are major patent issues facing Canon there may not be
any solution in the next few years.
cameron12x wrote:
Great photos and review, Fred! It confirms much of what has been postulated.
Bluntly speaking, the EXMOR sensor blows away the Canon at most low ISO settings, which is where most landscape photographers live. The usable detail is MUCH higher, and the ability to recover shadows is in a different league entirely.
So, it appears that if you're a landscape photographer and you make large prints the Nikon utterly destroys our beloved 5D3. And it is less expensive. But this only rings true if you can find the right Nikon glass for the job.
And therein lies the rub. Ultimately, when comparing bodies, once must consider available glass, "systems" and kits. And that's where any kind of choice becomes imperfect, and needs to be reflective of individual needs.
The 5D3 is still a VERY GOOD general purpose camera, but Canon would probably not be able to get away with $3500 for this body (even with the great AF) unless they had the great overall "system" (e.g. lenses) to keep loyal users loyal. It will be interesting to see what, if anything, Canon does with its sensor designs in the coming months to answer this most recent challenge from Nikon.
If one could only combine the best of both worlds... ...Show more →
It is a shame they didn't give us at least the D4 low ISO DR that doesn't need patents. 5D3 would really be complete the (other than MP, but if I had to I could live without that one thing until the next round).
5d3's last gen sensor isnt cutting it in direct comparison (though of course can still produce great images).
D800's live view and built in flash which interfere with the pc-e is a pain.
I was talking to a nikon user this morning, who had just sold his 14-24 (party to fund a d800) but mostly because it catches the sun and flares at every opportunity and he didnt like it...
I want to use canon glass on a 5d3 with the D800e sensor
RobDickinson wrote:
It seems to me neither camera is perfect.
The D800 sensor in the 5d3 is what we all want.
5d3's last gen sensor isnt cutting it in direct comparison (though of course can still produce great images).
D800's live view and built in flash which interfere with the pc-e is a pain.
I was talking to a nikon user this morning, who had just sold his 14-24 (party to fund a d800) but mostly because it catches the sun and flares at every opportunity and he didnt like it...
I want to use canon glass on a 5d3 with the D800e sensor
D800 also has some video issues with moire and noise.
Super write up Fred, well done. I wish I could just pop down to Yosemite for a shoot!
The D800 shadows are ungodly - it's like we are comparing a crappy P&S and DSLR. Too bad then they have crippled the LCD with interpolated crap when zooming in and they are seriously out-gunned on glass unless you stick to Zeiss or the 14-24 for landscape. We need that D800 sensor in the 5D IIIs and then we have the best of everything.
I hope Nikon wake up to themselves on the LV issue. Can it be fixed in fw or does it need a hardware fix? This and the lack of TS-E's to match Canon are a deal breaker, even though I could use Zeiss glass.
It was shocking how much difference there was when lifting the shadows. I mean I knew there would be a difference, but man, that Exmor is impressive.
I also was very interested to hear you verify that the 24TSEII is better than the 24PC-E. (I don't suppose you did any controlled tests shooting the TSE vs the PC-E?) That, plus plus the fact that there is no 17TSE equivalent from the Nikon, plus the live view issue with the D800...those things definitely reaffirm my choice to stick with the 5DII.
+1
Thank you Fred for the very informative review and amazing images as usual!
I had a chance to shoot a friend's D800 in comparison with my 5D2 and I came to your very same conclusions: D800 sensor is just great particularly with regard to shadow recovery. Difference in resolving details is there but visibile only if you print VERY BIG.
LiveView on D800 is poorly implemented and I had a hard time in focussing in low light.
I prefer some Canon glass so switching is not so easy.
One thing I noticed though is that while the 5D2 risolved over 90% of D800's, it seems always at the limit of what could be smoothly rendered. In a sense details seem more sharpened and digital compared to the D800 smooth rendition. I was told that the 5Dmk3 has the same smooth rendering, so despite no increased resolution compared to 5Dmk2 it is less "stressed" so to speak.
Do you find the same in your test?
