Canon or Nikon...

[Fishnose]

Yes something is going on. Either canon overstates the ISO true value or Nikon understates it. Thus although the canon has some advantage, at equivalent exposure the Nikon is at 1 to two stops brighter. Which means the canon needs to boost 2 stops to match it. So ISO 800 on Nikon is about 3200 on the 5d3 to get the same image. The real advantage of the mk3 is hard to measure then.

However the Mk 3 detail is so mushy, it does get killed for daytime shooting :|
nikon 422 hdmi out just blows it out of the water.

Not only that - I just looked at it on a really big, very well color-balanced monitor. The 5D3 is positively pink in comparison to the D800 (most obvious on the cards to the right). Either the guys doing the test missed setting WB properly, or the lighting changed in between, or the 5D3 is seriously pink in these conditions - which I kind of doubt.

 

[preppyak]

Not only that - I just looked at it on a really big, very well color-balanced monitor. The 5D3 is positively pink in comparison to the D800 (most obvious on the cards to the right). Either the guys doing the test missed setting WB properly, or the lighting changed in between, or the 5D3 is seriously pink in these conditions - which I kind of doubt.

I'm seeing the same thing, the Nikon has a slight yellow tinge (barely noticeable) and the Canon is really pink. And I agree about NR being on in the Canon cameras, I've seen 5dii video at those higher iso levels, and it isn't that clean.

That said, the mechanics of the cameras does make sense of Canon to be much better in low-light video. The video size, etc is better for downsampling

 

[psolberg]

CANON ALL THE WAY. . .

http://fstoppers.com/video-canon-5dm3-5dm2-and-nikon-d800-low-light-video-test

BEST ALL ROUND camera this side of the universe. . . heck I can't see ANY 95 in the Nikon example . . . . only in the Canon.

That test demonstrates that a certain N camera got Destroyed 800 times over and over!
but yeah some would argue it's *just* the low light video part..

Really?
How about trying what I tried, since it's obvious the light levels are very different in the 3 videos stacked on top of each other.

Run the video twice side by side, stop one video at ISO 800 and look at the D800. Let the other video run until the 5D3 comes up to the same level of light as the D800 has in the paused video.
What do you see? You need at least a full ISO stop to get the same amount of light.
Which means the D800 is being disadvantaged in this 'test' by a whole stop at least when you get to higher ISO levels, and of course you'll have more noise etc. Makes the 'test' rather pointless.
Or it means the 5D cameras is running way lower ISOs than they claim to be running.

Also, it appears that Canon has NR on and Nikjon has NR off. So the comparison is even more pointless.
Pity they couldn't do it properly when they did it at all. As the guy says who did it, "only had the cameras for a short time". Maybe he just doesn't know how to set them up exactly the same. Menu systems can be such a drag....

Yes something is going on. Either canon overstates the ISO true value or Nikon understates it. Thus although the canon has some advantage, at equivalent exposure the Nikon is at 1 to two stops brighter. Which means the canon needs to boost 2 stops to match it. So ISO 800 on Nikon is about 3200 on the 5d3 to get the same image. The real advantage of the mk3 is hard to measure then.

However the Mk 3 detail is so mushy, it does get killed for daytime shooting :|
nikon 422 hdmi out just blows it out of the water.

For what it's worth, according to the almighty DxO ISO sensitivity, for Nikon D800 and Canon 5d3, the 5d3 is closer to the norm than the D800 is...

http://www.dxomark.com/index.php/Cameras/Compare-Camera-Sensors/Compare-cameras-side-by-side/%28appareil1%29/795|0/%28brand%29/Canon/%28appareil2%29/792|0/%28brand2%29/Nikon

Cant quite figure out why it's so bright unless something wasn't quite equal.

DXO mark measures still images NOT video. It is clear that setting both cameras at ISO X, the nikon is brighter meaning one of the cameras is NOT reporting the ISO sensitivity correctly. As to which one it is, one would have to take a control camera, like a 5DmkII or D3s and repeat the test. I suspect nikon goofed off and what they report as ISO800 is really much higher...which is silly of them because this makes them look worse than they actually are on these candlelight tests. It can also seriously screw with your exposure if you use an external meter.

 

[awinphoto]

CANON ALL THE WAY. . .

http://fstoppers.com/video-canon-5dm3-5dm2-and-nikon-d800-low-light-video-test

BEST ALL ROUND camera this side of the universe. . . heck I can't see ANY 95 in the Nikon example . . . . only in the Canon.

That test demonstrates that a certain N camera got Destroyed 800 times over and over!
but yeah some would argue it's *just* the low light video part..

Really?
How about trying what I tried, since it's obvious the light levels are very different in the 3 videos stacked on top of each other.

Run the video twice side by side, stop one video at ISO 800 and look at the D800. Let the other video run until the 5D3 comes up to the same level of light as the D800 has in the paused video.
What do you see? You need at least a full ISO stop to get the same amount of light.
Which means the D800 is being disadvantaged in this 'test' by a whole stop at least when you get to higher ISO levels, and of course you'll have more noise etc. Makes the 'test' rather pointless.
Or it means the 5D cameras is running way lower ISOs than they claim to be running.

Also, it appears that Canon has NR on and Nikjon has NR off. So the comparison is even more pointless.
Pity they couldn't do it properly when they did it at all. As the guy says who did it, "only had the cameras for a short time". Maybe he just doesn't know how to set them up exactly the same. Menu systems can be such a drag....

Yes something is going on. Either canon overstates the ISO true value or Nikon understates it. Thus although the canon has some advantage, at equivalent exposure the Nikon is at 1 to two stops brighter. Which means the canon needs to boost 2 stops to match it. So ISO 800 on Nikon is about 3200 on the 5d3 to get the same image. The real advantage of the mk3 is hard to measure then.

However the Mk 3 detail is so mushy, it does get killed for daytime shooting :|
nikon 422 hdmi out just blows it out of the water.

For what it's worth, according to the almighty DxO ISO sensitivity, for Nikon D800 and Canon 5d3, the 5d3 is closer to the norm than the D800 is...

http://www.dxomark.com/index.php/Cameras/Compare-Camera-Sensors/Compare-cameras-side-by-side/%28appareil1%29/795|0/%28brand%29/Canon/%28appareil2%29/792|0/%28brand2%29/Nikon

Cant quite figure out why it's so bright unless something wasn't quite equal.

DXO mark measures still images NOT video. It is clear that setting both cameras at ISO X, the nikon is brighter meaning one of the cameras is NOT reporting the ISO sensitivity correctly. As to which one it is, one would have to take a control camera, like a 5DmkII or D3s and repeat the test. I suspect nikon goofed off and what they report as ISO800 is really much higher...which is silly of them because this makes them look worse than they actually are on these candlelight tests. It can also seriously screw with your exposure if you use an external meter.

The 5d2 was used in that test and looks similar to the 5d3 in terms of brightness but noisier... Of course if the 5d2 didn't report the ISO the same either, as you suggest may be happening in part by one, both, or all cameras, then we may never know unless we base them off of a still sample which should be pretty darn close within 1/3 of a stop. I dont know why they wouldn't use the still's ISO sensitivity levels, but i'm not an engineer or video guy so doesn't bug me either way.