Canon 7D VS Nikon D800 Low Light High ISO Comparison

Faymus Media

Faymus Media

Well-known member
So I rented the D800 because I am have been waiting for a camera like this to come around. Well before purchasing I decided to rent and get some good Nikkor glass as well. I picked up the 50mm, 85mm both 1.4g, and the 14-24mm 2.8g.

I compared the canon 7D in low light high ISO against the D800....you can guess what happened. I am definitely going to buy this camera, well worth the money! Enjoy!


 
Faymus Media said:
No problem man! I should have put that in the first post lol I was just super tired haha. But results are impressive tho!
Click to expand...
Why don't you compare it the D800 to something more in the same class and same year technology instead of comparing it to a 2 year old Canon, if low light is your only concern the Mark 3 would beat the living hell out of the D800.
 
PSA1 said:
Why don't you compare it the D800 to something more in the same class and same year technology instead of comparing it to a 2 year old Canon, if low light is your only concern the Mark 3 would beat the living hell out of the D800.
Click to expand...

I don't think so.
 
re: https://vimeo.com/40113110

"IZMO Producciones PLUS 8 months ago
The D800 needs less light to produce an image, the 800 ISO is the same 1600 ISO on 5dmIII, and it´s the first test that I see this grain on d800, too much.

Jean Daniel Villiers 8 months ago
Yes again we see 2/3 to a stop difference and many people still compare them using the same ISO. I think people forget why we use higher ISO, it is to get a certain exposure at a certain shutter speed. In the video case 1/50. Many don't understand something as basic as this."
 
The Canon 5d Mk3 is much better in very low light than the D800, especially in *extremely* low light, where it's *way* better, but from the comparison tests I've seen I prefer the look of the Nikon. In most actual conditions, which is to say not ultra-low light, it's a good deal sharper and has greater shadow detail, at least to my eye...

YMMV
 
Last edited:
There is a 2/3 stop difference, but doesn't matter if Nikon understated or Canon overstate the ISO, what matter is under extreme low light, which camera gives you a video quality that is clean and good color, no doubt for those that need low light performance, M3 is the way to go.
 
reygun said:
re: https://vimeo.com/40113110

"IZMO Producciones PLUS 8 months ago
The D800 needs less light to produce an image, the 800 ISO is the same 1600 ISO on 5dmIII, and it´s the first test that I see this grain on d800, too much.

Jean Daniel Villiers 8 months ago
Yes again we see 2/3 to a stop difference and many people still compare them using the same ISO. I think people forget why we use higher ISO, it is to get a certain exposure at a certain shutter speed. In the video case 1/50. Many don't understand something as basic as this."
Click to expand...

Exactly. The ISO on the d800 has almost a complete stop brightness increase over the mark 3.


PSA1 said:
There is a 2/3 stop difference, but doesn't matter if Nikon understated or Canon overstate the ISO, what matter is under extreme low light, which camera gives you a video quality that is clean and good color, no doubt for those that need low light performance, M3 is the way to go.
Click to expand...

In movies possibly, not photos. I have seen many comparisons and the d800 has better detail. I would rather apply noise reduction and save detail than have banding and softness.

The mark 3 exhibits banding in low light I high iso with a softer image. If you shoot in lowlight why not get a D4 it is the low light king or a 1DX? It has great dynamic range and sharpness.

i quite frankly disagree with this whole shooting in the dark the mark 3 is better than the d800 crap. If your shooting in the dark you should be a pro and own the 1dx.

Since I don't shoot in low light there is absolutely nothing beneficial for me to own the 5d mark III. In a well lit situation the 36 mp triumphs all. Case closed.
 
In the end the Canon 5dmark 3 might win but it is far from the general opinion that it is much better. For one thing the D800 is at least 2/3 to 1 stop brighter than the 5dmark 3. So you cannot compare them at the same ISO. It is not because a manufacturer put a badge on a camera, that it must be true. Which of the DSLR produce true 1080p for example. If someone has just a little knowledge about photography/video he knows it is the level of exposure that is important to compare two camera and not some rating by a manufacturer. At least when there is such a difference. You should every-time compare ISO 3200 on Nikon to at least 5400 to the Canon for example.

The second thing is resolution. Look at the test done by dxomark for sensor, they always normalize to a fix resolution of 8 megapixel. The reason is simple because in the end it is the output resolution or size that you have to compare to. For video it should be the same, the resolution should be normalised as the d800 is clearly sharper. Either you downscale the D800 to the mark3 that should reduce noise or a much more sophisticated solution is to use some noise reduction.

Look at this test done by Joe Marine from nofilmschool, in which he tried to normalise exposure and NR. You can clearly see in the test that there is a good one stop exposure difference between the D800 and 5d mark3. The second thing is the DR. In the D800 you can see much further in the shadows thus also seeing more noise. Tthe last example is the Noise reduction which make them about identical if you factor in the lost of resolution. The D800 could even be ahead because most side by side the D800 was clearly brighter. It also give a big hint that there are seroius NR going natively in the Canons because the image become so identical (very waxy like).

Kudos to Joe Marine, for going back to do another test after the first initial findings of his first test. The worst thing Nikon did was to give a D800 to philip Bloom to test. That was one of the worst test of a camera I have seen. Come on, test camera in a bar at night. How many variable like people moving around for example was there in that test. He said he was in a hurry, but you are supposed to be a dslr Guru and Nikon one of the major player brings a new line of camera and you do such a rushed job.

 
Faymus Media said:
In movies possibly, not photos. I have seen many comparisons and the d800 has better detail. I would rather apply noise reduction and save detail than have banding and softness.

The mark 3 exhibits banding in low light I high iso with a softer image. If you shoot in lowlight why not get a D4 it is the low light king or a 1DX? It has great dynamic range and sharpness.
Click to expand...

What is this banding stuff supposed to be? I had both the D800 and the 5dmarkIII and in videos, the 5dmarkIII had significantly less banding than the D800. In videos 5dmarkIII is simply, a lot better.


Since I don't shoot in low light there is absolutely nothing beneficial for me to own the 5d mark III. In a well lit situation the 36 mp triumphs all. Case closed.
Click to expand...

For stills the D800 is better for studio work and one shot landscape stuff. 5d mark III is better for event shooting. I shoot about 1000-2000 shots per event. I really liked the DR & mp of the D800 but in reality...the 5dmarkIII is no slouch either. And it's also considerably faster which is important when shooting events.
 
Last edited:
Danielvilliers said:
Kudos to Joe Marine, for going back to do another test after the first initial findings of his first test. The worst thing Nikon did was to give a D800 to philip Bloom to test.
Click to expand...

Mmm no. The D800 falls apart because it's not just noise that happens in lowlight. The colors go totally wonky in the shadows, especially when you go over 1600 ISO. Philip Blooms test shows the reality and those with both cams (and who actually shoot stuff) should prefer the 5dmarkIII.

Now you can say that the D800 has a smidgeon more dynamic range. I could also argue against this because the skin tones tend to go grey with the D800 quite a bit faster than with the 5d and the shadow detail is not really...all that much. I had to do a lot of correction on the D800 which we used to shoot along the 5dmkIII, 7d and nex-5n because the skintones were always quite odd and de-saturated with the D800.

Also the 720p is almost useless on the D800, the aliasing is that huge. On the 5dmarkIII, it's quite acceptable.

I also don't see the sharpness difference. The d800 has aliasing like the 5dmarkII so it will look "sharper" on edges but that's just that. Aliasing. The 5dmarkII looks also "sharper" but in reality they are not. Especially once you start applying sharpening. It's quite amazing how much the 5dmkIII can take it compared to cams that alias.

Also I really hate the way shooting stills and movies at the same time was done on the D800. That was maybe the biggest dealbreaker (along with the really low fps liveview zoom).

One thing that skews that nofilmschoolvideo is the Ki Pro. People don't use those recorders in real life shoots. There is a reason for it. They do tests with them. They complain if HDMI is not clean. But one thing people don't do is actually use those recorders with HDSLR's in a real production. Why? It's not usable. There is no backup to card at the same time. If the connector falls out (and HDMI tends to do that), the shot is ruined. It's quite bulky and not very friendly to use. Etc.
 
Last edited:
maarek said:
Mmm no. The D800 falls apart because it's not just noise that happens in lowlight. The colors go totally wonky in the shadows, especially when you go over 1600 ISO. Philip Blooms test shows the reality and those with both cams (and who actually shoot stuff) should prefer the 5dmarkIII.

Now you can say that the D800 has a smidgeon more dynamic range. I could also argue against this because the skin tones tend to go grey with the D800 quite a bit faster than with the 5d and the shadow detail is not really...all that much. I had to do a lot of correction on the D800 which we used to shoot along the 5dmkIII, 7d and nex-5n because the skintones were always quite odd and de-saturated with the D800.

Also the 720p is almost useless on the D800, the aliasing is that huge. On the 5dmarkIII, it's quite acceptable.

I also don't see the sharpness difference. The d800 has aliasing like the 5dmarkII so it will look "sharper" on edges but that's just that. Aliasing. The 5dmarkII looks also "sharper" but in reality they are not. Especially once you start applying sharpening. It's quite amazing how much the 5dmkIII can take it compared to cams that alias.

Also I really hate the way shooting stills and movies at the same time was done on the D800. That was maybe the biggest dealbreaker (along with the really low fps liveview zoom).

One thing that skews that nofilmschoolvideo is the Ki Pro. People don't use those recorders in real life shoots. There is a reason for it.
Click to expand...

I like comments like that. Everytime one camera trouces the other one in one domain and on the other where the other one is better it is just... "shadow detail is not really...all that much". I see a huge difference in the example of the side by side video above done by Joe Marine. The second thing I see in the test is that the resolution is not just aliasing, or my I should retire after 20 years as a graphic designer/ web designer and now professional photographer. The level of sharpening some people are claiming would magically make the 5dmark3 a true 1080p camera are useless in many cases. Do that on human faces and see the halos etc.

And I don't see the colors going wonky at all. Many people using Canon complaining about the skin tone, they just don't understand DR. What the Nikon exibit is much better DR so a much flatter look. Just apply a dam curve and a little saturation. I think 99% of the Canon Fanboys would puke at the Alexa with it Arri-log images. In fact I see the Arri Alexa very close to the Nikon neutral look. The last thing you have a ton of picture profile and you can even create your own. It is just amateurs that can't understand those simple things that images can be processed and I prefer a flat images that will preserve my highlight and shadows that I can tweak to my desire afterward. That a push button one that might look good in normal lighting and fall apart in chalenging one.

If you have work with o much camera, you should enlighten us with some footage so that we can see what you are sharing with us. At least the test above, people can make up there own mind. As for the real world test of philip bloom. Tell me how can a crowded bar at night be a proper test ground.

So tester star to do the test with camera one.

- 10 guys come along and are curious about it. They block a good part of the scene.

- 15 minute later tester instal camera 2. Just at this moment 5 pretty ladies comes into the bar.

- 5 guys bored by a guy with a camera move along to meet the 5 ladies at the Bar. So much more light in the scene.

- 30 minutes later tester changes to camera 3.

- Kobe Bryant dunks on the head of Kevin garnet on the TV bar. No one around the camera guy.

So this is a conclusive real world test. Because me tester of this world I give you real world tesssssssssssssssssst. You know real world meaning real man test. Not like those wimpy in there controlled environment in there controlled condition with there miserable life. It is the only test photo video where I did not see the 2/3 to 1 stop difference.
 
Danielvilliers said:
I like comments like that. Everytime one camera trouces the other one in one domain and on the other where the other one is better it is just... "shadow detail is not really...all that much". I see a huge difference in the example of the side by side video above done by Joe Marine. The second thing I see in the test is that the resolution is not just aliasing, or my I should retire after 20 years as a graphic designer/ web designer and now professional photographer.
Click to expand...

Well my profession completely trounces your 20 years of completely unrelated stuff as I actually do edit these cameras for a living.

The level of sharpening some people are claiming would magically make the 5dmark3 a true 1080p camera are useless in many cases. Do that on human faces and see the halos etc.
Click to expand...

No, it's not on the level of a BMC or GH2. But neither is the D800. Double standards? A well done slight unsharp mask actually does wonders without halos. It seems to me that a lot of cams actually apply a significant (huge and huge) amounts of in cam sharpening. For example, the Sony RX100.

And I don't see the colors going wonky at all. Many people using Canon complaining about the skin tone, they just don't understand DR. What the Nikon exibit is much better DR so a much flatter look. Just apply a dam curve and a little saturation.
Click to expand...

We shot this video with four cams, the D800, 5dmarkIII, sony nex-5n and the canon 7d.

https://www.youtube.com/watch?v=6FQUMngIP6I

D800 and 5dmarkIII were the two best ones. Still the D800 wasn't as good as the 5d. In the end though, it's very difficult to spot any of those cams from the finished version.

The colors wen't wonky this way: Guy standing in daylight. Skin color is grey with white balance set to daylight. That's just... wrong. Maybe there was another setting somewhere (we had two D800's for two days of shooting). A lot action stuff was quite useless with the 720p on the D800. Same as the 7d. But on the 5d they were quite good. Even the nex-5n started aliasing quite a bit when set in 1080p50. Also the compression (we did a bunch of explosions) was just hands down way better in the 5dmarkIII ALL-I than on any of the other cams.

You can't really create your own profile with the D800 as it really starts banding. It doesn't use the original raw-stuff but it actually FIRST converts everything to 8-bit, then applies some curves. So you will get a BUNCH of banding when you start messing with the profiles of the D800.

It just seems very, very odd to me to praise the D800 as a better cam as the 5dmarkIII. It comes quite close but in real-world usage it's just not there. Maybe the D900 or whatever will do it, this is just - to me - basically a 5dmarkII with better lowlight and stills.
 
Last edited:
You must log in or register to reply here.
Back
Top