Nikon D5 real-world low light, high ISO samples

I'd be very into the idea of doing camera testing at the Olympics.

I just re-checked with files from the studio comparisons: Imatest suggests that the D750 seems to be a bit MORE noisy than the D5 at 51200 ISO, albeit not by very much.

It seems like my first impression misled me. Still I think the D5 sensor is not that much better to fret about it. I'm not even sure whether the A7R2 sensor would be.

@Deliverator: These are stunning pictures from a pro photographer, no doubt. Unfortunately, the pictures don't prove much about his gear, except maybe the quality of the glass, plus autofocus speed perhaps. The thing is, when you scale down a FF sensor to 2.4 MP resolution, even the output from a 2004 Canon would look good enough in terms of sharpness and noise.

I have the A7rii and A7ii. When comparing to the A7sii I found the downsized A7rii to be as good until ISO 25600, which is my emergency high-ISO. Measurements (DXO) are in line with this. So unless you are talking about the extreme ISOs or video, I don't believe your 3-stop claim.

You really cannot be serious can you?
Do you seriously believe the only part of a camera that effects IQ is DR?
Try looking at autofocus and frame rate and you will see why not one sports photographer uses a Sony.
A Ferarri engine in a Ford Chassis is exactly what Sony deliver(sorry Sony guys).
In short while the A7 us fumbling along missing shots because it's AF is deficient and its frame rate can't match real action cameras the 1Dx and D5/D4 will continue to capture the worlds biggest action moments.
It takes more than a sensor to make a great camera.

Please noone hit me but unless something is very bad with either camer I think they will be very close in review. Also not sure if DPR will be able to finish review in time for olympics.

Typical grandiose comments about death of DSLR's..

Marike6, that's so well said. There are people who think a viewfinder is better just because it is electronic. I wonder if they have ever looked into a good, bright optical viewfinder.

actually i was talking about 1DX II,good luck

If you compare the Raw samples under the artificial light studio test, the A7s II is not as good as the D5. The D5 shows more fine detail as high ISO.

The price difference has nothing to do with IQ, it has to do with build quality, AF, battery life, ergonomics, and burst processing. That's what you are really paying for in a D5.

On the other hand, you are paying $3k for the A7sII, a camera that couldn't track a moving subject to save it's life.

The sensor in the D5 is actually designed by Nikon and not Sony this time. That information is all over and it may be why we're seeing degraded dynamic range at lower ISOs.

this is for fan boys that always talking fanaticalty
https://www.youtube.com/watch?v=9JjP7LlMZq0
even a7R II beat the D5 in low light,a7S II is still the king
sony sensors are the best thats a real

The D4s sensor was designed by Nikon as well. The difference in low ISO DR we are seeing are the result of design choices by Nikon.

Totally agree.
My frist thought watching these samples, was that I would like to participate in a proper concert shoot with my Olympus EM1 and my four-thirds 35-100 mm F2 against a Nikon D5.
I expect that the D5 would produce better results when used properly (Shutter speeds below 1/500), but not 4 times better !

If you use equivalence, i.e. same FOV, DOF, etc., of course. However, if the scene allows you to use the same aperture (easily possible with a 35 or 50/1.4, you can't overcome the deficit in having a sensor with only 1/4th the light-gathering area. Look, for the latter case at S/N ratio measurements between EM1 and D4s, there is a 6db difference at ISO 3200, for example (2stops): http://www.dxomark.com/Cameras/Compare/Side-by-side/Nikon-D4s-versus-Olympus-OM-D-E-M1-versus-Nikon-D3s___945_909_628 .

I doubt someone would claim that this is typical use scenario - I think it's rather done so in order to eliminate blur that may show at lower speeds and corner softness that may appear at wider aperture settings and thus only showcase the body's imaging capabilities.

disasterpiece, higher shutter speed means that the light was too low to require that high ISO speeds. In concerts (especially like the ones in photos) you would not need 1/1000.
I remember very well when D7000 was released, there were high ISO samples on internet and everyone was claiming that this camera would be better than D700. Then it became obvious that D700 was still better, those samples were taken under good light (say ISO6400, f8, 1/1000) to make the high ISO performance look good.
I'm not saying that D5 is bad, it's probably the best Nikon DSLR ever in high ISO.

I tend to agree with above. 1/125 won't freeze. I would argue you'd need 1/200 at least.

To echo the folks above, I'd hesitate going below 1/250 for any musical performance. 1/1000 certainly isn't outside the realm of possibility for really animated bands. The wide-angle lens will help with hand-shake at slower shutter speeds, but not necessarily with subject motion.

I often use 1/1000 when my kid is flapping about like a lunatic, it's an entirely appropriate shutter speed for an animated rock band for example.

People seem to get confused between something being impossible or impractical in the past and that same thing being unnecessary, there's a distinct difference.

If a concert photographer from the 60's could see people arguing that 1/1000 isn't necessary I wonder what they would say on the matter. Jusy because people did the best with what they had doesn't mean many amazing shots were missed.

Well, I freeze cars in motion at 1/125. Most rock musicians don't move that fast on a stage...
Sorry I'm not impressed with the pictures. I know I should, but I'm not.

I shoot a lot of jazz musicians for their publicity stills and cd covers, in live performance, and the only way that 1/125 is fast enough is if you're really good at catching the tiny still moments that occur in all musical performances. And that's for relatively laid back performers. I have had to shoot at slower speeds, and the only reason I got anything is because I knew the music, knew the musicians, so knew exactly when they'd be the most still.

ManuelVilardeMacedo, freezing cars is one part of the picture, what does the road and non moving subject matter look like?

You think you can get a moving car and the background sharp at 1/125 and It's ME who doesn't know what they are talking about?

PS: My gallery consists of screen captures from a youtube video to demonstrate the compression artifacts. Good work, detective.

Thank you for posting comments on this story here at DPR. It allowed me to realise you have no idea what you're talking about.

That all seems very subjective to me, to say motion blur has more artistic value than capturing a moment in freeze frame is purely an opinion and doesn't mean 1/1000 could have no purpose.

It doesn't matter how gifted you are, there some things you won't be able to pull off well at 1/125 even if you do want to introduce motion blur.

@fatdeeman
might be correct about controlled motion blur being subjective in terms of output preference, yet controlled blur is a display of superior photographic skills and will, in general, impress more and deliver more content, unless special effects are desirable with freezing fluids in the air or similar. Still 1/1000 is excessive. 1/320 can handle a concert no problem. That is almost 2 full stops of light. See the following link, not a special pic at all but easy for me to get and upload as I shot it 2 days ago. Instead of going lazy with a 16-35/2.8 zoom and a freezing speed of 1/250 or above, I chosen to shoot 20mm/1.6 @1/60. The clapping delivers the message far more effectively with blur and I shot with about 4 stops advantage going ISO 800 instead of ISO 12,800.
http://www.dpreview.com/galleries/7479542344/photos/3422235/dsc04870

ManuelVilardeMacedo If you held the camera still as the car went through the frame at 1/125 it would blur so you must be panning and hence the background would be blurred unless the car is stationary.

Panning a car nulls out the subject speed in relation to the camera, and the car is moving in one direction also. Taking photos of a flailing guitarist will result in portions of the subject having a much high speed relative to the camera than a car you're panning would and also parts of the subject would be moving In different directions.

This is all very basic stuff, there's plenty of nice friendly info graphics out there telling you which shutter speeds are appropriate for freezing particular motion. I remember our class had to make our own info card back in the 80's.

ManuelVilardeMacedo: Armandino is in no way validating your claims of being able to freeze cars and their backgrounds at 1/125 shutter speeds.

Armandino: I agree that's the correct choice for that particular photo but literally the only motion there is the hands anyway, a chaotic rock or metal band may well provide situations where frozen motion would be just as valid a choice artistically and technically.

@ManuelVilardeMacedo You and Armandino literally have nothing In common. Notice the part where he doesn't claim at any point that 1/125 can freeze a moving vehicle AND it's background at the same time?

@Armandino you definitely made the right choice with that photo but it might not necessarily be the right one for a lively band in every situation. I think we can both agree that knowing what approach to use on a case by case basis is the important factor.

If you have complete control, perhaps.

If you're like most people taking photos at a concert, you're not immediately in front of the stage, you don't have the fastest lens available and the artist is probably moving all over the place.

"fatdeeman: "...moving vehicle AND it's background..."
That's ITS background, actually."

Nothing says "I'm completely wrong and I know it" quite like attempting to completely change the subject of a discussion.

*cringe* how awkward.

Maybe next time it would be easier make sure you're not claiming to be able to break a fundemental rule of photography before you post and then you won't have to resort to cheap shots in a failed attempt to maintain some kind of dignity (it didn't work by the way, it actually made you look less dignified)

You very very seldom want freeze everything at a concert. If you have a headbaning deathmetal group you want to capture the motion, not freeze it completly. I have been photographing a lot of concerts in murky conditions and i dont think i ever go above 1/250 or 1/320. Sure its always easier if you know the music or what type if music they are playing. Its realy easy to capture good photos if you love the music then you can predict when something big is comming. But for a nikon d5 i guess its for those gigant stageshows with U2 or the like where you have 5 minutes to get a good image. Then its more spray and pray

I have taken shots where 51,200 was still too dark at f/1.2. Live shows that are very dark. Would have loved to have had 204,800 available. Just because you never shoot in the dark, doesn't mean others don't

It's called photography, if there isn't enough "fos" you can not "graph". Read the article on photon shot noise. It's curable only with a faster glass, not by ISO increase which will make matters worse. Cause there isn't any lens faster than f/1.2 commercially available for Nikon, you have to use some sort of additional light source.

I shoot NEF always. I don't think you have to tell me something.

The point was to push conditions to the high ISO... so I am sure you can shoot at the loss of DOF at F2.8, just imagine what shutter speed and light you can shoot that in, sorry your cr@p sensor won't match that, probably one stop or more worse ;-)

Why do you care? The images were supposed to demonstrate high ISO from the camera and they do exactly that.

it´s the LED lights of the venue. not the photographers fault.

the LEDs used today to light stages don´t have a continuous spectrum and therefore you end up with nasty one-colored shapes. very annoying.

Oh, I know there were crazy lights at the concert. What I meant was if you download and look at the raw files for those photos you'd see that the blues were not that saturated in the raw files but are over-saturated in the processed photos.

That's pretty strange. But you're right.

I suspect he's saying shooting in light levels that need these ISOs would be more useful, as then you'll have more noise in the light to mix with camera noise.

"more noise in the light" love it.

I don't follow your comment? You do understand that at any level of light intensity the average noise level is the square root of that intensity. Hence if you have 100 photons on average 10 of those will be noise? (Atoms don't chat together about giving out photons at exact intervals, they just do it at an average rate, which will vary around that average, i.e. be noisy.) Hence if you illuminate a totally uniform surface with an exact light level the pixel values recorded in a photograph will vary, regardless of camera performance, due to that noise in the light. In my example if the average is 100 you will get 95, 98, 101, 104 as well, with camera read noise added on afterwards. That's just physics.

For these testing purposes you want to visit a murky club with crap lighting, and try to get a usable pic of the drummer at the back lurking in the shadows - that's the real high iso acid test.
Not this pro lit scene - most of them are actually overexposed - I want to see this camera in an environment that would really benefit from it.

If you try that stupid ISO values above 100K in a difficult light, the result is going to be more or less random noise. However, pretty impressive in the ISO12800 - 51200 range.

must be only you, thoose look just awesome and will print very well.

This camera will be very handy for photo journalism
Also keep in mind that these are OOC JPG and that shooting RAW will likely yield better IQ at equal sensitivities.

Curious as to what camera does high ISO better in your view?

Useable 60.000 shots. What are you talking about ? The noise is not really visible on print and resized for web, most of them work.

Not to comment on the actual usability of the images (haven't had a chance to shoot the camera myself), but if we had used different aperture and shutter combinations, then some might say we didn't test the higher ISO's intensely enough.

In any case, the ability to use higher ISO's with a deeper depth of field is something worth exploring.

I know what you meant, but what you do not seem to understand is that the camera must be tested to show its limits. Otherwise how would you know how well it performs?

yes, but there´s a chance that this camera will be used on more than just this occasion. and maybe a situation will come up where someone will combine iso 40.000 with f2 and 1/200s to get a shot. would you be satisfied then? or would you then say a shot like that wouldn´t have been possible before so it´s useless now!

In this case they used short shutter speeds and small apertures just to get to those high ISOs they were aiming to test. In a normal situation longer times and faster lenses would naturally be used.

The venue was lighted too well...

I recently shot 2 big production American acts with a D810 and a 24-120 f4 and the 200-500 f5.6 and was very impressed. I have a Canon 1Dx as well and am considering completely selling it along with my Canon lenses and going fully Nikon. Am watching these D5 and D500 threads with a great deal of interest. Having seen these samples I am impressed but not sure if going the jump for a D5 is going to give me much more than I already have.