Top Nikon High ISO performers for DX and FX.

obajoba

Here's a somewhat poor example (but all I have available at the moment) of the D4 at ISO6400. A slightly sharpened and highly compressed jpeg but still better than my D7000 ever was, even at ISO1600.

My D4 import preset in LR4.3 sets my NR sliders to 0, and sharpening to 40 at .8 size, 10 detail, with a 50% mask. With the sliders at 0, LR does not apply NR by "default" or otherwise; "Zero = Zero" from my understanding.

TaoTeJared

It isn't a question of the effectiveness of the system as it stands now seven, I think they should couple it to the VR system so it ups the ISO if you have the shakes. Easier said than done with the current lens connection system perhaps but it would be unique if they could do it.

They have changed it a bit with the new bodies where you have just one more control to adjust it. Works better for sure.

I think what you are getting at is a system that 1st - tells if shaking exists, 2nd then applies the correct compensation, 3rd "user choice" to have priority of sharpness or IQ (lowest iso) or middle of the two.
At least that is what I would like to see or even more detailed than that.

TaoTeJared

Of course, spnp, but then we should as well consider that different RAW converters will lead to different results even if no additional NR is added. So, it's always the combination of RAW converter with cam.

And even when we add NR in post - different RAW converters use different ways of NR. So, +15 in AA could lead to similiar results as +4 in LR (just picked up numbers wild guessing)

That builds a bit on what I said above. I use LR 4.3 mostly but on really high iso images, or where I need more control I use Nik's Define. It is much better and you can control different areas etc. I can't get my D800 close to a D4, but I can use at least 1 stop higher (generally lock auto iso to 2000 - 3000 for work) and given the situation, I will let the system go to 6400 and just plan the extra time in post. It all depends on how much "fine" detail needs to be seen. 5000 still gets me fairly defined eyelashes.

obajoba

That builds a bit on what I said above. I use LR 4.3 mostly but on really high iso images, or where I need more control I use Nik's Define. It is much better and you can control different areas etc. I can't get my D800 close to a D4, but I can use at least 1 stop higher (generally lock auto iso to 2000 - 3000 for work) and given the situation, I will let the system go to 6400 and just plan the extra time in post. It all depends on how much "fine" detail needs to be seen. 5000 still gets me fairly defined eyelashes.

+1 for Dfine 2.0 - Love it. My only issue with the Nik plug-ins + LR4.3 is the inability to reopen the image and tweak adjustments (or maybe I am doing something wrong?)

I may be misinterpreting, but I have been under the impression that the D800 actually outperforms the D4 for retention of detail at high ISO. At 6400 the D4 is amazing, and even up to 10,000 I would guess that one could get an "acceptable" 8x10 print.

TaoTeJared

I may be misinterpreting, but I have been under the impression that the D800 actually outperforms the D4 for retention of detail at high ISO. At 6400 the D4 is amazing, and even up to 10,000 I would guess that one could get an "acceptable" 8x10 print.

I'm not sure about that. I look at color accuracy lumped into "detail" - If you can see an eyelash but it is green rather than black... I'll defer a comparison to those with the D4, but to my eyes the D4 has it above 5000 or so. That is one area that is really difficult to judge since every factor (lighting, shutter speed the scene allows, colors in details, contrast, etc) all multiply in the various directions of good and bad. For instance, if the detail is in the "high-mid to highlights range" of a dimly lit scene the D800 grabs the detail in amazing fashion. If the detail is in the "mid to low" range the D4 does a much better job to my eye. That is a general statement though.

When shooting at extremes everything makes a difference. F8 at 6400 will be better at details than f1.4 at 6400. I know that is assumed, but sometimes forgotten as so many (including me) want to shoot wide open in dimly lit situations which can actually be a lost opportunity to get a better image. But then of course that depends on what shutter speed you need to freeze action/movements.

heartyfisher

Thanks all for the responses so far..

So to summerise we have
1) D3S is the champ but only by a small margin that is more than made up by the other attributes of the D4 ie Colour depth Dynamic range etc.
2) D4 - really is the top choice because of the much higher DR etc.
3) D600 !! - This was a bit surprising to me. that it beat out the D800. Of course the D800 is much more of a camera than the D600, however for those from the DX world that really want to improve their High ISO capability but the D4/D3S $$ are way out of reach, the D600 being in effect the second best High ISO performer would be a great attraction (if you dont need the features of the D800 eg: better focusing system etc.)
4) D800 - That this 36MP sensor beats the D700's 12MP in High ISO is a marvel !
5) D3/D700
6) D3X
So do we have consensus for this list and order for the High ISO capability of the FX sensors from Nikon ??

haroldp

I am getting excellent results with a D800e at ISO 6400 by down ressing to 12mp in post.

The D800 using this process is much better than the 12mp D700 at 6400, and at ISO 6400 looks like the noise of the D700 at ISO 2400, but retains more detail.

The D3x down ressed to 12mp is the noise equal of the 12mp D700 at high ISO's but retains more detail.

I do not have a D4 or D3s but Thom Hogan (who inspired me to try this technique, thank you Thom ) claims that up to 6400 the downressed D800 is about the equal of the D4, which does then pick up a DR advantage above 6400.

I do a lot of theater work under stage lights, and was planning a D4, but since I do not need frame rates higher than 4 fps, the versatility of the D800e is the deciding factor.

The worlds top landscape DSLR at 36mp FX, a 15 mp DX for wildlife (at 5-6 fps), and a 12-18 mp very good low light camera.

I still use both the D700 and the D3x since most of my requirements are not at any extreme, but in the middle, where they are all good and ( at ISO 400 and 11 X 14 size print) practically indistinguishable.

However if I could only have one Nikon body, today it would be the D800e, and If I were adding one it would be another D800 (not e).

Regards .... Harold

heartyfisher

@ haroldp : Thanks for that ! very good information and process.

RenardRichie13

D600 beat d800 by 1 stop? No way where did you ever get that idea. You're definitely right if we're looking at each photo 100% output (36mp vs 24mp) and looking at it down to pixel size. But Lets be honest. Exporting at full Megapixel output for ANYTHING (wedding in my instance) would be stupid. I wouldn't be able to fit 500-700 photos on a 8gb usb stick if i ever want to deliver to clients.

I've been outputting wedding photos to about 12 inches (3872 width) Long edge limit checked on the output on LR. That is about 10-10.5 megapixels. If they want to print their prints say wall sized, I'll reoutput it separately

Now imagine d800 compressed even d600 compressed to a 10mp photos. By that logic alone a photo that i just took from my d800 at 6400 would have equal or less noise than even d4 taken at 6400. I might be crucified for this but so far I have no fear shooting my d800 around 4000 if i need to. Problem is when it hits iso 3200 Up i usually take my flash anyway which basically making my iso adjustment around 200-400 then adjust flash accordingly.

studio460

That looks about right to me, Hearty. Where did you get the D600/D800 one-stop figure? [FWIW: DxO's sensor score for the D600, D800, D800E bodies are 2980, 2853, 2979, respectively (higher number is better)].

I would again like to post a comment I've heard from at least two Getty photographers I know, "[The D4] is noisier than the D3s at ISO 3,200." All Getty staffers were shooting dual D3s bodies, prior to the D4's release. Now, all are shooting with dual D4 bodies. This is largely the reason I went with another D3s for my second FX body (for when I need two matched FX bodies). Of course, another reason for buying two D3s bodies (one refurbed, one used) was that this route was significantly cheaper than buying two D4 bodies.

To me, what this means, is that working photographers are noticing a visible difference between the D3s' and D4's noise performance, however slight, under real-world conditions (which as DxO measures it, is only about a 1/8th-stop performance benefit for the D3s). And, believe me, these are no pixel-peepers--these guys are always in a hurry. I think it's significant that they make these remarks, unprompted, after being asked only, "How do you like your D4?"

heartyfisher

?? Where is this 1 stop difference between D600 and D800 that you guys are talking about ?

TaoTeJared

?? Where is this 1 stop difference between D600 and D800 that you guys are talking about ?

There isn't.

ads

"Where is this 1 stop difference between D600 and D800 that you guys are talking about ?"

They are talking about pixel-level performance, which is right from my limited testing of the 2 bodies.

BUT when you print or view on screen the difference is almost impossible to see because the image is resized.

haroldp

It is only valid to compare images displayed or printed at the same size.

... H

TaoTeJared

?? Where is this 1 stop difference between D600 and D800 that you guys are talking about ?

Heartyfisher - you are the one who said the D600 had a 1-stop difference in your original post. Where did you come up with that since so far you are the only one who has claimed it.

There is no such thing as pixel-level performance, there is just performance - there is not sub-categories in what we have been talking about. 100% zoom view, is not the same as performance.

Golf007sd

I would again like to post a comment I've heard from at least two Getty photographers I know, "[The D4] is noisier than the D3s at ISO 3,200."....."How do you like your D4?"

First things first, I love my D4 and the manner in which it handles noise at high ISO.

With respect to those comments the the two Getty photographer...well here are some shots for you to look at and then decide for yourself if there comments are valid.

Note: These photo's where provided by Photographylife.

D4 ISO 3200


D3s ISO 3200


D4 ISO 6400


D3s ISO 6400

ads

?? Where is this 1 stop difference between D600 and D800 that you guys are talking about ?

There is no such thing as pixel-level performance, there is just performance - there is not sub-categories in what we have been talking about. 100% zoom view, is not the same as performance.

Well each pixel performs to a certain level, so there is such thing as pixel level performance - 100% view shows that performance. Eg if we were comparing high iso poster prints done with a D300 with a D3s (ie same number of pixels) the difference in pixel level performance would be obvious.

What makes the issue less relevant when we're comparing the D800 and D600 is that the D800 shots will be downsampled when viewed at the same size.

TaoTeJared

We are talking about photos not micron sized digital parts.
As haroldp said...

It is only valid to compare images displayed or printed at the same size.

... H

ads

Well the size of the pixels and their relevance depends entirely on how big you print, but let's not derail the thread with a semantic argument.

The take away from this thread for me is that, when viewed at a similar size, high ISO performance is not a decisive factor in choosing from the current FX lineup (unless you intend shooting above 25600!) so the choice should be made according to features/form factor.

From that perspective it does make the D600 great value (well, comparatively - its still a $2k+ camera after all) if low light/high ISO is your main requirement

sevencrossing

One question
When using high ISO values, one is usually work in low light conditions
I am very impressed with my D800 ability to focus in low light
is the there any comparison between the D600 and the D800 focusing in low light

focusing seems to be a D600 weakness

heartyfisher

@ TaoTeJared : Re the 1 stop performance between the D800 and D600 i claimed in my first post. I was trying to confirm this assertion which I found stated on the internet a few times but I myself was skeptical about it. I think Ade clarified this by saying that at pixel level the D600 is indeed better than the D800 by almost 1 stop.

However, I have been looking into this a bit more and for down re sized files files from both the D600 and D800 it looks to me like they are almost identical, maybe a slight edge to the D600 but no where near 1 stop. maybe 1/8 stop to my eye. So you really need be seriously pixel peeping to see the difference ! That still makes the D600 a huge bang for buck if High ISO performance is a Key Performance attribute for you.

TaoTeJared

Grumble - there is no pixel level... That thought process is the misunderstanding and incorrect application of technology onto a end product. Image Noise is a result of the sensor diodes reading light & SOFTWARE interpretation of the results.

Upsample the D600 image if you want to see how it performs as well. People forget it is a two way road. Everything I have seen, tested difference measured is about 1/16th of a stop. The actual math is 1/15th of a stop of iso values (between 1600 & 3200). DXoMark Sports (Low-Light ISO) (about the only test I do look at from them) - D600 = 2980 - D800 = 2853 = difference of 107. The actual visible noise is no difference once I ran image samples through LR Until you start to crop the images - then the D800 wins due to the resolution.

I would echo sevencrossing's low light focusing - the D800 is great. From what I have seen, the D600 seems to focus like my D300 in low light (similar results) and my D800 kicks my D300's butt in lower light focusing. Even indoors at night with lamps lighting a room, I can tell a big difference. The D800 locks correctly much more often than the D300.

DXV_Photo

Here is a review I found comparing high ISO between D600, D700, D800E and D4. It covers dynamic range between all the cameras and resampling the images down to 7MP.

heartyfisher

2 DXV_Photo : Thanks for finding and sharing that review video. ! Very nice info there. It looks like upto 6400 iso there is really no significant difference between the D600, D800 and D4. although the D600 is a tiny insignificant bit better than the D800. So although it doesn't change the order of ISO performance we had listed previously it does show that its much closer than I previously thought.

haroldp

Except for very specialized circumstances, the other properties of these 3 cameras (size, weight, control system, focus, resolution, cost, frame rate, versatility, body robustness etc) are the decision points for purchase or suitability. Relative high ISO performance is a trivial factor in this comparison.

I am not even sure that these very small differences in quantum sensor efficiency will ever actually translate to a noticeable and repeatable difference in real photography and PP.

Many still use D300, even though the D7100 is a much better high ISO camera because of the above reasons.

Regards .... H