D3 vs D300 Dynamic Range

[Greg Seitz]

Hi Folks,

As a spin off of the D3 vs D300 thread I decided to do some dynamic range comparisons between the two. My approach to testing the dynamic range was to expose a bracketed set of shots on both cameras and then test the ability to push and pull them in Adobe Camera Raw. This gives me a real world idea of how much flexibility and range both cameras have at the extremes which translates to real world shots where I need to pull up the shadows and pull back the highlights.

The test setup was shot on a tripod using the Nikon 70-200 VR for both cameras. I equalized the field of view by using a zoom of 70mm for the D300 and 105mm for the D3 (actually the lens reported 110mm to get the framing as close as possible - probably in reality somewhere between 105 and 110). Both sets of shots were focused using live view. Once focused, I bracketed 9 shots from -4EV to +4EV using mirror lockup with a remote release for each shot. To minimize the depth of field differences I shot the D300 at f/7.1 and the D3 at f/10 (1 stop difference) and adjusted the shutter speed accordingly.

To test the ability to push up the shadows, I pushed the -4EV shots to 0EV by adding +4 in ACR to match the +0 shot.

To test the ability to bring back the highlights I pulled the +2EV shot back to 0EV by setting -2 in ACR to match the +0 shot.

The default settings were used in ACR with the exception of the black level at 0 and I used the following for camera calibration:

D300 calibration:

Shadows: 0

Red Hue: 0
Red Saturation: +8

Green Hue: +11
Green Saturation: +4

Blue Hue: 0
Blue Saturation: -22

D3 calibration:

Shadows: 0

Red Hue: -2
Red Saturation: +13

Green Hue: +18
Green Saturation: +9

Blue Hue: -1
Blue Saturation: -18


I'm also showing the +0EV shot as a reference.

Below I'll post the shots by camera and exposure so you can compare them. For those that want to dig into the raw files I'm enclosing a zip to a subset of each with -4EV,-2EV,-1EV, 0EV, +1EV, +2EV. I've also got a set of shots on the D3 at f/7.1 for reference.

Here is the raw file set for the D300 at f/7.1:

http://www.mediafire.com/?wbcyhqwdtdc

Now the D3 at f/10:

http://www.mediafire.com/?jtwzznvl0tt

and finally for reference but not shown here the D3 at f/7.1:

http://www.mediafire.com/?jtwzznvl0tt

Enjoy!

Greg

 

[Greg Seitz]

D300 & D3 as shot

In all of the following the order will be D300, D3, D300 crop, D3 crop:

This is the +0EV exposure:

 

[Greg Seitz]

D300 & D3 -4EV as shot

This one is the -4EV as shot, again D300, D3, D300 crop, D3 crop:

 

[Greg Seitz]

D300 & D3 +2EV as shot

This one is the +2EV as shot, again D300, D3, D300 crop, D3 crop:

Had to edit this one, had the wrong D300 as shot. It's now the last one since I couldn't figure out how to place the edited one first. Sorry for any confusion...

 

[Greg Seitz]

D300 & D3 -4EV corrected with +4

This one is the -4EV corrected to 0EV with a +4EV push in ACR, again D300, D3, D300 crop, D3 crop:

 

[Greg Seitz]

D300 & D3 +2EV corrected with -2

This one is the +2EV corrected to 0EV with a -2EV push in ACR, again D300, D3, D300 crop, D3 crop:

 

[Greg Seitz]

Ok, that should do it. Let me know what you think...

Greg

 

[asabet]

Excellent test Greg! This is how I like to examine DR also. Gives a practical sense of what the RAW shooter has to work with. To me, it seems that some of the recoverable shadow detail in the pushed underexposed D3 shot is lost amongst the shadow noise in the D300 shot. The differences on the highlight end are more subtle, but I'd again give the edge to the D3. Regards, Amin

 

[Greg Seitz]

Excellent test Greg! This is how I like to examine DR also. Gives a practical sense of what the RAW shooter has to work with. Your test demonstrates very clearly how much more the D3 offers in the shadows. Shadow detail which is usable in the pushed underexposed D3 shot is lost amongst the shadow noise in the D300 shot. The differences on the highlight end are more subtle, but I'd give the edge to the D3. Regards, Amin

Amin,

Not sure I'm seeing what you are regarding the shadows. I see maybe a small advantage on the shadow side but not much. What surprised the heck out of me is that I'd say the D300 pulled up the shadows very well and I'd dare say may have outdone the D3 there since the D3 is showing a hint of banding and the D300 shows none.

Greg

 

[asabet]

I went back and edited my comment after looking at it for a while longer (and before I saw your response, or I wouldn't have edited it ). I still think the D3 recovers more shadow detail in the bottom left corner of the crop, but it's not as clear a difference as I initially though.

 

[jonoslack]

HI Greg
Interesting stuff - I think the link for the f7.1 on the D3 is the same as that for the f10.

I do understand why you used f10 . . . . but I'm not sure that you were right - you've equated the dof, but I can't get my head around whether you were right to reduce the aperture or not . . . so it'd be nice to see the 7.1 if you have the link.

I thought I'd have a look at them in Aperture, Incidentally - why did you only overexpose by +2 and not +4 (it always seems to me where there is a lot of headroom in the D3 files.)

Incidentally, I think you're VERY brave to publish this on dpreview . . . or VERY foolish . The toys are flying out of the prams as we speak!

 

[LCT]

....and finally for reference but not shown here the D3 at f/7.1: http://www.mediafire.com/?jtwzznvl0tt...

Thank you Greg but the files you're quoting above are f/10's. Did you keep the f/7.1's by chance?

 

[jonoslack]

Re: D300 & D3 -4EV corrected with +4

This one is the -4EV corrected to 0EV with a +4EV push in ACR, again D300, D3, D300 crop, D3 crop:

Hi Greg
I've also done this one in Aperture - principle is the same - however, what I have noticed is that on the D300 shot there is a splash of sunlight in front of the chart - on the D3 this is further away, and would certainly have reflected less light on to the chart.

What do you reckon?

 

[jonoslack]

This is what I got with default settings (+4 and 0 black point)
aperture left ACR right
D300 top
D3 bottom

Then with -2 on the overexposed (still would have liked to have seen +4)

For what it's worth!

 

[robmac]

Greg, excellent. Thank you. I see what you mean re: vertical banding on the pushed D3 shot, the D300 has more chroma noise when pushed, but at least it's irregular.

Jono - Aperture does a sweet job against ACR - especially in the pushed shots. Have been considering it as DAM since 2.0 came out and price dropped. Would be curious how the push/pull would compare against NX.

 

[Steven Philips]

Hi.

I've used both Aperture and Lightroom, and they are different enough, and these test results are similar enough, that I wouldn't base my decision on which software to use based on the DR handling question. Maybe my eyes aren't that good, but the test shots look extremely similar to me.

Steven

 

[Greg Seitz]

Re: D300 & D3 -4EV corrected with +4

Hi Greg
I've also done this one in Aperture - principle is the same - however, what I have noticed is that on the D300 shot there is a splash of sunlight in front of the chart - on the D3 this is further away, and would certainly have reflected less light on to the chart.

What do you reckon?

Hi Jono,

Regarding lighting, the shots were taken roughly mid day 4 minutes apart. What you're seeing in terms of the lighting change are shadows from some 100+ ft tall Douglas Firs off to the left of the scene. So obviously areas where the shadows changed would show differently between the two. In the crop I provided, the chart remained in shadow on both, but examining other areas of the originals reveals the same amount of shadow recovery. Even in the same crop, take a look at the branch detail at the base of the bush to the left of the chart. Same for the highlight detail. Were you able to spot any major differences in other areas that remained in deep shadow?

Thanks,

Greg

 

[Greg Seitz]

This is what I got with default settings (+4 and 0 black point)
aperture left ACR right
D300 top
D3 bottom

For what it's worth!

Thanks for posting these. I've noticed myself that Aperture seems to pull a bit more detail out of higher ISO shots as well as when pulling the shadows up. But everytime I find and example where Aperture does better I seem to find another where ACR does a bit better. Generally they are pretty close most of the time.

By the way, I do have the +3 and +4 shots, I didn't include them due to the 100MB limit of mediafire. I can package them up separately and upload them. I'll post a link when they are up. I did find however that +2 was about the limit for true highlight recovery on both.


Greg

 

[jonoslack]

Re: D300 & D3 -4EV corrected with +4

Hi Jono,

Regarding lighting, the shots were taken roughly mid day 4 minutes apart. What you're seeing in terms of the lighting change are shadows from some 100+ ft tall Douglas Firs off to the left of the scene. So obviously areas where the shadows changed would show differently between the two. In the crop I provided, the chart remained in shadow on both, but examining other areas of the originals reveals the same amount of shadow recovery. Even in the same crop, take a look at the branch detail at the base of the bush to the left of the chart. Same for the highlight detail. Were you able to spot any major differences in other areas that remained in deep shadow?

Thanks,

Greg

HI Greg
Maybe you're right - I would have been interested in seeing the +4 shots, and also the f7.1 D3 shots (have you got them?), as it's highlight recovery which has really impressed me with the D3.

Otherwise - I certainly 'like' the d3 files better, but it's certainly pretty close

 

[Greg Seitz]

HI Greg
Interesting stuff - I think the link for the f7.1 on the D3 is the same as that for the f10.

I do understand why you used f10 . . . . but I'm not sure that you were right - you've equated the dof, but I can't get my head around whether you were right to reduce the aperture or not . . . so it'd be nice to see the 7.1 if you have the link.

I thought I'd have a look at them in Aperture, Incidentally - why did you only overexpose by +2 and not +4 (it always seems to me where there is a lot of headroom in the D3 files.)

Incidentally, I think you're VERY brave to publish this on dpreview . . . or VERY foolish . The toys are flying out of the prams as we speak!

Whoops, here's the correct link to the f7.1 shots on the D3 - no wonder no one downloaded them. :ROTFL:

http://www.mediafire.com/?l1gdt0j1ylz

Regarding 10 vs 7.1 I didn't want people to focus on the depth of field differences and cry foul. Being wise to the ways of these tests I took the easy way out and provided both (well at least I thought I did!).


Greg