Mini-Review: Portrait Lens Shootout on D3

[Jonathan]

Excellent test and very helpful. Thanks once again for taking the time to do this.

 

[SMH77]

Why did you ditch the 200VR?

A couple reasons:

a.) I'm focusing mostly on weddings and portraits--200 is just too long to get 'stuck with' when the action changes unexpectedly during a ceremony where you want to capture more of the scene, so it was very difficult to justify grabbing it during a ceremony

b.) My wife would shoot with it during the ceremony, but was finding being fixed at 200 mm too limiting to get proper composition, at times, from the places she was shooting

c.) Because of A & B above, and deciding that I wanted to go with a second D3 to put my wife on level playing ground as me, I decided that the cash tied up in the lens was better spent on D3 #2

d.) definitely too long for an indoor studio

It was very difficult to part with since the 200 f/2 VR is such a special lens. I'll probably come back to discuss the 200 f/2 some more as I get some time to peruse some of the portraits I took with the 200 f/2 on the D3 before I parted with it. (Sorry Keith for not getting back to your PM yet--I'm really trying to juggle my time).

The 200 f/2 is such an awesome lens. I have posted many portraits with it in "THE" thread here on the Cafe.


Sean

 

[SMH77]

I wonder how the 28-70 would do in this test since it is supposed to render skin tones better than the 24-70?

You know, as I was studying the 'studio' round of this comparison, I (and my wife) could see a distinct increase in the 'yellow/orange' of the two zooms tested--with the 24-70 being the most extreme. The two primes looked close to one another, though the 105 DC has 'just the right amount' of yellow in the skintones to make the skin look 'just right' while the 85 f/1.4 has less and makes it look a little 'cool'.

While pondering above observations, I thought about the 28-70 and thought, based on the age of the lens, it would better match the two primes rather than the zooms--which would lead me to believe that it would be improved in skintones over the 24-70. I'm certain it wouldn't match the 105 DC, but it would probably be closer to the 70-200 than to the 24-70, so it wouldn't surprise me if it did test out as being 'better' in the skintones department than the 24-70. I'm speculating of course, but it's 'educated speculation' so it's valid I think...


Sean

 

[SMH77]

Round II: in the Studio, same lighting, same WB preset, skin rendition discussion

As promised, today I was able to battle through another comparison round in the studio to test these 4 lenses again--under more controlled conditions. This time, I pre-set the WB with the 70-200 on, under the flash lights using a Whibal, and used that pre-set for all 4 lenses to compare the skintone / color rendition differences of the 4 lenses tested. Again, I tried to match the composition as best I could by moving the whole setup forward and back until it was just about right. I wasn't perfect though as my 'subject' had less patience today than my subject from yesterday. :frown:

A few more notes about todays test:

Lighting was via two SB800's, main fired into a ~30" umbrella, hairlight shot through a 'mini-softbox' (BTW, don't be thrown off--the 70-200 shot had the hairlight a little too hot compared to the other three photos so it also affects the background brightness).

The D3 had Active D-Lighting turned on 'Low' in camera, later turned to 'Off' in NX. 14 bit NEF files were captured and converted in NX (same as yesterday's test). In post I compared the avg 'RGB' values (5 pxl X 5 pxl avg in NX) of the whites of her left eye (just to the right of the dark part of the eye) in order to adjust the exposure to get an 'equivalent' comparison where exposure wasn't influencing the skintones. Focus was on her left eye (right eye in the picture) in each case. 24-70 was still not 'AF fine tune adjusted' so that may have an impact on the sharpness ranking below.

Again, originals can be downloaded from the same link as in the original post.

All shots below were shot at f/2.8 (DC set to '0').

The Data


70-200 VR @ 105mm & f/2.8

Whole scene:

Subscribe to see EXIF info for this image (if available)

Face Crop (30% this time):

Subscribe to see EXIF info for this image (if available)

105 f/2 DC @ f/2.8 and DC neutral

Whole scene:

Subscribe to see EXIF info for this image (if available)

Face Crop (30% this time):

Subscribe to see EXIF info for this image (if available)

85 f/1.4 @ f/2.8

Whole scene:

Subscribe to see EXIF info for this image (if available)

Face Crop (30% this time):

Subscribe to see EXIF info for this image (if available)

24-70 @ 70 & f/2.8

Whole scene:

Subscribe to see EXIF info for this image (if available)

Face Crop (30% this time):

Subscribe to see EXIF info for this image (if available)

Observations

My wife and I spent quite some time studying the 30% views side by side in NX and both clumped the lenses as follows: the colors of the 70-200 & 24-70 are more 'yellowish / orange' (Jaundice tendency), while the two primes are close to one another and more realistic--with the 105 being picked straight out by both my wife and I in terms of accuracy of color/tones. Overall the sharpness is very close between the 70-200 and 105 DC with the 24-70 being the softest (could VERY WELL be focus tuning issue) and 85 f/1.4 sharpest. Micro-contrast: the 85 f/1.4 is definitely the most harsh of the 4 lenses with the 105 DC seemingly able to 'hide' the very small shadows from the texture on the skin--yet all the while keeping the resolution (it's pretty difficult to try to explain) very high. I think this trait of the 105 DC is what leads me to rank the 70-200 as being slightly sharper than the 105 DC (below)--I think it isn't actually resolution, but micro contrast that makes the 70-200 appear to have a very slight--and I mean VERY slight--edge in 'sharpness'.

So here are my rankings in terms of a couple different lens traits:

Color rendition / skin tones (coolest on top, 'warmest' / 'yellowist' on the bottom):

Coolest
85 f/1.4 (too cool)
105 f/2 DC (perfect)
70-200 f/2.8 VR (tainted 'yellow')
24-70 f/2.8 (even more 'yellow')
Warmest / yellowist

Note: it was interesting to see that the 24-70 actually blew out the red channel on the 'show lost highlights' in NX on her cheeks and nose, while none of the others blew any channels at all (except for the hot hairlight on the 70-200, which was my fault). Perhaps this 'blown read channel' that we're experiencing on the D3 is more related to the wide use of the 24-70 lens on the D3 than the D3 itself? Just speculation after witnessing this first hand...

Micro-Contrast: tendency to show shadows from very fine skin 'bumps'/texture in the well lit and transitioning into shadow areas

Nicest to the skin
105 f/2 DC (almost looked like some blems were removed from the skin, it was so good!)
(blank to emphasize the separation this lens has over the others in this area)
24-70 f/2.8 (may be a function of not having been 'focus tuned' though)
70-200 f/2.8 VR
(blank for emphasis)
85 f/1.4 (Brutal?)
Most harsh rendition of skin 'features'


Sharpness (or perceived sharpness)

Sharpest
85 f/1.4
70-200 f/2.8 (tied with 105 DC?, micro contrast gives it the apparent edge)
105 f/2 DC
24-70 (may rank higher with focus tuning)
Softest

Conclusion

My wife and I both picked the 105 f/2 DC as the best 'skintone' lens out of the bunch, but this really shouldn't be a surprise as Nikon designed it as such! I really like having the 105 DC on the body as it's nice and light and gives good reach. Interestingly, I had the most 'resistance to focus' with the 105 DC under darker focusing conditions than I did with the other lenses--not sure what that is all about, but it was annoying at times.

When it's all said and done (and after looking at my finances), I think I'll be hanging onto the 105 f/2 DC and parting company with the 85 f/1.4 and 70-200 f/2.8 VR. I own two 70-200's right now, so parting with one shouldn't be a big deal. The 85 f/1.4 is a great lens, but I think the 105 mm focal length works better on the D3 when combined with the 24-70.

Thoughts, comments?


Sean

 

[Paul.r.lindqvist]

How did you set the wb ?

 

[SMH77]

How did you set the wb ?

I used the Whibal (a 'neutral' card), held it where the flash would hit it when pressing the shutter, and set a 'PRE' set WB setting. This was set with the 70-200 VR and was not changed for the remainder of the testing.


Sean

 

[Paul.r.lindqvist]

So no custom wb for each lens ?

 

[Jonathan]

Super job!

The 85 f/1.4 is a great lens, but I think the 105 mm focal length works better on the D3 when combined with the 24-70.

Couldn't agree more on the above statement.

Sean, excellent test and results(great model too!). I've owned all of these and at this time have the 105 f2 and the 24-70. I'm in the process of going back to the 28-70 because I really think it's better with skin tones than the 24-70. The 24-70 I have is super sharp, but I've been bothered by the skin tones and the contrast being a bit over the top. I bought it for portraits, and it should be A+ in that area. I don't think it is. I know most are very happy with this lens, so I'll clam up on that.

On the D3 the focal length is just right with the 105. I never liked it on the
D300.

Thanks again for posting this.

I plan to do a similar test this week between the 24-70 and the 28-70, although my 'subject', won't sit still for very long.:smile: I'll post the results.

 

[SMH77]

Couldn't agree more on the above statement.

Sean, excellent test and results(great model too!). I've owned all of these and at this time have the 105 f2 and the 24-70. I'm in the process of going back to the 28-70 because I really think it's better with skin tones than the 24-70. The 24-70 I have is super sharp, but I've been bothered by the skin tones and the contrast being a bit over the top. I bought it for portraits, and it should be A+ in that area. I don't think it is. I know most are very happy with this lens, so I'll clam up on that.

On the D3 the focal length is just right with the 105. I never liked it on the
D300.

Thanks again for posting this.

I plan to do a similar test this week between the 24-70 and the 28-70, although my 'subject', won't sit still for very long.:smile: I'll post the results.

I'd love to see the comparison between those two zooms! I've thought about picking up (sometime in the future) a 28-70 for my wife as a mid-range zoom--but alas, the bank account is currently 'feeling the heat' to put it mildly so that will be a ways off... :biggrin:

It's good to get confirmation on my observations though...


Sean

 

[SMH77]

So no custom wb for each lens ?

That's correct Paul, for the studio version of this test--and that was intentional as well to demonstrate the differences in the lens' renditions of colors. I thought about duplicating it with custom WB's as well for each lens, but was getting 'frustrated' trying to keep the whole process going (imaging 4 kids 8 and under running around the house (2 boys) with a whole studio setup, lenses out and accessible, a couple D3's, etc. and you'll get the point about 'frustrated'... :biggrin.


Sean

 

[Paul.r.lindqvist]

Ok understand.

 

[SMH77]

Ok understand.

I should also add that the first segment of the test (outdoors) did use a custom WB for each lens--though with the changing lighting conditions it may not have been as effective.


Sean

 

[Paul.r.lindqvist]

Yes i noitced the apparent shift in lightning, a fixed wb wouldnt have improved the consistency imo.



I dont know the exact colors of your daughters dress so its impossible for me to evalute wich one is closest to be "correct" eventhough im using a calibrated CG.

A color chart, wich isnt alot of fun for most people to look at, will tell me alot more about the colors each lens produce, since i know the fixed rgb values.


Its a big diffrence what is "correct" and what is most "pleasing"
(needless to say, pleasing is very subjective)

The thing with skintones is that most people dont like the "correct" tone.. often one want someting a little warmer.


Nice test with some real world examples, and this test confirm that the 85/1,4D does have a cool cast.

Edit:What picture control did you use ? (since you used NX)

 

[SMH77]

Yes i noitced the apparent shift in lightning, a fixed wb wouldnt have improved the consistency imo.



I dont know the exact colors of your daughters dress so its impossible for me to evalute wich one is closest to be "correct" eventhough im using a calibrated CG.

A color chart, wich isnt alot of fun for most people to look at, will tell me alot more about the colors each lens produce, since i know the fixed rgb values.


Its a big diffrence what is "correct" and what is most "pleasing"
(needless to say, pleasing is very subjective)

The thing with skintones is that most people dont like the "correct" tone.. often one want someting a little warmer.


Nice test with some real world examples, and this test confirm that the 85/1,4D does have a cool cast.

Edit:What picture control did you use ? (since you used NX)

Paul,

Thanks for the comments. Yes, I agree that 'correct' wasn't the 'correct' term for me to use, but rather 'most pleasing'. Yes, 'just the right' (subjective) warm is the actual 'correct' tone...

Picture control was 'Standard', with only the sharpening turned up to '5'. Yes, this also has an effect. So many parameters...


Sean

 

[Paul.r.lindqvist]

I dont use NX on a daily basis, im curious to know what camera did you use before the D3, and how do you feel the colors compare to your old body ? I personally dont like the standard mode in NX at all. Way to much contrast and saturation (to begin with) this has offcourse to do with ones workflow i do most of my color correction/adjustment in CS3.

So netural or the D2Xmode1 is the one i find most pleasing to start with as a base.(when i use NX)

Kindest

 

[SMH77]

D2X Mode I conversions in NX

Here are the same pictures converted to D2X Mode I in NX, adjusting only the sharpening up to '5' like was set on the 'Standard' style of the previous examples.

70-200

Subscribe to see EXIF info for this image (if available)

105 f/2 DC

Subscribe to see EXIF info for this image (if available)

85 f/1.4

Subscribe to see EXIF info for this image (if available)

24-70 f/2.8

Subscribe to see EXIF info for this image (if available)

Sean

 

[arela]

very informative ....and your latest model deserves an icecream

 

[Radiohead]

For me we're talking about microscopic differences here. A worthwhile exercise though. I personally don't consider the 24-70 a true portrait lens - it's a high-class standard zoom aimed at PJs and event shooters that can do many things very well, but is outperformed as expected by the specialists.

Interestingly, the D2x modes make the 105DC look the coolest, here at least.

 

[Paul.r.lindqvist]

Sean, thanx.

The 105/2 is to cool for my taste, out of the three i would pick the 24-70/2,8.

Kindest

 

[Crazy Cat Lady]

Very interesting comparison. I just bought the 85 1.4, should be here today :biggrin: I have been looking at the 135 f/2 DC. Should the 135 f/2 DC give the same skin tone's as the 105 DC?

Thanks!