Does the D3 cheat ISO-wise? (Very serious)

michaeladawson said:

D3 D200
=== ===
137 160
143 165
141 163

I'll think about this test overnight. What I want to determine next
is what ISO do I need to boost the D3 to in order to get the same RGB
readings as the D200?

Click to expand...

i would leave it in color and the same picture control - and probabaly manually white balance and shoot with a white balance card in the scene

i would shot a gray scale or the gray card with different exposures (zones)- bracket- i would want to make sure i am not just looking at a tone curve difference

i would use the exposure slide in capture NX to see how much exposure comp it would take to make them eqaul

good luck!

David

I see you enjoy leaping to conclusions. Again.

Anthony

MackII said:

After your first post I had already thought you had nor read my post
carefully.
Now I know for sure.
Plus I know a thing or two about your character.

Click to expand...

--
check out my blog at http://anthonyonphotography.blogspot.com

Thanks for the feedback. Comments below.

David314 said:

i would leave it in color and the same picture control - and
probabaly manually white balance and shoot with a white balance card
in the scene

Click to expand...

I can leave it in color. What I'd really like to do is measure the Lightness channel in LAB mode. Is there a LAB mode in NX? I'd do it in Photoshop but I only have CS2 and I don't want to introduce the Adobe DNG converter as a variable.

I am using a white balance card. The white balance (grey card) IS the subject.

David314 said:

i would shot a gray scale or the gray card with different exposures
(zones)- bracket- i would want to make sure i am not just looking at
a tone curve difference

Click to expand...

I will be doing the bracketing. I had planned on doing ISO bracketing to find the ISOs on the two cameras that gave equal RGB values when the exposure was the same. I can also do exposure bracketing with shutter speed and same ISO I guess.

David314 said:

i would use the exposure slide in capture NX to see how much exposure
comp it would take to make them eqaul

Click to expand...

Yes, I just did that with my first dry run samples. It appears that I have to underexpose the D200 by about 0.4 stops to equal the D3. If I turn it around I find I need to increase the D3 by about 0.4 stops to equal the D200.
--
Mike Dawson

Pardon me if this was asked already. What was the light source, and what was the colour temperature?

--
no text

I would agree with your statement that the D3 does not cheat. I don't think it does.

But as you state with your example of the 5D I think one has to be careful when making statements of a camera being "x" stops better than another.

I am in the middle of some tests comparing the D3 with a D200. My crude initial tests show the D200 to be between 1/3 to 1/2 stop more sensitive than the D3 at ISO 200. That's only at ISO 200, I haven't yet looked at the high ISOs.

So at the end of the day, if my testing shows a 1/3 stop difference I'm not really going to care one whit. That's a far cry from the 1 stop difference the OP is claiming.
--
Mike Dawson

The question is not whether the D3 meter is accurate. The question is whether the sensor sensitivity (ISO rating?) is accurate. Plus, how does the sensitivity of the D3 sensor compare with other cameras.
--
Mike Dawson

This gets more interesting. I started to repeat my tests this evening indoors under incandescent light. The light is not the interesting item.

Using a D200 and a D3 with the 50mm f/1.8 lens at ISO 1600. Took pictures of two Kodak grey cards set up side by side to give a larger target using completely manual exposure. Used 1/3 stop bracketing of the shutter speed. Exposures ended up being in the range of 1/20s (plus or minus) at f/4 with a preset WB on the grey card with each camera.

The preset WB turned out fine with the grey cards from each camera showing equal RGB values when loaded into Capture NX. Now here's where it gets interesting.

Initial load of a D3 and a D200 image into Capture NX, each at 1/20s, f/4, showed identical RBG values in the center area of the frame of around 120. This is just a tad brighter than middle gray. Visually on screen the images looked identical.

But then I stopped and thought about it. The D200 was loaded using non-Picture controls. The D3 image by default was loaded using Picture controls. If I switched the D200 over to use Picture controls it was consistently 0.4 stops brighter than the D3 image using the same Picture control.

Hmmm, not so simple to test afterall. So it seems to me that you can't really use Capture NX to do this test as NX seems to treat a D3 file differently than a D200 file.

So what to conclude from all this? I guess I'll try ACR in CS2, but it means I have to convert both to DNG first as otherwise I can't load the D3 file.

One thing I am comfortable concluding is that any ISO sensitivity difference, if it does exist, is far less than the 1 stop reported by the OP.
--
Mike Dawson

Same scene, Same lens, Matrix, All other factory default settings.
D3 D2h D200 Flash meter IV
at 200
1/60 1/60 1/30-1/40 1/60
2.8 2.8 2.8 2.8

at 3200
1/60 1/60 1/30 1/60
11 11 11 11

D3 is right on and my d200 saying it needs a little more exposure. opening up all files in NX resulted in near identical histograms, with the D200 a hair lighter. So, my D200 is about a 1/3 stop off.

--
David

For A Pessimist, I'm Pretty Optimistic

This time took the files and converted them to DNG format so I could load into Photoshop CS2.

Loading the 1/20s, f/4 image (ISO 1600) from each camera shows that they are very close. But you know what? The D3 image is brighter. In Lab mode the lightness channel shows the D200 file to be exactly 50% whereas the D3 file is about 53-54% in Lightness. So ACR handles the file differently than does NX.

So I'm leaning towards the conclusion that Nikon is not cheating with the ISO specifications of the D3, at least not compared with my D200.

Now let's be clear, these were not metering tests. But then, that was not the claim of the original OP.
--
Mike Dawson

Perhaps we should let the thread fall off the page as others have established the D3 is metering correctly. But I said I would compare the D3 with my D2Hs when it came today, so I did. Boy, are there other things to talk about the D3, and I'll share those thoughts later. But here's my findings. Using the same lens (70-200 VR), Iso 800, f2.8 in A priority, same lighting, I spot metered middle C on my piano, and got a shutter speed of 1/40 on both cameras. No problem with false Iso rating or other meter discrepancies. Meanwhile, what a revolutionary camera. best....Peter
--
http://www.innerimager.com

no text.

no text

--
regards
Janne Mankila, Finland

15.3.08
What's going on? A well known Finnish MP made
a bill on introducing 'love holidays'! If that isn't
a sign of a brighter future, nothing is. Read more...
http://jannemankila.googlepages.com

You asked an important question based on your experience and the forum was able to answer it. That's what the forum is all about! I'm sure Nikon will have your meter adjusted perfectly in no time. best....Peter
--
http://www.innerimager.com

David R Lane said:

Same scene, Same lens, Matrix, All other factory default settings.
D3 D2h D200 Flash meter IV
at 200
1/60 1/60 1/30-1/40 1/60
2.8 2.8 2.8 2.8

at 3200
1/60 1/60 1/30 1/60
11 11 11 11

D3 is right on and my d200 saying it needs a little more exposure.
opening up all files in NX resulted in near identical histograms,
with the D200 a hair lighter. So, my D200 is about a 1/3 stop off.

Click to expand...

According to David's test the D3 meter is spot on, which is consistent with previous DPReview tests on Nikon camera that have shown their metering in general to be very close to nominal. The value for the green channel in ACR 4.4 is 163, considerably higher than the value obtained from a straight 2.2 gamma tone curve.

According to the ISO standard 12232:2006, photographing an 18-percent reflective surface will result in an image with a grey level of 18% √2 = 12.7% of saturation.

http://en.wikipedia.org/wiki/Film_speed

I tested the system response by taking exposures at ISO 200 from a gray card illuminated by daylight at the nominal exposure (manual exposure, central average metering) and bracketing up and down in third EV increments. Data were recorded as 14 bit lossless compressed. I then used Iris to examine the raw data in the green1, blue, red and green2 channels and determined the mean pixel value in a 200 by 200 crop of the central portion of the image. Since the two green channels were nearly identical, I used green1 for analysis. In this test, the nominal exposure was 1/125 (0.008) second at f/5.

The raw data are shown tabular form. The green channel saturates at 15930. The exposure time, data numbers (raw pixel values), percent saturation of the green1 channel, and the corresponding pixel value for a gamma 2.2 space with no tone curve are shown. The green1 channel clips at 15930, not 16383 as some would expect.



A plot of the data confirms linearity and shows a good fit.



One can then plot the percent of saturation in the green1 channel versus the exposure time. According to the ISO standard, the saturation should be 12.7% and the observed value is 11.7%, very close. From the regression equation, the exposure time for 12.7% would be 0.00863 (1/116) s, or -8% from nominal.

The system is performing close to nominal, and there is no serious problem as the OP alleged. It is important to look at the raw data to avoid problems with tone curves, active D lighting, etc, introduced by the raw converter.

Bill



--
Bill Janes

My pathetically crude test (above) ran into exactly what you're talking about with questions of tone curves, RAW converters, etc. Even so, I was satisfied that the D3 is not somehow overrated when compared to another Nikon body.

Your tests are much better.
--
Mike Dawson

I totally share the opinion of the OP.

So sorry that forums are filled with people aggressive and close minded.

The OP was posing a serious question. Your answers tell nothing and are insulting.

Alessandro

From your starting post in this thread,
http://forums.dpreview.com/forums/read.asp?forum=1021&message=27257971

"I also thought I should look at the photos in Imaging-Resource Comparometer.

I downloaded the D3 and D300 photos at ISO 1600, both were taken at f/8, but the D300 at 1/250 and the D3 at 1/160 (both with -0.3 bias)...exactly as I thought."

--
no text

Tested my D3 on the day it arrived, as I do all my new cameras to check for any possible 'out of spec calibration' situation; perfect.

If it'll make you feel better, have someone who knows what they're doing (you don't), like Nikon Service check it. Or continue to gnash your teeth down to your bloody gums with angst over this (Very serious) situation...

Either way, switch to decaf and perhaps obtain a prescription for Valium.

--
'Eternity is a very long time, especially towards the end.'
-Stephen Hawking

'The fox knows many things, but the hedgehog knows one big thing.' -Archilochus

'The media doesn't dictate the message. The need dictates the media.'
-Melvin Van Peebles

--
don't wait for technology -- it won't wait for you

CRH said:

If it'll make you feel better, have someone who knows what they're
doing (you don't), like Nikon Service check it. Or continue to gnash
your teeth down to your bloody gums with angst over this (Very
serious) situation...

Either way, switch to decaf and perhaps obtain a prescription for
Valium.

Click to expand...

Dear CRH,

why on earth did you react like this?

When I said "it must be my camera" that's what I honestly thought. I was not being ironic. I think the replies were very clear and conclusive to what the problem is.

Funny thing is you read my reply as if it had been written by you...ironic, over-reactive and aggressive. In the end this tells a lot about you and about me.
You know that decaf and Valium?...maybe you should consider it seriously.

I really wish you Peace on this Easter Day.