D200 vs D70 Dynamic Range graph

Started Jan 30, 2006 | Discussions
Greg Matty Senior Member • Posts: 1,386
D2X By Chance?

I don't suppose you have run one of those on the D2X by chance?

I would love to see how my D2X fairs as there are times I am not happy with the DR.

Greg

OP davexl Senior Member • Posts: 1,058
Re: Wow, lotta data there.

Bill Janes wrote:

Yes, the thread has been productive. Have you seen the post by
Photogentic? He uses a method similar to yours, and I suggested to
him that he should do a log(pixel) plot to compare to H&D curves.
Unfortunately, he has not been active lately since he is a Naval
Officer at sea. BTW Ansel Adams used similar tests with negative
films and prints many years ago.

http://fotogenetic.dearingfilm.com/custom_tone_curves.html

Yeah, he helped us all out a few years ago with the D70 curves.

(fotognetic, not Ansel

Now what Ansel would do with a D200 and some stitching software - I ask myself that everyday!

-- hide signature --

Regards,

David F.
Melbourne, Australia.
http://www.pbase.com/davexl/folio

 davexl's gear list:davexl's gear list
Nikon D810 Nikon AF-S Nikkor 14-24mm f/2.8G ED Nikon AF-S Nikkor 24-70mm f/2.8G ED Nikon AF-S Nikkor 85mm f/1.4G Nikon AF-S Nikkor 50mm f/1.8G +3 more
Bill Janes Senior Member • Posts: 1,848
Re: Imitest D200 vs David F's Method

Here is a graphical comparison of the Imitest plot versus the plot that David F posted for his D200. NC was used to process the NEF with normal contrast to aRGB for the Imitest procedure.

In the Imitest plot the value for Zone M on the Kodak card (19.95% reflectance, 0.7 optical density) is 132 and David's nominal exposure reads 126 (presumably from an 18% gray card). Zone M is listed as 7 in Imitest. The higher tone values are about the same in both plots, but the shadow values are dramatically different. I conclude that the the rendering is different in David's exposures for low tone values. Perhaps Julia or another expert can comment.

Since I don't have a complete set of values for David's D200, Excel will not connect the markers on the graph where there are gaps in the data points.

-- hide signature --

Bill Janes

Julia Borg Veteran Member • Posts: 7,280
Re: Imitest D200 vs David F's Method

Bill, may I suggest something. check with spotmeter that the wedge lit as you do it has 1/3 eV step difference between rectangles.
--
Julia

OP davexl Senior Member • Posts: 1,058
Re: Imitest D200 vs David F's Method

Hi Bill - what tone curve have you go the camera set to? Normal or Auto?

And spotmetering those shadows would tell us a lot about the dynamic range of the target.

When does your transmissive test target arrive?

-- hide signature --

Regards,

David F.
Melbourne, Australia.
http://www.pbase.com/davexl/folio

 davexl's gear list:davexl's gear list
Nikon D810 Nikon AF-S Nikkor 14-24mm f/2.8G ED Nikon AF-S Nikkor 24-70mm f/2.8G ED Nikon AF-S Nikkor 85mm f/1.4G Nikon AF-S Nikkor 50mm f/1.8G +3 more
Julia Borg Veteran Member • Posts: 7,280
Re: Imitest D200 vs David F's Method

And spotmetering those shadows would tell us a lot about the
dynamic range of the target.

we know it. we do not know glare.

-- hide signature --

Julia

cuonghuutran Contributing Member • Posts: 513
Re: Confirms my thoughts and findings

If D70 has the modified curve then does D200 still have more useful dynamic range (leaving aside the noise issue when boosting the midtone)?

davexl wrote:
It DOES have more useful dynamic range.

The total is about the same, but a D200 shot will look brighter in
the midtones/shadows. The very deepest shadows are about the same,
but that is less noticable compared to the gains across most of the
tonal range.

Don't take this as a mythbusting finding - it proves what people
have been seeing - you will see more shadow detail with a D200 than
a D70.

SergioRZ wrote:

There are so many users reporting how much better Dynamic Range on
the D200 than on the D70 or the Fuji S3... Some with pretty good
tests it seems...

Internet is just like that

OP davexl Senior Member • Posts: 1,058
Re: Imitest D200 vs David F's Method

Julia Borg wrote:

And spotmetering those shadows would tell us a lot about the
dynamic range of the target.

we know it. we do not know glare.

Sigh.

Which means in fact we DON'T know the effective DR of the target, which is what I meant - and metering would shed some light on that - pun intended

-- hide signature --

Regards,

David F.
Melbourne, Australia.
http://www.pbase.com/davexl/folio

 davexl's gear list:davexl's gear list
Nikon D810 Nikon AF-S Nikkor 14-24mm f/2.8G ED Nikon AF-S Nikkor 24-70mm f/2.8G ED Nikon AF-S Nikkor 85mm f/1.4G Nikon AF-S Nikkor 50mm f/1.8G +3 more
drejc Senior Member • Posts: 2,075
better DR and noise perfor. on D70s???

from samples on this page(move to page 6):
http://www.e-fotografija.com/artman/publish/article_952.shtml

looks like that D70s have better DR than D200. But that is just one sample. Also noise is much lower on D70s.

OP davexl Senior Member • Posts: 1,058
Re: better DR and noise perfor. on D70s???

Got an English page?

Do they compare noise of D200 downsampled to 6MP? - because when comparing systems, the D200 hypothetically could be worse than the D70 on a pixel level, but printed the same size you would still be better off due to the extra MP.

-- hide signature --

Regards,

David F.
Melbourne, Australia.
http://www.pbase.com/davexl/folio

 davexl's gear list:davexl's gear list
Nikon D810 Nikon AF-S Nikkor 14-24mm f/2.8G ED Nikon AF-S Nikkor 24-70mm f/2.8G ED Nikon AF-S Nikkor 85mm f/1.4G Nikon AF-S Nikkor 50mm f/1.8G +3 more
OP davexl Senior Member • Posts: 1,058
Re: Confirms my thoughts and findings

Yes and no. Yes similar "look", but will be noisier.

Dynamic range of the camera is really how much noise you can stand by boosting shadows.

But there are a lot of subtleties that have been improved in D200 IQ - DR is one of them - they all add up to a much nicer result.

cuonghuutran wrote:

If D70 has the modified curve then does D200 still have more useful
dynamic range (leaving aside the noise issue when boosting the
midtone)?

-- hide signature --

Regards,

David F.
Melbourne, Australia.
http://www.pbase.com/davexl/folio

 davexl's gear list:davexl's gear list
Nikon D810 Nikon AF-S Nikkor 14-24mm f/2.8G ED Nikon AF-S Nikkor 24-70mm f/2.8G ED Nikon AF-S Nikkor 85mm f/1.4G Nikon AF-S Nikkor 50mm f/1.8G +3 more
Bill Janes Senior Member • Posts: 1,848
Re: Imitest D200 vs David F's Method

David & Julia,

I repeated the Q14 test and took spot meter (1 degree) readings from the target. I found that the values below B were not reproducalbe and could not be distinguished by the spot meter, ? glare or deterioration of the target from smudging, etc. I repeated the Imitest with the D70 and no longer got a straight line as before. Therefore, the current Imitest results are not valid for the lower tone values.

However, on looking at David's data, I see the low tones fall off more rapidly than they should. This table shows David's results with the raw values supplied Zkruk also shown (the average of the two green values is used).

The RAW file is linear as it should be. Each halving of exposure halves the pixel level. At EV -5 the raw pixel value is 20 and I extended it further to EV -9.

The column 2.2 corr shows the gamma 2.2 corrected pixel values for the given raw value (similar to what Julia showed on her web site for the step wedge,
http://www.pochtar.com/gamut_view/gamma.xls ).

At EV -4 the calculated pixel value should be 30 (Julia gets 34.91--she may be using sRGB with an offset rather than aRGB), whereas David gets a value of 6.

AT EV -5 David's value is 1 pixel, whereas the calculated value should be 23 using a gamma of 2.2 with no offset as in aRGB. One has to go to EV -9 to get the predicted raw pixel value down to 1, corresponding to a pixel value of 6 in the gamma 2.2 space.

David's results track with the gamma 2.2 results until EV -2 (linear exposure 0.25), and then they drop sharply from predicted. This is well shown in the graph, which is log-log. As one can see, this type of plot linearizes the plot (also shown in the Imitest plots).

I don't know what is going on, but it may be that the camera or converter is underflowing with 8 bit calculations. It might be advisable to repeat the process using 16 bit conversions in Nikon Capture or ACR.

I will have to repeat my Imitest plots when the step-wedge comes. It should have been here by now.

Table:

graphical:

-- hide signature --

Bill Janes

drejc Senior Member • Posts: 2,075
Re: better DR and noise perfor. on D70s???

English translate will be in a few days I think.

No I think that D200 picture is not downsampled to 6MP. It is 10mpix with crop and 6mpix with crop.

I also think that they should downsampled D200 picture to the same size as D70s picture is and then make a crop. There should be visible real difference.

davexl wrote:

Got an English page?

Do they compare noise of D200 downsampled to 6MP? - because when
comparing systems, the D200 hypothetically could be worse than the
D70 on a pixel level, but printed the same size you would still be
better off due to the extra MP.

Steve Bingham
Steve Bingham Forum Pro • Posts: 26,175
Dave

Now THIS takes me back!!! B&W Voodoo.
" pre-fogging film in the old days to boost shadow response"

-- hide signature --
 Steve Bingham's gear list:Steve Bingham's gear list
Nikon D810 Nikon D7200 Nikon AF Nikkor 85mm f/1.8D Nikon AF-S DX Nikkor 18-140mm F3.5-5.6G ED VR Nikon AF-S DX Nikkor 16-80mm F2.8-4E ED VR +20 more
Steve Bingham
Steve Bingham Forum Pro • Posts: 26,175
Bill

Am I interpreting your results correctly as saying the D200 has a theoritical maximum DR of 12 stops in raw? The Fuji S2 stopped at 11 (and I thought that was awesome). Of course real world would depend on your chozen noise floor (noise figure). What say?

Steve Bingham
http://www.dustylens.com

 Steve Bingham's gear list:Steve Bingham's gear list
Nikon D810 Nikon D7200 Nikon AF Nikkor 85mm f/1.8D Nikon AF-S DX Nikkor 18-140mm F3.5-5.6G ED VR Nikon AF-S DX Nikkor 16-80mm F2.8-4E ED VR +20 more
Bill Janes Senior Member • Posts: 1,848
Re: Bill

Steve Bingham wrote:

Am I interpreting your results correctly as saying the D200 has a
theoritical maximum DR of 12 stops in raw? The Fuji S2 stopped at
11 (and I thought that was awesome). Of course real world would
depend on your chozen noise floor (noise figure). What say?

Steve Bingham
http://www.dustylens.com

Steve,

I had not intended to make such a claim. Here are some results from Imitest with the Stouffer target: 12.3 stops total. Of course the noise floor could be limiting. I need to read the Imitest documentation some more.

-- hide signature --

Bill Janes

Bill Janes Senior Member • Posts: 1,848
Re: Problems resolved

davexl wrote:

Hi Bill - what tone curve have you go the camera set to? Normal or
Auto?

Normal

And spotmetering those shadows would tell us a lot about the
dynamic range of the target.

When does your transmissive test target arrive?

-- hide signature --

Regards,

David F.
Melbourne, Australia.
http://www.pbase.com/davexl/folio

The transmissive target arrived and further testing has resolved the previously noted differences between my results with Imitest and David's characteristic curve from multiple exposures of a gray card. Actually, they give the same results if you use the normal Nikon tone curve for both tests.

The Nikon curve rolls off the darker values toward zero much more rapidly than Adobe Camera Raw or a literally applied gamma of 2.2

Here is the plot of the RAW file with no gamma correction tone curve. The plot is linear with some flattening in the shadows consistent with flare in the lens and camera box:

Here is the Imitest plot of the Stouffer target with Nikon Capture using the normal contrast setting.

And here is the plot of the same image converted with ACR using default settings. Note that there is much less roll of of the darker tones.

Finally, here is an Excel chart showing all of the above results together in a log-log plot similar to Imitest's. For simplicity all pixel values are expressed in 8 bit (0..255) units. David's results are presented as yellow triangles and closely track the Nikon Capture conversion. Also shown is a literal application of gamma 2.2 to the raw data. It is necessary to roll off the shadows at some point, each EV halves the exposure and it would take 12 halvings to go from a raw pixel value of 4096 to 1 and a print does not have that many zones.

Here is the same graph with linear-log scaling (similar to what David showed). IMHO, the relationships are much better appreciated in the log-log plot, which is probably why Imitest uses it.

As Julia correctly stated: "David, that in no way characterizes cameras. What you presented is dynamic range of your preferred processing of the images." The converter software has a profound influence on the results.

-- hide signature --

Bill Janes

Julia Borg Veteran Member • Posts: 7,280
Re: Problems resolved

Hi Bill,

thank you for taking the trouble....

may I ask what was used for lighting?

the wedge you used - is it of calibrated batch?

raw data - was it from internal processing in Imatest?

your results are very consistent with mine, except for shadows. I have somewhat better linearity there.

-- hide signature --

Julia

Bill Janes Senior Member • Posts: 1,848
Re: Problems resolved

Julia Borg wrote:

Hi Bill,

thank you for taking the trouble....

may I ask what was used for lighting?

3200K quarz lights

the wedge you used - is it of calibrated batch?

No, it is not calibrated. For my personal use I did not think the extra expense of calibration would be worth it.

raw data - was it from internal processing in Imatest?

Yes, I used the built in Imitest raw processing. I also used DCRaw with the -h -m -n -3 switches as you suggested and substituted the green channel into the R and B channels. However, the maximum density was supoptimal as shown in the Imitest results below. With 0.3EV more exposure, the second step was clipped and could no longer be distinguished from the first

your results are very consistent with mine, except for shadows. I
have somewhat better linearity there.

Do you think it is flare or stray light affecting the shadows?

-- hide signature --

Bill Janes

Julia Borg Veteran Member • Posts: 7,280
Re: Problems resolved

may I ask what was used for lighting?

3200K quarz lights

any diffusers? was the wedge placed into the window in a black opaque plate?

I also used DCRaw
with the -h -m -n -3 switches as you suggested and substituted the
green channel into the R and B channels. However, the maximum
density was supoptimal

can you send me the nef?

Do you think it is flare or stray light affecting the shadows?

looks like stray light. eliminating it you will probably get more "high qality" steps resolved. do not know if it is worth any trouble as essentially the questions are already clearly answered.

-- hide signature --

Julia

Keyboard shortcuts:
FForum MMy threads