This is only the first revision. Right now it comes up "empty" and you choose the camera model(s) you want to compare. Feedback is welcome. These revision are a "work in progress".
There is also new data covering the DX Crop Mode of the FX bodies. This data is relevant because Image Quality (IQ) is reduced when you use DX Crop Mode and print at your usual FX size. This data is indicated by the DX after the camera model in the chart legend.
The good news is that the "penalty" for using DX Crop Mode is less than the usual quick calculation would predict.
Start geek alert This is not intuitive and even caught me by surprise but there is a good sound reason for it. The situation is analogous to the trade-off between raising ISO and "pushing" digitally. Usually raising ISO produces less additional noise but for some cameras, above a certain ISO, it becomes break-even. There is a similar relationship for cropping. Usually letting the camera perform the crop produces less noise than cropping later in post processing. I'm preparing an article on the subject and will post a link when it is available on my site. End geek alert
Sun 26-Feb-12 01:15 AM | edited Sun 26-Feb-12 01:20 AM by KnightPhoto
> I have updated the Photographic Dynamic Range Chart] > at my site to be interactive. >The old static chart was getting far too cluttered. > >This is only the first revision. ... > >Bill >
Fantastic Bill, it was getting to be a problem and since I rarely look at the oldest cameras anymore, the new version is much improved.
I may have commented before, but your ideal CX/DX/FX is extremely important. For me it changes the way I look at my current cameras, my purchasing decisions for newer models, and my long-term kit plans. Very very helpful! In general it has also made me more tolerant of the megapixel march, since there is less and less to be done on the low light side, especially on FX.
I had noticed one thing a while back, we seem further off our ideal DX than we are our ideal FX. Hopefully this bodes well for future DX releases!
Your comments about FX in DX mode are interesting, keep us posted.
it does not seem logical that letting a camera perform a DX crop would have less noise than doing it in post
I thought the same thing until I saw the numbers and worked it through.
As a preview, try to visualize this. We have two curves and the distance between them vertically is in the proportion that we would expect; for example 1.5:1 But the distance horizontally between the curves does *not* match our naive expectation. That would only be true if they were straight lines starting at the origin with slopes in the 1.5:1 ratio.
More to follow. Other projects like my site update regarding charts is keeping me from getting this from my head and onto paper.
Thu 01-Mar-12 08:31 AM | edited Thu 01-Mar-12 08:32 AM by richardd300
Bill, Great work. Sorry to be vague, but am I interpreting the graphs correctly please? I see that a D800 in DX mode has a better Photographic Dynamic Range than the D7000, about equal to the D700 and almost equivilent to the D3?
If correct then my reasoning is that as the D800 in DX mode produces a higher number of pixels than the D7000, exceeds the dynamic range capabilities and as good as my D700, then this may be a more capable camera noise wise for wildlife crops than the D7000 is?
I look forward to seeing your report and how it effects the issues you mention in your "Geek" report. I do realise however that nothing can equal the in user camera tests, which will be elusive until folks have ownership of the new camera
Another naive belief I'm subject to is that differences of less than 1/3 of a stop are not photographically significant. To put it another way - given a choice between a fixed crop and a variable crop at the cost of less than 1/3 stop of DR I may well choose the latter. This is particularly true as the variable crop might well be larger than the fixed DX crop, resulting in a gain in DR.
differences of less than 1/3 of a stop are not photographically significant.
I agree Although if your exposure is off by 1/3 EV it can make the difference between nailing it or clipping your highlights I don't think you see the difference in dynamic range unless you "pixel peep" the shadows.
It would be challenging to objectively find the amount where DR is visibly different !
Thu 01-Mar-12 02:38 PM | edited Thu 01-Mar-12 02:40 PM by Jim Pearce
Broadly speaking, Nikon cameras already had enough shadow recovery for wildlife photography (it's very different from landscape in that respect) since the D300. But highlight recovery is a real challenge in wildlife photography (the coot's beak!) in normal morning and late afternoon light. Marriane Oelund mentioned (in that other place) a concern that the D800 might have less capacity for highlight recovery than the D700. Can you reach any conclusions on this at say ISO 800 (corresponding to ISO 400 on DX, same shutter speed and DOF), compared to a D700 at ISO 800 (or ISO 400 in crop mode) and a D300s at ISO 400?
"Recovery" is a different topic and I'll defer to Marianne on that for now. I'm more interested in fundamental dynamic range while "recovery" implies "pushing the envelope".
FWIW, for the D700 ISO 800 FX is not equivalent to ISO 400 DX, more like ISO 500. You would read this as the horizontal difference between the two PDR curves. Note for example that for FX ISO 400 and less there is no DX equivalent ISO (at the same image quality). It makes sense that you can't get FX low ISO performance out of DX Crop Mode.
I was referring to equal DOF and shutter speed, FX vs. DX. The invariable (albeit approximate) rule is to double the ISO. However, I understand your point, that your numbers show a slight DX advantage in this range. What I'm questioning is the value of more DR for wildlife photography unless highlight recovery is improved - unless you cash it in by deliberately underexposing to protect the highlights.
I'm don't think either camera is better or worse. The curves on both FX and DX modes on both cameras follow each other and although there seems to be a small gap between curves, then perhaps this is just done so both curves are visible on the graph! Otherwise one would sit over the other.
That's my thinking, but I'm sure Bill will correct me if I'm wrong.
I don't understand why the D3 would be better than the D700
What you're seeing is that I show the actual data (unless there is an (e) after the camera then it might be an estimate from other values). So you're seeing the slight variation that reflects the accuracy of the test and sample variation.
Also, FWIW, even if the sensor is identical, the electronics of the D3 and D700 could be slightly different resulting in slightly different performance.
Bill, looking again at the curves and reading your comments, would I be correct in assuming that DX mode with the D800, is liable to better better dynamically against the D700 in DX mode? I see the D800 in DX mode beats both the D700 and D7000 although not by a lot. I would be keen on a D800, but only if the DX mode could be utilised for wildlife.
Sorry to appear a bit daft, but I am trying to understand the curves and what they may mean to me in terms of the D800 in DX mode. I use FX for landscapes, portraits, architectural (so no issues there), but if the D800 DX mode can exceed the capabilities of the D7000 and if the DX element gives me equal or improved results against my current D700/D7000 then that interests me. Do you think that on paper the D800 with it's extra crop/pixels abilities will do the job well please?
Basically, if the D800 can equal my D7000 in terms of the photographic dynamic range in DX mode, then it will open up many previously unavailable opportunities all in one camera. If not much would be gained then I'll wait for a d300s replacement which was my original intention.
You are not "daft" at all. I have not been clear and will be revising the article.
Regarding the curves. Yes, it looks like the D800 in DX Crop Mode will be about equivalent to the D700 in FX and ahead of the D7000. But it's not a cheap decision and waiting to see what the D400 brings might be a reasonable unless you're looking to get into FX and see DX Crop Mode as a big plus.
There is an issue separate from dynamic range regarding the D800. Will the lower diffraction limit have any practical effect on your images? I suspect this will depend significantly on print size (and viewing distance) as well as subject matter.
Thanks Bill. That helps clarify my thinking somewhat. I intend to do nothing until the D300s replacement is announced in fear of making a serious and expensive blunder! However, your comment regarding "the D800 in DX Crop Mode will be about equivalent to the D700 in FX and ahead of the D7000" is extremely interesting. Mostly because if the DX crop on the D800 is equivilent to the D700 in FX, I'd certainly settle for that.
I've used my D700 for wildlife with higher ISO settings in FX when I can get closer, in other words making up for the x1.5 APS-C crop distance, then conducting crops with great results noise wise. Far better results than with the D7000 at the same ISO's x1.5 distance further away. If that makes any sense
Interesting stuff and valuable too, better to talk about it here and make a more informed decision.
Awesome! Thanks for the hard work. I just compared my D200 to a D800 and saw the D800 has over 1.5x better dynamic range than my D200.
At ISO 400:
D200 - 6.11 D800 - 9.39 estimated
Even better is that I find ISO 400 about as high as I can do the D200 and like it. For the same dynamic range, the D800 can be at an estimated ISO 3200! I believe that is a 3 stop improvement at a minimum!
Shoot nature with respect and don't trample it or startle its inhabitants. :)
To compare camera models the best comparison is to hold quality steady and look at the horizontal differences. To aid in this I drew a line at 6.5 on the y-axis and came up with this separate chart: ISO at PDR of 6.5EV Using this chart the advantage would be log2(3832/303)=3.66 rather than 9.39-6.11=3.28 vertically
Either way a D800 blows a D200 away. (This should not surprise anyone!)
Bill, Thanks for all the info. I am curious as to what the difference is between your DR # and those at Sensorgen. I believe that Sensorgen gets their data from DXO. Are your figures based on results from the camera as opposed to testing a sensor's capabilities.
Marc There are always two people in every picture: the photographer and the viewer.-Ansel Adams
I'm not sure what you mean by camera as opposed to sensor. Everyone uses the camera in order to test the sensor. Sometimes results are stated for a particular camera & lens combination. Usually that's done for lens testing rather than sensor testing.
Sensorgen definitely sets their data from DxOMark. I do something similar with DxOMark data but don't publish my results.
It's unclear to me whether DxOMark actually collects "good" data. On paper they have very good "protocols". But in truth the data they expose at the web site is not suitable for sensor analysis. Both Sensorgen and myself often get "screwy" numbers when using DxOMark data. (FWIW I have made inquiries at DxOMark but they go without reply.)
If you were going to compare Photographic Dynamic Range (PDR) with anything it would be with the DxOMark Landscape Dynamic Range. They compare quite favorably only they are on different scales.
The Sensorgen numbers are "engineering" dynamic range. So even if I trusted them they would not be right for comparing to PDR.