Here's a test of a street lamp to see how much DR I could get without tone curve/contrast/blackpoint manipulation. I was interested in how much effect flare & glare has. In my case, the answer is "a lot".
Pentax K-5 & SMC Pentax-M 50/F1.7 at 1/500s, f/4.5, ISO80 (base sensitivity). 14-bit raw.
Sensel counts for each raw stop (using the RAWSHACK program):
Stop #/Range | Red | Green | Blue | Green_2 |
--------------- | ---------- | ---------- | ---------- | ---------- |
0 [00000-00001] | 0 | 0 | 0 | 0 |
1 [00002-00003] | 0 | 0 | 0 | 0 |
2 [00004-00007] | 0 | 0 | 0 | 0 |
3 [00008-00015] | 0 | 0 | 0 | 0 |
4 [00016-00031] | 0 | 0 | 0 | 0 |
5 [00032-00063] | 1,776 | 1,915 | 1,809 | 1,919 |
6 [00064-00127] | 4,060,809 | 4,060,498 | 4,060,736 | 4,060,492 |
7 [00128-00255] | 766 | 181 | 719 | 182 |
8 [00256-00511] | 493 | 659 | 552 | 662 |
9 [00512-01023] | 31 | 477 | 52 | 477 |
10 [01024-02047] | 20 | 137 | 23 | 129 |
11 [02048-04095] | 23 | 17 | 24 | 25 |
12 [04096-08191] | 32 | 26 | 35 | 23 |
13 [08192-16383] | 0 | 40 | 0 | 41 |
I kept the shutter speed very high (for night-time) to keep the top stop sparsely populated to reduce the possibility of any sensor blooming-induced glare, so the image appears quite dull. (This is mainly a problem with CCD sensors, usually in older video cameras where you'll see phantom vertical lines centred on lights in the frame where the whole column has been affected, but it can effect CMOS slightly too, but very localised. See
http://micro.magnet.fsu.edu/primer/digitalimaging/concepts/ccdsatandblooming.html)
There are 9 stops DR in total, but with such a vast amount of darkish stuff at the 8th stop, I think you'd have to call this 8 stops of DR.
Raw processing in RT4 starting with the Neutral profile:
Completely neutral (no NR, EV boost etc)
100% crops (all starting with the Neutral profile in RT4):
Completely neutral
Boosted +5EV (blown highlights). Chroma & luminance noise is now visible, but no pattern noise.
Now with highlight-recovery applied to keep the boosted highlights below the clipping point.
Finally, with NR applied. The glare is quite obvious.
I was hoping to achieve a wider DR, but that was not the case here. I don't have any modern primes (the lens under test is about 30 years old). It's possible the aged inter-element lens cement might increase glare & flare, so a modern prime with modern coatings may do better.
Dan.
Last edited by dosdan; 10-15-2011 at 02:07 PM.