k20d Expanded Dynamic Range opinions - Page 6

Confused · 05-30-2008, 07:15 PM

Hi again lbam

I reckon my slightly warped sense of black humour finally got the better of me there.....although I'm sure I sensed Darth Vader murmuring something like "Beware the Dark Side", so in hindsight I guess he wasn't referring to Histograms.......lol !

Best regards
Richard

photog · 05-31-2008, 03:29 AM

I have been reading this post with quite sime interest and not a little confusion!
I am considering updating my K10D to the K20D. The benefit I am most anxious to gain is an increase in dynamis range. The K10 is very prone to have blown highlights when exposing to the right (Manual always) when subjects include highly saturated reds - yellows. I have been somewhat perturbed to note that a number of reputable test sites test the dynamic range of the K20 to be LESS than the K10; and this despite the change to CMOS sensor!
My question is: Has anyone carried out accurately measured tests and PRINTED from them?
re some of the posters I would mention that the only way to gusrantee that you have increased the exposure by one f-stop stop is to simply use the next f-stop. I am a (fairly) experienced cs3 user and ACR4.1 and I would not bet any money on the 1 fstop calibration on the exposure slider being a precise one stop change. The gamma of the image can increase / decrease the exposure change as read in ACR. and btw when you change brightness of the image you are actually changing midrange gamma.
Can anyone say that photographing a display of bright yellow, orange and red flowers in brilliant sunshing you can hold the RGB layers short of blowing. I have found that, even if undersposing 1-2 stops the extremely saturated colours mentioned can lose information.
HopeI havent been too tedious,
Ron McDermott (Sunny Queensland)
PS Actually drought stricked!

*falconeye* · 05-31-2008, 04:11 AM

Originally Posted by photog

a number of reputable test sites test the dynamic range of the K20 to be LESS than the K10; [...] I would not bet any money on the 1 fstop calibration on the exposure slider being a precise one stop change. [...]
Can anyone say that photographing a display of bright yellow, orange and red flowers in brilliant sunshing you can hold the RGB layers short of blowing.

Your questions are truely valid but difficult to answer, actually...

First, the dynamic range tests that I know of are done in JPG (8 Bit output anyway) and are nothing more than a test of non-linearity of the image. Personally, I consider 8 to be the ideal value here (exact linear curve with precise clipping). Software tricks like K20D's D-Range increase it at will.

LightRoom 'exposure' slider: Adobe says it affects the bright colors more than the shadows...

For the reds... I didn't notice any particular problem here. I attach two images with red in it, one done at normal exposure, the second at -1EV correction, both from RAW, with LR default settings. The first has minimal clipping. To really answer your question, you must define a test to be carried out...

flyer · 05-31-2008, 05:56 AM

There is one thing bugging me a little bit here. Some people bought the 20 due to the specs they saw advertised, and then some other people ask them to test this, then that and so on for them. I don't mind helping other people, and I know it is an expensive piece of equipment, but, at some point those people will have to take a decision. It seems (to me, at least) that some (NOT that many, I am not trying to generalize) people keep asking for one test or another (in quite a few posting), but never seem to make a decision wether they should invest or not. There are already quite a few test that have been done to help people take a stand as to decide to buy or not. Discussing a topic to figure out how it works and how it can improve our ability to take pictures is fine, but it seems (to me) that some members are just asking test for the sake of asking.

*isteve · 05-31-2008, 08:38 AM

Originally Posted by PentaxPoke

I'm still at a loss to see the 200%. Wouldn't a 200% increase mean a 2-stop increase in DR? There is not even a 1-stop increase as far as I can tell, on either the high or the low end. If EDR were going to show itself, it would have done so here. The effect is very subtle, bordering on unobserveable.
Comments?

No - its not a 200% (3X) increase, its 200% of the original (2X) increase hence a 1 stop increase in highlight range.

It works quite simply by underexposing by 1 stop (it actually uses a sensitivity of half the stated amount eg. ISO 200 is actually shot at ISO 100 with the exposure settings for ISO 200) and then using a processing algorithm than applies zero boost to the brightest tones then progressively applies up to 1EV boost for the darkest tones giving you a 1EV "compression". You will not see more shadow detail this way because you are actually compressing the mid tone range into fewer stops. What you will get is a camera with an extra 1EV of highlight range - great for wedding dresses and avoiding highlights on shiny skin - but with more compressed mid-tones hence reduced contrast.

lbam · 05-31-2008, 08:45 AM

Originally Posted by Confused

I reckon my slightly warped sense of black humour finally got the better of me there.....although I'm sure I sensed Darth Vader murmuring something like "Beware the Dark Side", so in hindsight I guess he wasn't referring to Histograms.......lol !

No worries mate, I'm actually a Brit myself so am questioned on my sense of "humour" at times.
I wonder if the EDR is pushing my pics towards the dark side, or if it's merely just crushing the (high)light side. Vader would be impressed either way...

Myself, I don't know whether to be impressed or not to be honest.

*falconeye* · 05-31-2008, 08:48 AM

Originally Posted by flyer

it seems (to me) that some members are just asking test for the sake of asking.

I agree with the statement in general.

photog's question, however, seemed novel (with its emphasis on red clipping), he didn't post too many questions before, and the thread discussed red clipping already). So, maybe exclude him from the general critics

*falconeye* · 05-31-2008, 08:56 AM

Originally Posted by *isteve

applies zero boost to the brightest tones then progressively applies up to 1EV boost for the darkest tones giving you a 1EV "compression".

@*isteve, I completely agree.

My gut feeling is that the 1EV compression isn't equally distributed over the available tone range. It seems to be pretty strong in the highlights and quite mild in the rest of the tone range -- so that the contrast impression doesn't change much, either.

flyer · 05-31-2008, 10:43 AM

Originally Posted by falconeye

I agree with the statement in general.

photog's question, however, seemed novel (with its emphasis on red clipping), he didn't post too many questions before, and the thread discussed red clipping already). So, maybe exclude him from the general critics

I wasn't referring to him. It comes from way further than this thread.

Ben_Edict · 06-01-2008, 01:32 PM

Originally Posted by *isteve

No - its not a 200% (3X) increase, its 200% of the original (2X) increase hence a 1 stop increase in highlight range.

It works quite simply by underexposing by 1 stop (it actually uses a sensitivity of half the stated amount eg. ISO 200 is actually shot at ISO 100 with the exposure settings for ISO 200) and then using a processing algorithm than applies zero boost to the brightest tones then progressively applies up to 1EV boost for the darkest tones giving you a 1EV "compression". You will not see more shadow detail this way because you are actually compressing the mid tone range into fewer stops. What you will get is a camera with an extra 1EV of highlight range - great for wedding dresses and avoiding highlights on shiny skin - but with more compressed mid-tones hence reduced contrast.

Steve, this is exactly how EDR works and as far as I remember, Pentax has also publicly explained it that way. I made some comparisson shots with EDR and without it and finally decided to skip this feature. The problem is just shifted, because it is very hard to get a good mid-tone contrast with EDR images. Yes the highlights tend to be within the range and not be blown, but a EDR's shadow shows more noise and I can have that with post-processing, if I wish. But the mid-tones (depending on the image) can be a real problem with EDR. May be with weddings it's not much of a problem, but with landscapes I definitely find it is - and as such won't use EDR anymore...

Ben

DAZ · 06-25-2008, 02:50 PM

Is this the answer?

Pentax K20D Review: 21. Photographic tests (DR):

Aaa what was the question?

DAZ

*jbinpg* · 06-25-2008, 09:26 PM

Here is my "garden" test, a white petunia under a heavy overcast sky so no ambient light changes between exposures. Both exposures at 1/60s,f11 but 2nd with 200% DR applied, taken one right after the other. SMCP-M 100mm f4 with 2x doubler. The taming of the blownout highlights with DR application is quite evident. I use it for wedding dress shots, for sure.

Jack