[wildman]

Pure, unadulterated, image data.

At least that is the conventional wisdom.

For all practical purposes this is the way I think of shooting RAW - see my crude schematic. It's what I, in theory, think is the final irreducible pure RAW file.

But I can't help but think it's not this simple and there is in a real world camera more factors effecting the final RAW ouput than this.

I'm not thinking about things that are intrinsic to the design like bayer filters etc because the photographer has no control over this anyway.

Any additional thoughts from the RAW gurus? - How RAW is RAW really?

---

Last edited by wildman; 09-30-2013 at 04:13 AM.

[StefanMikes]

I'd like to see the outcome of the following test: A high contrast scene (well beyond 8 EV) is taken under controlled conditions with the HC on and off. Then the RAW would be opened by an application such as RawTherapee and the histograms of both takes compared. Such a scene could be a black electronic appliance with a red or blue bright LED on. As I say, I am not in the position to make such a test.

Something tells me it is up to the manufacturer such as Pentax to handle the sensor data before they are stored as RAW. We are talking about a 14-bit sensor in the K-5. We also know it is very hard to make a photo with a dynamic range wider than 8EV, and every bit in the sensor can theoretically store 1 EV. We never get 14-bit (14 EV) DR, and many bits of the data are just used to properly render colour transitions. Now, we all know the K-5 really does something with 1 bit of the data because HC on allows using ISO160 as the lowest ISO setting. What Pentax does to that very 1 data bit must be a trade secret. Now, is it not possible that up to three top bits of the sensor data are specially handled for SC? With some trade secret of Pentax maker? (Simple thinking: "If HC is on then divide the signal from the "highlight"-bit of the sensor by two" - giving effectively 1/2 EV correction for highlights).

I can only say about own practical experience: Scenes under similar high-contrast lighting taken with the corrections off were damaged and not to be repaired with LR, and the same scenes taken with HC and SC on were perfectly manageable. I do not care how it is done in K-5; I simply wonder why it happens so.

[kh1234567890]

So I am not sure. What i am sure is the HC affects the RAW in positive manner and I have the feature on all the time. What about Shadow Compensation in RAW? Did anybody make some tests?

Unless things have changed HC does affect RAW, SC doesn't.

[StefanMikes]

Unless things have changed HC does affect RAW, SC doesn't.

Let us stick to this thought ))

[wildman]

Let us stick to this thought ))

OK...

so far then -

Must effect RAW - ISO
May effect RAW depending on configuration - HC

Anything else?

[kh1234567890]

I can only say about own practical experience: Scenes under similar high-contrast lighting taken with the corrections off were damaged and not to be repaired with LR, and the same scenes taken with HC and SC on were perfectly manageable. I do not care how it is done in K-5; I simply wonder why it happens so.

I've just had a quick play with a brand new K-5II, looking at the RAW files with RawDigger. The RAW black levels (lowest counts) were typically like this :

ISO......Black..Black(HC)
80.........64.......-
100.......64.......-
125.......64.......-
160.......64......64
200.....128......64
250.....128......64
320.....128......64
400.....256....128
500.....256....128
640.....256....128
800.....512....256
1000...512....256
1250.......0....256
1600.......0....512
2000.......0....512
2500.......0........0
...

So from ISO1250 the camera subtracts the black level internally. And at ISO800/ISO1000 you only have a five bit dynamic range (highest possible count / lowest possible count). Not that any of this makes much difference in practice, since those low counts would be noise anyway.

[adwb]

Unless things have changed HC does affect RAW, SC doesn't.

Looking at images I have taken with both hc and sc turned off and then on I can see a difference in the images with out any processing in LR at both ends. If it did not affect raw why would you be able to select them both ?