[BruceBanner]

Recently I have been curious about the Dynamic Range Settings in my K-1 and KP. From what I gather then Highlight Protection is actually baked into the RAW. Typically I have every piece of additional processing in the camera turned off, but I thought I might explore this feature more thoroughly. I thought I'd share my findings with you and hopefully spur some further comments.

The Test.

I chose a scene of high dynamic range, a blue sky sunny aussie winter's day, dark shadows on lawn, a white car reflecting sun, yup, very dynamic!

I did a series of comparisons and have paired them together so hopefully you can see them better and what's happening in the shots. The EXIF is in the data of each shot. Essentially I took 5 images of the same scene, they consisted of;

1) An image with no EV bias or any D-Range option on.

2) An image with -0.7 EV bias, no D-Range on (this is what I typically have my settings at, a small effort to preserve highlights).

3) An image with -0.7 EV bias with D Range settings on (set to max).

4) An image with -1.0 EV bias with D Range settings on (set to max).

5) An image with -1.3 EV bias with D Range settings on (set to max).

I shot each image RAW, but then used the in camera RAW Development of my K-1 to generate Jpgs of them without further processing. The jpgs have not been edited at all, just imported into photoshop to do the comparisons.

I am curious of how much the RAW's differ from these Jpgs and am still to process that part.

My initial feelings are that 0.0 EV is not ideal for dynamic range, red blinkie highlights are very plentiful. Perhaps I am being overly critical because these are only jpg red blinkies firing and not RAW ones, but it always made me under expose by 2/3rds of a stop (-0.7EV) for most of my shooting. What I cannot stand is someones portrait I have shot in natural light to have a highlight blown on a forehead or cheek hence my default go to value of underexposing every shot (at least as a starting point) of -0.7EV.

The last comparison (Image 5) where I show my current -0.7 EV with no D-Range on vs a shot with D-Range On and with an additional 1/3rd stop (-1.0 EV) actually makes me feel this is a better starting point. To me this looks better.

Thoughts?

The only downside as I can see it is shooting at ISO 200, but I think that's an ok trade off.

Colours look a bit better as well on Image 5 final frame, the -1.0 EV and D-Range on. Does dropping a 1/3 stop assist with saturation and colours?

Image 1


Image 2


Image 3


Image 4


Image 5

[Larrymc]

Shadow correction seems to affect RAW files in a positive way, at least to my eye. I just shot a couple of shots in my living room with my K-3 and the difference was pretty dramatic in FRV. I don't see the advantage of Highlight correction though.

[BruceBanner]

Shadow correction seems to affect RAW files in a positive way, at least to my eye. I just shot a couple of shots in my living room with my K-3 and the difference was pretty dramatic in FRV. I don't see the advantage of Highlight correction though.

I have just loaded two of the RAW files into LR.

The two files are the images contained in Image 2 in this thread.

Photo one.
Shutter speed 1/800
f5.6
ISO 200
EV Bias -0.7
HL Correction OFF
Shadow Correction OFF

Photo two.
Shutter speed 1/800
f5.6
ISO 200
EV Bias -0.7
HL Correction ON
Shadow Correction Max

As I am looking at them in Develop module and have them side by side in Reference mode, there is a definite difference between the two RAW files despite having the exact same exposure variables.
When I max out the shadow and highlight bars (highlights -100, shadows + 100) the difference is still apparent.
In fact... to match the images closely to being the same the slider values are;

Photo One (no HL Correction or Shadow Correction On)
Highlights -100
Shadows +100

Photo Two (HL & Shadow Correction On)
Highlights -78
Shadows +87

These values now make the photos have a similar dynamic range look, shadows being matched and car at same brightness/details (though I'm eyeballing it here).
So if all this stuff is nonsense for RAW then how come two RAW files with the exact same Exposure triangle variables are giving the LR user different range of scope for playing with the dynamic range sliders.

(PS happy to provide the RAW files for download and inspection for anyone who wants them).

[biz-engineer]

My initial feelings are that 0.0 EV is not ideal for dynamic range

The dynamic range is the same , regardless of the exposure compensation. Negative exposure compensation favors high light details, while positive exposure compensation favors shadow details.

Shadow correction seems to affect RAW files in a positive way, at least to my eye.

Shadow compensation doesn't affect raw files, it affect the JPEG preview only. High light protection on the K1 does compress the high lights of the raw files.

---------- Post added 23-07-19 at 07:47 ----------

I don't see the advantage of Highlight correction though.

Here we have a diverging opinion. Looking close, high-light protection on the K1 does work pretty well for me. The camera AE sensor will detect if an area of the images will be clipped, and if yes, the camera underexpose by 1 stop and apply a ton compression curve to the raw file, this can save a lot of files that wouldn't otherwise be recoverable in post. I personally find high-light protection more useful than shadow correction.

[Trickortreat]

Turning the highlight protection to auto and leaving shadow protection off works best for most of the images.

[BruceBanner]

Shadow correction seems to affect RAW files in a positive way, at least to my eye. I just shot a couple of shots in my living room with my K-3 and the difference was pretty dramatic in FRV. I don't see the advantage of Highlight correction though.

There seems to be some difference in opinion of whether Shadow Correction is baked in RAW. Everyone seems agreeable that Hightlight Correction is, but some say Shadow is and isn't.

With my examples above I can report that there is definitely a difference between files shot with and without D-Range settings on (even in RAW), however I always had HL Correction on along with Shadow Correction, so the difference may purely be just HL and nothing to do with Shadow...

However,

I'll say it again, look at Image 2;

"The two files are the images contained in Image 2 in this thread.

Photo one.
Shutter speed 1/800
f5.6
ISO 200
EV Bias -0.7
HL Correction OFF
Shadow Correction OFF

Photo two.
Shutter speed 1/800
f5.6
ISO 200
EV Bias -0.7
HL Correction ON
Shadow Correction Max

As I am looking at them in Develop module and have them side by side in Reference mode, there is a definite difference between the two RAW files despite having the exact same exposure variables.
When I max out the shadow and highlight bars (highlights -100, shadows + 100) the difference is still apparent.
In fact... to match the images closely to being the same the slider values are;

Photo One (no HL Correction or Shadow Correction On)
Highlights -100
Shadows +100

Photo Two (HL & Shadow Correction On)
Highlights -78
Shadows +87

These values now make the photos have a similar dynamic range look, shadows being matched and car at same brightness/details (though I'm eyeballing it here).
So if all this stuff is nonsense for RAW then how come two RAW files with the exact same Exposure triangle variables are giving the LR user different range of scope for playing with the dynamic range sliders.

(PS happy to provide the RAW files for download and inspection for anyone who wants them)."

I am seeing more dynamic range to play with in LR with the file that had both HL Correction & Shadow Correction on for sure, the question for me was that purely to do with HL Correction only. I wouldn't have thought so because I seem to have more room to move with the Shadows.

The dynamic range is the same , regardless of the exposure compensation. Negative exposure compensation favors high light details, while positive exposure compensation favors shadow details.


Shadow compensation doesn't affect raw files, it affect the JPEG preview only. High light protection on the K1 does compress the high lights of the raw files.

---------- Post added 23-07-19 at 07:47 ----------


Here we have a diverging opinion. Looking close, high-light protection on the K1 does work pretty well for me. The camera AE sensor will detect if an area of the images will be clipped, and if yes, the camera underexpose by 1 stop and apply a ton compression curve to the raw file, this can save a lot of files that wouldn't otherwise be recoverable in post. I personally find high-light protection more useful than shadow correction.

Turning the highlight protection to auto and leaving shadow protection off works best for most of the images.

Sorry, yes I mean I get that, I didn't explain myself properly.

When the camera is looking at a scene with very strong dynamic light, to use 0.0 EV Bias I think is not the best as Highlights will definitely be blown. Lowering to something like -0.7 EV Bias will protect the HL more and although the exposure will be lower and shadows darker they are still not blown neither (and very recoverable in post). On a cloudy overcast day whereby lighting is not so dynamic but more uniformed then shooting at -0.7 EV Bias may be overkill and unnecessary and an EV Bias set back to 0.0 may be fine.
I'm concerned with a setting for difficult high dynamics that will work the best, treading that fine line of getting as much shadow detail as possible without blowing Highlights.


If Shadow compensation doesn't affect RAWs then why in my second post whereby I compare a shot with the same settings but one that has HL Correction on and Shadow on seems to have more dynamic room to move in post (please see above). Not only do I get more Highlight breathing space but also Shadows.
Larrymc is not the first member to perceive Shadow Correction to be baked into RAW.

[stevebrot]

Thoughts?

Exposure compensation merely changes the actual stop-wise exposure with no modification of the sensor response curves, otherwise, the results would not be consistent. You can demonstrate by using EC in M mode vs. simply dialing in the compensation manually where the camera is ignorant of intent.

If Shadow compensation doesn't affect RAWs then why in my second post whereby I compare a shot with the same settings but one that has HL Correction on and Shadow on seems to have more dynamic room to move in post (please see above). Not only do I get more Highlight breathing space but also Shadows.

Not my experience, though there may be a difference in what is considered "breathing space". I don't have time today (heading to the mountains to escape the heat and get alpine flower photos), but can do some work in RawDigger later to see if there are any histogram changes in the sensor data as written with those two settings.


Steve

[swanlefitte]

Isn't
1/800
F/5.6
-0.7 ec
Iso 200
Just another form of saying 1/800
F/5.6
Iso something like 133? I thought it just told you how you got there.

I was thinking it was like a+a+a is the same as 3a.

I am also under the impression the RAW is not baked different but the preview can be. It could be like a Rubik cube that can't get back to the same point. I would need to check every setting in ACR under every panel seen and hidden before I would assume the RAW is baked. I mean if the yellow luminance was dropped to -3 things can change and I don't know the Pentax algorithms and their variance.

[Larrymc]

There seems to be some difference in opinion of whether Shadow Correction is baked in RAW. Everyone seems agreeable that Hightlight Correction is, but some say Shadow is and isn't.

With my examples above I can report that there is definitely a difference between files shot with and without D-Range settings on (even in RAW), however I always had HL Correction on along with Shadow Correction, so the difference may purely be just HL and nothing to do with Shadow...

However,

I'll say it again, look at Image 2;

"The two files are the images contained in Image 2 in this thread.

Photo one.
Shutter speed 1/800
f5.6
ISO 200
EV Bias -0.7
HL Correction OFF
Shadow Correction OFF

Photo two.
Shutter speed 1/800
f5.6
ISO 200
EV Bias -0.7
HL Correction ON
Shadow Correction Max

As I am looking at them in Develop module and have them side by side in Reference mode, there is a definite difference between the two RAW files despite having the exact same exposure variables.
When I max out the shadow and highlight bars (highlights -100, shadows + 100) the difference is still apparent.
In fact... to match the images closely to being the same the slider values are;

Photo One (no HL Correction or Shadow Correction On)
Highlights -100
Shadows +100

Photo Two (HL & Shadow Correction On)
Highlights -78
Shadows +87

These values now make the photos have a similar dynamic range look, shadows being matched and car at same brightness/details (though I'm eyeballing it here).
So if all this stuff is nonsense for RAW then how come two RAW files with the exact same Exposure triangle variables are giving the LR user different range of scope for playing with the dynamic range sliders.

(PS happy to provide the RAW files for download and inspection for anyone who wants them)."

I am seeing more dynamic range to play with in LR with the file that had both HL Correction & Shadow Correction on for sure, the question for me was that purely to do with HL Correction only. I wouldn't have thought so because I seem to have more room to move with the Shadows.





Sorry, yes I mean I get that, I didn't explain myself properly.

When the camera is looking at a scene with very strong dynamic light, to use 0.0 EV Bias I think is not the best as Highlights will definitely be blown. Lowering to something like -0.7 EV Bias will protect the HL more and although the exposure will be lower and shadows darker they are still not blown neither (and very recoverable in post). On a cloudy overcast day whereby lighting is not so dynamic but more uniformed then shooting at -0.7 EV Bias may be overkill and unnecessary and an EV Bias set back to 0.0 may be fine.
I'm concerned with a setting for difficult high dynamics that will work the best, treading that fine line of getting as much shadow detail as possible without blowing Highlights.


If Shadow compensation doesn't affect RAWs then why in my second post whereby I compare a shot with the same settings but one that has HL Correction on and Shadow on seems to have more dynamic room to move in post (please see above). Not only do I get more Highlight breathing space but also Shadows.
Larrymc is not the first member to perceive Shadow Correction to be baked into RAW.

Hey Eddy, after closely examining my DNG files from last night, I've changed my mind regarding the shadow correction. I had a hard time finding much difference in my two RAW files upon close inspection today. What I thought I saw can probably be chalked up to a slight difference in the light from shot to shot, the RAW histograms in Fast RAW viewer are essentially the same from no shadow correction to full shadow correction so my conclusion is that the shadow correction does not affect the RAW files. I may do some more testing at a later date but I doubt that the outcome will be different.

Larry

[BruceBanner]

Exposure compensation merely changes the actual stop-wise exposure with no modification of the sensor response curves, otherwise, the results would not be consistent. You can demonstrate by using EC in M mode vs. simply dialing in the compensation manually where the camera is ignorant of intent.



Not my experience, though there may be a difference in what is considered "breathing space". I don't have time today (heading to the mountains to escape the heat and get alpine flower photos), but can do some work in RawDigger later to see if there are any histogram changes in the sensor data as written with those two settings.


Steve

Yup I get that, I mean like if you leave the Exposure Compensation alone, and the Pentax camera is faced with a highly dynamic scene, then it's a little too hot to get a balanced shot. Do other brands have it different? Some are saying they do, so set a Nikon set against my same scene with same ev comp (or none) would it exposure the shot differently (all things being considered equal) to a Pentax? How does a camera decide what 0.0 ev comp 'looks like'?

My point is I typically always run (at least last year) with -0.7 EV comp in my Av mode as sunny australia often has very bright parts in the image and very dark. I get a little more highlight protection this way and can lift shadows. I have found however on a overcast day that 0.0 is better for capturing more and you need not worry so much about blown highlights etc.

Enjoy the mountains! We can pick this up later.

Right now I'm thinking of changing my strategy to shoot -0.7 (or -1.0) + HL on and Shadow On, though the shadow may be irrelevant (and shoot ISO 200). I'm basically looking for an ultra conservative mode of quick shooting and hopefuly never really blowing an important highlight in an image.

EDIT: Just to clarify, I'm not meaning to say that using these settings is somehow better than not using them, just that incorporating them might help a user quickly take a shot of a highly dynamic lighted scene. It's just the 'Australian' way to get constant dynamic range from those blue sky days. A bright part in the scene where the sunrays fully illuminate an area and then another part in the scene completely void of any sunrays so extremely dark shadows. Yes I could continue to chimp and reduce EV Compensation for the scene, but I was kinda curious if I set up a 'Dynamic Light' user mode and perhaps use a setting of -0.7 (or -1.0) EV with the HL and Shadow Correction on then perhaps this would almost guarantee no blown highlights in the shot (but riding that exposure wave to being as close to a blown highlight as possible).
I understand there is no 'correct exposure' for any shot really, for this purpose it's just about quickly squeezing off a shot in difficult conditions with a good degree of prearranged settings that hopefully should provide the user with maximum dynamic range to work with in post with no highlights blown.

What are the impacts of using HL Correction (And/or) Shadow Correction, does it slow the shooting down, buffer worse?