Adobe ACR showing raw IR images as being over exposed

[george013]

Pixel values are easy enough to see in ACR. The sampler tool provides a readout of wherever anyone wants to sample the values at any particular location. I'm not particularly sure why this is relevant? That being said for an 8-bit image, each channel has a maximum value of 256. For 16-bit data, each channel can have a maximum value of 65,536.



A dng is a strange construct and there are two types. The first tupe is simply a wrapper for raw data, the metadata and a jpeg. It does not contain any RGB image other than the JPEG and the raw convertor has to turn it into RGB data.

The second type is what Donald is dealing with where processed data can be stored in the file; i.e. there is RGB data. The problem here is I have no idea how ACR handles this data and suspect that something is going on in the background that is causing the issue that Donald sees.

Regardless, this has nothing to do with what Donald is trying to figure out. His question is still trying to understand what has changed that ACR is reporting clipped highlights, where in the past it has not.

The why is easy to answer, pixel values are exceeding a limit. If you want figure out why 2 different programs are showing two different clippings the first question should be: when do they show clipping. Only then you can go further.

George

[xpatUSA]

Haven't used ACR since Elements 6, so: can ACR now show a raw histogram? In other words, does the ACR histogram actually show sensor exposure or is it just showing the brightness levels of a review (converted) image?

[Manfred M]

Haven't used ACR since Elements 6, so: can ACR now show a raw histogram? In other words, does the ACR histogram actually show sensor exposure or is it just showing the brightness levels of a review (converted) image?

I would suggest it uses a histogram based on the imported raw data.


This is a SOOC JPEG opened in Camera Raw.





This is the raw file opened in ACR




The two histograms are different, so that leads me to believe that ACR generates a histogram based on the converted raw data, not the JPEG.

[george013]

I would suggest it uses a histogram based on the imported raw data.


This is a SOOC JPEG opened in Camera Raw.





This is the raw file opened in ACR




The two histograms are different, so that leads me to believe that ACR generates a histogram based on the converted raw data, not the JPEG.

A histogram based on the raw, sensor, data will contain only one color per pixel. Besides Ted's gear.
I just can't resist anymore.


Main part is the rgb raster block. Your jpg was created by the in-camera converter, acr create his own raster image. Two different pictures by definition.
Still don't know about the pixel values of Donald's blinkies.

George

[Manfred M]

A histogram based on the raw, sensor, data will contain only one color per pixel. Besides Ted's gear.
I just can't resist anymore.


Main part is the rgb raster block. Your jpg was created by the in-camera converter, acr create his own raster image. Two different pictures by definition.
Still don't know about the pixel values of Donald's blinkies.

George

George - please read Ted's question. It simply asked if ACR used the JPEG to create the histogram or if it was created from the raw data. Clearly it uses the raw data because while the histograms have the same basic shape, there is space on both the left hand side and the right hand side of the one generated from the raw data. With the JPEG, we clearly see highlight clipping and less space on the left hand side.

I do not see the relevance of your diagram in this discussion. As I (and others) have mentioned before, it definitely incorrect.

[DanK]

I think what George meant was that the histogram shown for a raw file is for the raw file as rendered. This is correct: you can see this by importing a raw image into Lightroom or ACR and changing the profile used to render the photo. The histogram will change, even though the raw file obviously doesn't. This is also entirely irrelevant.

I think this thread has been hijacked. The OP poses a specific question: why is it that when bringing photos from DxO into ACR, ACR now shows an obviously incorrect histogram that does not display of you close ACR and thereby show the same image in photoshop. In addition, this phenomenon is new: Donald didn't have this experience in the past, even though he has been moving files from DxO into ACR for some time.

Most of this thread has no relevance to that question.

[Manfred M]

Donald - while I don't have any IR shots, I have a number of images shot against a white background that is close to being pure white. I have pushed some so that they have started to clip in PhotoLab. I have been processing them in PhotoLab 2 and exporting them to ACR as dng files directly from PhotoLab. My Photoshop and Photolab are up to date and the most recent versions.

I have not been able to replicate your issue.

I start in PhotoLab and have changed the exposure to clip. The area that is largely yellow corresponds to where the large softbox was pointing to on the white seamless paper.




I then exported the image to Photoshop directly as a dng file. This is what ACR opened up with; even the areas that were clipped in PhotoLab came across without clipping.




I opened this file in Photoshop and dropped colour samplers in areas that should have been hottest. The maximum reading I got was 254, 254, 254.

[george013]

George - please read Ted's question. It simply asked if ACR used the JPEG to create the histogram or if it was created from the raw data. Clearly it uses the raw data because while the histograms have the same basic shape, there is space on both the left hand side and the right hand side of the one generated from the raw data. With the JPEG, we clearly see highlight clipping and less space on the left hand side.

I do not see the relevance of your diagram in this discussion. As I (and others) have mentioned before, it definitely incorrect.

Look at my diagram. That raster image in the centre is the image where all the editing etc. is done on. If you look at a picture, it's that image. If I use Ifanview to view images, even raw files, then the embedded jpg will be loaded. But if you use a converter the raw data will be used to create a rgb raster image. The shown histogram is based on that raster image in memory, and has been fresh builded, with all the specific settings for the used program.

George

[Donald]

I have not been able to replicate your issue.

It looks as if it's an issue pertaining to IR files.

[Manfred M]

It looks as if it's an issue pertaining to IR files.

Donald - I just ran into this in ACR.



There are no "blinkies" but the indicator at the top right corner of the histogram shows there are clipped highlights as the triangle is white. When I explore the image in areas that are very bright the maximum value shown is 254,254, 254. When I <Alt> click on the Highlights or Whites slider, this would normally give a view of the clipped areas but nothing shows up.

There is definitely something wrong with this functionality in ACR.

[Stagecoach]

Donald - I just ran into this in ACR.



There are no "blinkies" but the indicator at the top right corner of the histogram shows there are clipped highlights as the triangle is white. When I explore the image in areas that are very bright the maximum value shown is 254,254, 254. When I <Alt> click on the Highlights or Whites slider, this would normally give a view of the clipped areas but nothing shows up.

There is definitely something wrong with this functionality in ACR.

I suspect that example may be nothing new Manfred, I have seen it happen before.

Donald mentioned "Adjusting the Exposure slider suggests that I have over-exposed by anything between 0.60 and 0.95 stops.", that is a pretty significant amount.

[Manfred M]

I suspect that example may be nothing new Manfred, I have seen it happen before.

Donald mentioned "Adjusting the Exposure slider suggests that I have over-exposed by anything between 0.60 and 0.95 stops.", that is a pretty significant amount.

Thanks - this is the first time I have noticed this, but on the other hand, it's also the first time I was looking for something.

[Stagecoach]

Since the last update of CC, I've got this problem. I use DxO Photolab to process my raw files. But, when I convert them to .dng or otherwise take the raw file into ACR, the 'blinkies' show up, indicating over-exposure. Adjusting the Exposure slider suggests that I have over-exposed by anything between 0.60 and 0.95 stops.

Donald,

I opened a NEF into DXO Photolab2 and made a very minor adjustment to remove the 'highlight clipping', then exported it to file as a DNG.

This was then opened in ACR and there was clipping evident that required around -1.0 'exposure' to remove the red non blinkies completely.

ACR was set in 'sRGB Colour Space' and 'Adobe Color profile'.

Findings;

a) Changing between Adobe Color and Adobe Standard profiles clipping blinkies could be removed by -1.0 & -0.5 exposure respectively.

b) Changing to 'Adobe RGB Color Space' there was no highlight clipping at all in either Adobe Color or Standard Profiles.

In case you have not checked them have your colour spaces changed somehow?

[george013]

Donald - I just ran into this in ACR.



There are no "blinkies" but the indicator at the top right corner of the histogram shows there are clipped highlights as the triangle is white. When I explore the image in areas that are very bright the maximum value shown is 254,254, 254. When I <Alt> click on the Highlights or Whites slider, this would normally give a view of the clipped areas but nothing shows up.

There is definitely something wrong with this functionality in ACR.

A good start to examine the values of the blinkies Now waiting for Donald's values.

May I also ask your attention to the following.

Since the last update of CC, I've got this problem. I use DxO Photolab to process my raw files. But, when I convert them to .dng or otherwise take the raw file into ACR, the 'blinkies' show up, indicating over-exposure. Adjusting the Exposure slider suggests that I have over-exposed by anything between 0.60 and 0.95 stops.

Donald does do the converting in DxO. His histogram is based on that raster image, you know somewhere in the middle of my diagram. Then he's exporting that imag as a dng. And is opening that dng in acr. I don't know anything of dng files. By example if they have to be interpreted and converted to a rgb raster image. If so there is always an influence of that used program during that process..

Why not opening that dng file in dxo again? And examine the differences then?

For those who state that Donald didn't experience that behaviour before, he states he got this problem after the last update.

George

ps.
I see Graham could have good point too.

[Donald]

Donald - I just ran into this in ACR.

>Pic>

There are no "blinkies" but the indicator at the top right corner of the histogram shows there are clipped highlights as the triangle is white. When I explore the image in areas that are very bright the maximum value shown is 254,254, 254. When I <Alt> click on the Highlights or Whites slider, this would normally give a view of the clipped areas but nothing shows up.

There is definitely something wrong with this functionality in ACR.

Delighted that you found this. As I have said, it seems to be more pronounced with an IR file, but you finding the same thing shows that there is something wrong with the functionality and that I'm not going mad! I am grateful.

[Donald]

In case you have not checked them have your colour spaces changed somehow?

I don't think so, but let me go off and check that.

[davidedric]

Been reading without entirely following. However, I feel sure that in Lightroom I have also seen a histogram with apparently clipped highlights, but without blinkies. I have pushed the whites (I think!) Just so I could see where they would first appear, and then gone back again to verify that they weren't there.

[Manfred M]

Delighted that you found this. As I have said, it seems to be more pronounced with an IR file, but you finding the same thing shows that there is something wrong with the functionality and that I'm not going mad! I am grateful.

I suspect that the IR file per se has nothing to do with it as the software has no way of knowing that this technique was in use. Rather, I suspect you are seeing it because monochrome IR images have a lot more very light grays in them, so would be more susceptible to this showing up. This is why I went to some of my images that had lots of values close to being white in them. The main difference between my image and yours is that the "near white" values in your shots will be quite random and fairly small as edges of leaves will break up the area whereas in mine, the values are fairly evenly distributed and the only edges that occur are near the subject.

I was starting to explore along the same thought process as Grahame and was looking for settings that could push individual pixels from being "near white" to white like the SmartLighting or the Vignetting (correction), but found that these implementations in PhotoLab seem to protect highlights.

The other area I was starting to explore, much like Grahame, were the settings that could not be disabled in one or the other piece of software. The PhotoLab Color Rendering and ACR Profile settings fall into this category and some profile has to be selected, so by going with the linear dng output from PhotoLab it is possible that two different adjustments will be made to the file; one by PhotoLab and one ACR and I see no way around that.

[george013]

Been reading without entirely following. However, I feel sure that in Lightroom I have also seen a histogram with apparently clipped highlights, but without blinkies. I have pushed the whites (I think!) Just so I could see where they would first appear, and then gone back again to verify that they weren't there.

It looks like highlight warning is in 2 steps. You could find the max value for just that sign by lowering the exposure and find out at what value it disappears. As I understood highlight warning is on when just one of the channels is to high.

George