Nikon D200 Highlights representation.

I just shoot two test images of a very contrasty scene.

 

I understand the principle of overexposing the most I can without blowing the

highlights in order to have more tonal information in the RAW image.

 

So my pictures came out like this:

 

#1: Dark but didn't blow the highlights (according to D200 and ACR histograms).

 

#2: Well exposed but with blown highlights (also according to both histograms).

 

But guess what, changing the exposure in AdobeCameraRaw allowed me to get the

highlights back from the image #2. The highlights were blown but it was

recoverable.

 

The blown-highlight-version became the best one, with more shadow information.

 

So I guess the blinking highlights and histograms on the D200 are not that

real. I had concluded that the blinking highlights meant lost information

there. But it didn't.

 

Is this normal?

 

Cheers

Your contrast in camera might be set high. That would explain the blown out highlights in

the jpeg, but not in raw.

Tristan,

 

The constrast is default (normal) and the format is RAW.

 

What I questioned is the way that the camera histogram induces the user to believe that the blown-out highlights are forever lost when they are still in the file.

Just to make it clear, both the #1 and #2 pictures were taken in RAW mode.

Marcio, are you sure your monitor is calibrated?

Hi, Marcio,

 

I would say that Nikon's blown-highlight-warning is rather "conservative" and not "not that real".

 

I have once used one CF card for both D100 and Coolpix 4500. I happened to monitor some landscape shots taken by D100 which were perfectly exposed (no highlight blinking on D100's monitor) on 4500 and found that 4500's monitor warned that some parts of clouds of the same images were overexposed (with highlight blinking).

 

Obviously, the dynamic range of D100 is much wider than that of 4500, so 4500 judged the exposure of the images "conservatively".

If you read your manual, the blinking highlights indicate areas that *may* have been blown- not necessarily truly blown. If you zoom into the image and can see your detail just fine, go with that exposure. If you underexpose you are placing your highlights where the shadows should be and you lose shadow detail, and increase your noise when you have to boost the image in your RAW processor. I'm glad you are learning this early on! This is sooo critical to understanding anything digital! (including audio, video, photo, you name it!)

 

Aaron Lee

Thank you all for the help so far.

 

Yaron, I guess in this case it has little to do with monitor calibration. I am evaluating these results according to the highlights alerts and histograms on the AdobeCameraRaw and D200.

 

At the end of it I think it's a matter of the D200 being really conservative in this matter, what can be bad.

 

For instance, in this example I pointed, I would discard the #2 picture because the camera said that it probably lost detail in highlights. But, after using ACR exposure, the #2 became the best of them with more shadow detail AND highlight detail.

Look at the R, G and B histograms separately. You will see that you got better dynamic range with the overexposed RAW file because higlights were not blown in all three channels. The information for the blown highlights was recovered from the remaining one or two.

 

Regards, Marko

Marcio,

 

an earlier poster referred to the contrast setting, and I think this might be the key here. AFAIK the histograms (on the D200's monitor) are based on the JPG preview that is embedded within the RAW file. While your contrast setting has no effect on the RAW data itself, the preview would be different.

 

Actually it's easy to try; shoot the same scene twice with different contrast settings and check if the histograms are different. In fact I will just get my D200 out of the bag and try it...

 

Carsten

 

http://www.cabophoto.com/

just tried it: the histograms change with different contrast settings even in RAW mode

 

Carsten

Marcio said that he shot using the default contrast setting, which is pretty flat and is the one that I use.

 

The key to making good use of flashing highlights is not necessarily having no flashing highlights, but making sure that critical parts of the image are not flashing. For instance if I'm shooting a bride I don't want flashing highlights on her face, minimal in the dress and I don't care about clouds in the sky.

 

The D200 also has the separate, 3 individual RGB histograms display. You still have to think about what you're looking at here. If you are shooting AWB under tungsten lighting, the red channel will show clipping with a good exposure.

Marko and Bruce: I think maybe the key is here. Maybe the flashing RGB alert was indicating clipping somewhere but the G and B channel were not clipped, for instance.

 

I will try to re-do the test.

 

Carsten: So in order to have a better RAW flashing highlights in camera I would have to lower it's contrast? Is that it?

As Marco suggested looking at the individual channels is a good idea. The problem is that "recovery" of the lost information in one or two channels with blown highlights is not a recovery it is a guess! If the information is not there you can not "revover" it. You can make an "educated guess" what the color in these pixels originally was. It may be ok for most cases and give you the result you like. For documentation it is certainly questionable.