How far to the right are you?

[george013]

George if you do a simple test such as the one Allan mentioned in post 26, you will get a good idea of the margin between white clipping point in the camera jpeg and the raw file.

I found a similar result to Allan with my D610 - about 1 stop. I find this a bit surprising and am not sure why the difference would be so great.

Dave

The way Allen is doing it, is using two different RAW-converters, the in-camera and a out-camera, with different settings. The more I think about it, the less sence it makes. A JPG is a compressed rasterimage. That image is the same image that comes out of the RAW-converter.
When it's written to file, it's compressed. Groups of pixels are averaged. That means the maximum value of the highest pixel wil not be exeeded. And that's what I see when I compare the NEF and JPG in Capture. Less clipping in the JPG.
Capture is taking in account the camerasettings which another converter doesn't.

BTW, did you try it with Capture?

George

[dje]

The way Allen is doing it, is using two different RAW-converters, the in-camera and a out-camera, with different settings. The more I think about it, the less sence it makes. A JPG is a compressed rasterimage. That image is the same image that comes out of the RAW-converter.

When it's written to file, it's compressed. Groups of pixels are averaged. That means the maximum value of the highest pixel wil not be exeeded. And that's what I see when I compare the NEF and JPG in Capture. Less clipping in the JPG.
Capture is taking in account the camerasettings which another converter doesn't.

BTW, did you try it with Capture?

George

No I don't have Capture, I'm using ACR.

Yes there are two different raw processors at work - the in-camera one uses a certain white point when creating the jpeg and the external raw processor uses a different white point. It's difficult to know where the external processor gets the white point from but I think it's provided by the camera maker in the raw file.

Yes camera settings will affect the amount of clipping but I Can't see that running to one stop.

Dave

[ajohnw]

No I don't have Capture, I'm using ACR.

Yes there are two different raw processors at work - the in-camera one uses a certain white point when creating the jpeg and the external raw processor uses a different white point. It's difficult to know where the external processor gets the white point from but I think it's provided by the camera maker in the raw file.

Yes camera settings will affect the amount of clipping but I Can't see that running to one stop.

Dave

The difference between the raw converter and the camera is more likely to be down to a different mid grey point.. It wouldn't surprise me if Nikon, Canon, others and Adobe use entirely different values. It's often shifted around in various jpg modes as well.

I don't think Allen's way of telling the difference is ideal either. A white card isn't really suitable other than it's likely to have some highlight areas that are brighter than the rest of it.

A better test would be a colour test card or something similar*. Ok camera blinkies wise but when the raw converter is used the exposure should be pulled back to make sure there isn't any more available - that can be detected with it's histogram fairly closely.

The other point taking a D610 as it was mentioned switch to adl extra high and nearly 1 full stop is gained in highlights but the camera also shift the mid grey point so some of that is lost after a fashion image wise but the total sensor output that gets into the jpg is very nearly 1 stop more. The black end is more or less the same in all modes but this setting might loose about a 1/3 stop meaning the headroom could drop by the same amount.

*Including a nice bright buttercup yellow with shape without much tonal range and looking at the output might be interesting as well as channel behaviour influences how it looks and seems to often fool metering systems. Debayering too at times I feel as well.

John
-

[xpatUSA]

Incidentally, there is one raw converter that will give a raw histogram - RawTherapee.
Dave

Also there's a freebie called UFraw and a paid prog. called RawDigger.

UFraw is interesting in that you can see the raw histogram and converted+edited histogram at the same time.

[Polar01]

John when I did my test I used a sheet of paper from my printer is was Epson Exhibition Fibre so there is a very high standard of quality put into this high paper so you would not have to worry about some areas being brighter than others.
As for your buttercup yellow paper the standard of quality is no where near that of the photographic paper used, Coloured papers are on the lower end of quality and price and usually make of cheaper pulps. I know as I was involved in the paper and printing industry for a number of year, so experience was gained first hand.

Cheers: Allan

[purplehaze]

This has all been very informative and by all means, continue, if you're so inclined. I just wanted to step in and thank you all for responding. I have read enough that I am convinced that it is worth my while to further experiment with this method. What really stood out for me in the referenced article was the bit about 90% brightness in the software representing two stops underexposure and 97% brightness, one. I will never look at my histograms the same way again. I'm sure this is not the first time I read that, but it is only now hitting home for me, as I lament noisy exposure after noisy exposure. I don't like noise, but neither do I like hauling my tripod around all of the time and I have a crop frame camera, and small, not-so-steady hands. I have recently started using auto-ISO but until now have been putting a rather severe limit on it. Now I see that I might be able to push that limit and actually get better results than I have been getting up until now, and that makes me happy, happy, happy (less FF-envy). I understand that it is not the answer for all types of situations and that, more clearly now, thanks to you. Now, all I want for Christmas is a camera that displays my RAW histograms as interpreted by View NX, Aperture 3 and Lightroom. In the absence of that, I will try to mimic Allan's procedure for finding his OWP as best I'm able; thank you for pointing the way, Allan.

[Manfred M]

What really stood out for me in the referenced article was the bit about 90% brightness in the software representing two stops underexposure and 97% brightness, one.

This is of course the misleading part. If this is information is correct (I would like to see the workup to his assertion), this is only true sofar as what the camera sensor records, not in how you see the actual image. I seem to remember Colin often mentioning that a print shows a maximum of around 4 stops of dynamic range and a computer screen around 5 to 5-1/2 stops, so whatever the camera records is compressed to something that can be reproduced. The material is there in the RAW data and can be extracted, if needed.

The other misleading part is that this absolute brightness being such a major factor, when the ENTIRE range from darkest shadow details to the lightest highlights are actually what matters in an image. So while the numbers he states sound scary, they are really bogus and meant to confuse (and buy his book). If you are getting consistently noisy images, shot with as low an ISO as you can get away with. The base ISO of your camera has the lowest noise and highest colour range that your camera can produce; the higher you go in ISO, the more noise and lower colour range you get. When you are close to the base ISO, this difference will be meaningless, but as you raise the ISO, this will become an issue; but as with anything else in photography, you need to make a choice on how to capture the image and make the appropriate decision on how you will use the exposure triangle in your shot.

So go ahead, and explore the method. I think you will find that there is a lot more hype than substance in what he has written. Yes, ETTR does work, but only if you have the headroom at the high end of the data, because you do not want to clip the whites. Also remember that when you are looking at the histogram or the "blinkies" on your camera, these are based on jpeg data, not raw data, so they are going to be a bit conservative.

[george013]

...
Also remember that when you are looking at the histogram or the "blinkies" on your camera, these are based on jpeg data, not raw data, so they are going to be a bit conservative.

And that's where the last part of this thread was about. There is no reason why that should be. On contrary, blinking highlights should be and are less in JPG. Unless, and that's not mentioned jet, it's the thumbnail size that's making the difference. Reducing the size in pixels might have some influence.

George

[Stagecoach]

Ok, I'll throw in some real world observations into this thread that demonstrate that great word 'variables'

My camera, Nikon D300, this shot taken in RAW 14bit + JPEG Fine. Camera picture control set to 'standard'.

No 1 - RAW file opened in ACR (-2.2 Exp required to remove all clipping)


No 2 - JPEG file opened in ACR


No 3 - JPEG file 'exposure' reduced to remove clipping (-1.05 totally removed it).


No 4 - JPEG file in ACR 'highlights' reduced to remove clipping (-45 totally removed it)


As for the image, working on the RAW file in ACR there are two options, reduce exp or highlights. If exp is reduced 2.2 stops needs 100% shadows recovery to get adequate detail in the blacks, but fully recoverable good image.

Grahame

[Stagecoach]

And that's what I see when I compare the NEF and JPG in Capture. Less clipping in the JPG.
Capture is taking in account the camera settings which another converter doesn't.
George

I have just looked at the clipped image I've just posted in ViewNX2 and the results are opposite to what you have observed in Capture, the NEF shows slightly less clipping.

Just goes to show the variables between software types.

[ajohnw]

Also there's a freebie called UFraw and a paid prog. called RawDigger.

UFraw is interesting in that you can see the raw histogram and converted+edited histogram at the same time.

Some people look at the raw histogram in Ufraw and say oh goody I have a keeper before they have even done anything to the shot. I'd guess that they are wait for 7 stop conditions shooters and don't bother if it's much more than that. It will also show channel clipping which can be useful at times.

There are 2 styles of blinkies even in raw converters. A single channel clipping and not all channels clipping only some.

As I described a butter cup and then see photo paper mentioned may be best to leave that area at that. There is clearly some misunderstanding that would be hard to clear up.

John
-

[Inkanyezi]

There is clearly some misunderstanding that would be hard to clear up.

I couldn't agree more.
I have a pet peeve there, and it is the "exposure" slider. Of course it does nothing about exposure, but the method of operation can be different in different converters. Whatever it does, it does not change exposure.

But there is another problem here as well, and it goes around algorithms for white balancing. Many converters have multipliers for the channels, which can push a channel over the edge when the coefficient is too high. Instead of pulling down the other two, the red channel can be amplified in a way that resembles over-exposure, even though all data is there.

I have read, but not tried, that this is a problem of ACR, but not of Lightroom. I don't remember exactly, but I think it is in the discussion about UNI-WB. There's also where that "exposure" slider comes in. I really hate that it is called exposure, but not so much to do about it. All manipulations done in PP are digital, and they have nothing whatsoever to do with exposure. However they affect the way the image looks.

So it may be important just which RAW converter you use, whether you wish to use UNI-WB or not, and also whether you subscribe to the ETTR mantra or not. There are many converters out there, and it might be worth the while to do some converter shopping.

[ajohnw]

I have just looked at the clipped image I've just posted in ViewNX2 and the results are opposite to what you have observed in Capture, the NEF shows slightly less clipping.

Just goes to show the variables between software types.

It's down to choice of the mid grey point which I should add can be any colour not just shades of white Grahame, When you load the jpg for instance the camera has chosen one plus some degree of Nikon processing, when it gets displayed in ACR it might choose yet another one.

When you load a RAW in ACR it will be choosing the grey point. When you load it in NX it will do the same as the camera will but probably more thoroughly as there is more processing power available.

This is all down to using different LUMINOSITY curves in the conversion. In capitols as from another thread it seems that Adobe default to another type. The differences in the curves applied can be very subtle so are not that easy to adjust to give matching results.

Nikon actually have another trick but I am not sure how far through it's camera range it can be used. This is yet another curve that can be stored in the camera and applied. It also gets placed in the raw file so that subsequent processing can use it. I can show you what that does but not as well as Nikon can probably due to the camera profile I used but more likely because pre processing is trying to produce a nice image straight off. It would turn out extremely dark if it didn't.

No pre curve




With a style of pre curve that has replaced the white wedding one. Interesting name now. It gives subtle white detail surprisingly maybe looking at the curve but it's being applied to pure raw data before other things are done to it.



Notice that the channel clipping has changed. Next step would be probably a luminosity curve change, exposure change and then on to another editor ideally leaving histogram space for further work - not always possible.

Probably pointless me posting all of this but you never know.

Should add that the D600 profile used is probably Adobe Standard - not the one Nikon software would use.

John
-

[chauncey]

Minutia is what this discussion has morphed into...this software is better than that/my software shows
this/and so on. The exposure slider shouldn't be named "exposure"...so what? This minutia nonsense
matters not a whit.

ETTR is a good tool, but it comes with caveats...ya gotta know your gear and how it tracks light/exposure.
In camera picture-style settings only affects jpeg images...period. Your only goal should be to have
those LV or chimping exposures conform to you RAW converter, starting by neutralizing, sliding all
those picture-style setting to the left.

Then...determining proper exposure is based on experience...every scene is different and you must
determine what works the best in each scene, whether it's a black cat at night or an egret in the sun.

[xpatUSA]

Minutia is what this discussion has morphed into...this software is better than that/my software shows this/and so on. The exposure slider shouldn't be named "exposure"...so what? This minutia nonsense matters not a whit.

I disagree, William, especially with the word "nonsense". And I imagine that Urban is none too pleased with "so what?".

Attention to minutiae, both in theory and in practice, is what distinguishes the good photographer from the point-and-shoot guy.

[Inkanyezi]

Minutia is what this discussion has morphed into...this software is better than that/my software shows
this/and so on. The exposure slider shouldn't be named "exposure"...so what? This minutia nonsense
matters not a whit.

I would say that it does matter. As the erroneous designation of a software control is named to something no software can control, is not a minor thing, not at all minutia, because it clouds any serious discussion about the very subject, that regards exposure. Therefore, it is not minutia. It is an erroneous name that throws a smoke curtain over serious discussion.

And we have not come into the positions of "mine is larger than yours", but there are differences in software, and those differences can matter. Whether it matters for you is another thing altogether. For anyone interested in the matters of optimal exposure and the tools to use, also minutia have to be considered, but I have not seen anything pointing toward brand wars in this thread.

I think most of the info on UNI-WB was collected from Guillermo Luijk, but that the subject was discussed in depth somewhere at Luminous Landscape.

Obviously the subject is a can of worms, but there is indeed some substance that could be worth considering.

The Uni-WB concept calls for another kind of white balancing and raw processing than more common approaches, aiming at better control over exposure, in order to maximise dynamic range and enhance both shadow and highlight detail. ETTR is one of the tools in the toolbox of a resourceful photographer, whether others would subscribe to the idea or not. Many have tried it, and many misunderstand it, but there are still a few, who utilise it after learning the advantages it may have.

Regretfully, the reality behind those "minutia" distorts serious discussion, as people use terminology sloppily, talking about over-exposure and under-exposure and correct exposure, without thinking twice. When words are used in the wrong way, when one participator in a discussion uses words that are ill defined, sloppily, it is very difficult to partake. Therefore, it is not a minor thing, that there is a software control called exposure, as there is no software that can ever control exposure, except the software used in direct connection to the camera before exposure, be it firmware, tethered software or untethered.

[george013]

I couldn't agree more.
I have a pet peeve there, and it is the "exposure" slider. Of course it does nothing about exposure, but the method of operation can be different in different converters. Whatever it does, it does not change exposure.

But there is another problem here as well, and it goes around algorithms for white balancing. Many converters have multipliers for the channels, which can push a channel over the edge when the coefficient is too high. Instead of pulling down the other two, the red channel can be amplified in a way that resembles over-exposure, even though all data is there.

I have read, but not tried, that this is a problem of ACR, but not of Lightroom. I don't remember exactly, but I think it is in the discussion about UNI-WB. There's also where that "exposure" slider comes in. I really hate that it is called exposure, but not so much to do about it. All manipulations done in PP are digital, and they have nothing whatsoever to do with exposure. However they affect the way the image looks.

So it may be important just which RAW converter you use, whether you wish to use UNI-WB or not, and also whether you subscribe to the ETTR mantra or not. There are many converters out there, and it might be worth the while to do some converter shopping.

I totally agree with you. That's why I adviced Capture to do the experiment. Same converter in- and out-camera.

Also the exposure slider which doesn't change exposure but moves the midtones. I only can speak of Capture, but when I have an image without clipping up and down, I can move the slider from minimum to maximum without showing clipping up and down.

George

I saw you reacted before me. Just a remark about uni-wb.
Clipping is considered exceeding the boundaries of the digital representation of the tonal values. Let's say the number 0 and 255. The value of a pixel is the result of A/D conversion and a correctionfactor for the wb. Uni-wb has a correctionfactor of 1 for all channels, so you have the value of the A/D converter sec. But my goal as photographer is an image with "true" colours, thus with wb-correction.

Let's say the wb handles a correction of 1.1. That means that the unchanged vakue may not be bigger than 255/1.1= 232. Corrected for the wanted lightbalance it wil give 255.

The influence of white balance is easy to see. Show the clipping parts on your screen and then change the wb.

George

[ajohnw]

Minutia is what this discussion has morphed into...this software is better than that/my software shows
this/and so on. The exposure slider shouldn't be named "exposure"...so what? This minutia nonsense
matters not a whit.

I just used Ufraw to show what Nikon get up to from raw and to jpg. Also tried to explain why the earlier posts show different signs of clipping. Plus the fact as far as I am aware that various jpg modes will change metering levels. I am not aware of any camera that show a true raw histogram.

[ETTR is a good tool, but it comes with caveats...ya gotta know your gear and how it tracks light/exposure.
In camera picture-style settings only affects jpeg images...period. Your only goal should be to have
those LV or chimping exposures conform to you RAW converter, starting by neutralizing, sliding all
those picture-style setting to the left

Couldn't agree more.but some people look for fix all's and get attracted. All methods have their problems, some stick to a particular one and learn to get round the issues just as with any of the methods.

Then...determining proper exposure is based on experience...every scene is different and you must
determine what works the best in each scene, whether it's a black cat at night or an egret in the sun.

And all will get it wrong from time to time. That's why cameras have bracketing which can be used at times.

No way would I suggest that people should use Ufraw either. Try it if they like just as anyone can try anything they like including rolling about on a motor way for fun playing with the traffic.

John
-

[xpatUSA]

I think most of the info on UNI-WB was collected from Guillermo Luijk, but that the subject was discussed in depth somewhere at Luminous Landscape.

Yes, Urban, in his article he credits Iliah Borg with coining the name "UniWB". (Iliah is a member of LL and DPR and also a developer of RawDigger and FastRawView - see LibRaw.com)

Tried to give you a 'Like' for your post but CiC wouldn't let me. Ho hum.

[ajohnw]

I wonder how many know that raw converters in general process an image in a way that hopefully gives a decent image of sorts as soon as it's loaded. That comment follows from reading the Uni-Wb link - thanks for that Urban. It is possible to stop one package from doing that and then if there is any clipping it's actually on the sensor. The facility is there because it's what people who write PP packages might want also some PP'ers but I have never come across one. Can of worms is correct. If it's not clipped on the sensor that knows nothing what so ever about colour temperature it is still there apart from what software might do to it. If some manipulation can't achieve a correct colour balance without clipping it sounds like there is a bit of a problem to me as apart from actual sensor clipping the info is there but can't be used.

I get the impression that one package renames levels exposure if I read the comments correctly. Actually going on comments elsewhere I suspect the package does other things as well and some do favour it strongly. The person on photozone seems to always use it for instance. Others do to.

I suspect WM's comments relate to technical aspects but as tone curves are fundamental to PP in camera or out I don't see them as minutiae. They need to be understood. The fact that these can also shift mid grey points as well is hardly surprising. In practice recovery sliders are also manipulating tone curves as well. Not the type that changes saturation to any extent though or maybe some in camera does. No way of knowing. On the other hand some people stick to just recovery sliders so have no need to understand them at all. Some probably use both. Some might even tamper with individual channel curves when needed as well.

John
-