[house]

That's a great approach, and one I'd never thought of. Whenever I use the auto-matched tone curve feature in RT, the results always looks way too punchy - too much contrast and/or saturation. Shooting with the in-camera profile set to flat would tame that significantly, I imagine. Thanks for the tip - I'm going to have to try this!

The realisation opened up a whole new way of thinking about jpg styles for me. By setting the camera to produce the output you like you save considerable time and effort in postprocessing.

[Wingincamera]

I shoot raw but have jpeg set to a flat profile so I can better see blown highlights in the lcd viewing. If using the Natural profile the lcd view might indicate a blown highlight when it really isn't. The negative part of this is when someone wants to view the lcd to see the picture and think that is how the final picture will be, not understanding the reason for the flatness of the photo.

[Serkevan]

The realisation opened up a whole new way of thinking about jpg styles for me. By setting the camera to produce the output you like you save considerable time and effort in postprocessing.

Yep, I set RT up with Squirrelmafia's dcp profiles and it makes it so much easier to get a nice looking image... really streamlines the process.

[stevebrot]

The idea is to adjust the contrast level on the selected style , so that the camera indicate blown pixels at the same level of exposure that makes the raw data clipped.

OK...got it. You are referring to the RAW histogram that you will never see without special tools. If the "flat" custom image setting works similar to what @photoptimist suggests, it would be the obvious solution.


Steve

[Ian Stuart Forsyth]

An histogram and "blinkies" can be enabled and displayed on top of LCD preview (live view), and those rely on selected jpeg style. When I used "bright" or even "natural" style and pushed exposure compensation, the highlight warning came earlier than when using the "flat" style. That means there was still some exposure headroom even with seeing blinking spots on the rear display with Bright and Natural styles.

With what Wb are you relying to gauge the headroom, for daylight WB the red channel when viewed on the back of your camera is being multiplied by a factor of 2 time. the only real blinking spots or histogram you can rely on is the green channel as there is no multiplying going on with that channel.

That means there was still some exposure headroom even with seeing blinking spots on the rear display with Bright and Natural styles.

I

You also have to consider the baseline exposure that is being applied to the raw convert for many cameras this can be 3/4 to 1 1/2 stops

It can be done but first you have to setup the cameras profile that would best give you a more neutral low contrast image.
You would also need to find your cameras UniWB setting and load this into the camera WB. On quick way you can find out the UniWB is to look at the multipliers in the exif data


looking at the DNG you can see that line c627 the Red 0.4096 is being multiplied , Green 1 is not and the blue 0.587155 is also being multiplied. What you want to do is select a Wb that will give the Red 1, the green 1 and the blue 1 values
This is the UniWB settting and is different for every model of camera.

https://dslrbodies.com/cameras/nikon-and-dslr-camera-faq/what-is-uniwb.html

GUILLERMO LUIJK >> TUTORIALS >> UNIWB. MAKE CAMERA DISPLAY RELIABLE

What you are asking for can be done


Here is how the image would appear on the back of the camera, on the left is the profile & UniWB used to convert the raw file into a jpg view and on the right is the image when viewed using fast raw viewer

---------- Post added 10-06-2021 at 12:26 AM ----------

The idea is to adjust the contrast level on the selected style , so that the camera indicate blown pixels at the same level of exposure that makes the raw data clipped.

First you have to know what your looking at is truly a raw histogram

Here is an image taken using ETTR

Even within a raw converter you don't have a raw histogram, not until you remove some of the behind the scene processing

Here is a more accurate view of the raw file


Next you also have to consider the color space

[house]

Raw ettring isn't such a great idea imho. At least not if you go right to the edge. The reason is that the modifications you need to do to create a nice looking image may need a bit of a margin. I've shot UNIBW and attempted raw ettring a few years ago. Gave up on it because it wasn't worth the trouble. The resulting files were quite difficult to work with. For my shooting it makes more sense, as mentioned above, to set a style close to where I want to be when opening the file.

By spot metering highlights and adding exposure compensation to the level you feel comfortable is a pretty good low tech way of achieving the same thing.

[biz-engineer]

With what Wb are you relying to gauge the headroom, for daylight WB the red channel when viewed on the back of your camera is being multiplied by a factor of 2 time. the only real blinking spots or histogram you can rely on is the green channel as there is no multiplying going on with that channel.

Yes, I've search the web and found discussions around the so called UniWB, it seems something used by Nikonians. How can uniWB work on a Pentax camera. Now, I have an issue because the Pentax light meter looks at R, G and B channels to optimize exposure, so, depending on the actual incoming color mix, the exposure amount changes. How can I use UniWB on a Pentax camera? Does this mean I can look on the camera at G channel histogram and ignore R and B channel? Problem is , on the K1, the histrogram of LV preview doesn't show separate R,G and B; separate R,G and B histograms are only shown with picture review after taking it.

---------- Post added 06-10-21 at 10:23 ----------

By spot metering highlights and adding exposure compensation to the level you feel comfortable is a pretty good low tech way of achieving the same thing.

I've used that, mostly spot meter on brightest area of image with +2ev exp. comp., but, it's strange because in sometimes I can dial +3ev exp. comp. on spot meter without clipping the brightest point and sometimes even +2ev exp. comp. is too much. So, there is up to 1 full stop of wiggle to achieve ETTR and sometimes it's no better than a regular matrix metering exposure (too conservative) and sometimes the highlights are blownup.

---

Last edited by biz-engineer; 10-06-2021 at 01:30 AM.

[house]

Yes, I've search the web and found discussions around the so called UniWB, it seems something used by Nikonians. How can uniWB work on a Pentax camera. Now, I have an issue because the Pentax light meter looks at R, G and B channels to optimize exposure, so, depending on the actual incoming color mix, the exposure amount changes. How can I use UniWB on a Pentax camera? Does this mean I can look on the camera at G channel histogram and ignore R and B channel? Problem is , on the K1, the histrogram of LV preview doesn't show separate R,G and B; separate R,G and B histograms are only shown with picture review after taking it.

---------- Post added 06-10-21 at 10:23 ----------


I've used that, mostly spot meter on brightest area of image with +2ev exp. comp., but, it's strange because in sometimes I can dial +3ev exp. comp. on spot meter without clipping the brightest point and sometimes even +2ev exp. comp. is too much. So, there is up to 1 full stop of wiggle to achieve ETTR and sometimes it's no better than a regular matrix metering exposure (too conservative) and sometimes the highlights are blownup.

I've settled on +1,5ev but as you say there is more headroom sometimes. I've not figured out why either but you do get a bit of "feel" for it.

As I said I've found that headroom is good to manage contrast in the highlights. If your're up against the edge I've found it limiting in practice (in theory it shouldn't matter).

[BigMackCam]

I've used that, mostly spot meter on brightest area of image with +2ev exp. comp., but, it's strange because in sometimes I can dial +3ev exp. comp. on spot meter without clipping the brightest point and sometimes even +2ev exp. comp. is too much. So, there is up to 1 full stop of wiggle to achieve ETTR and sometimes it's no better than a regular matrix metering exposure (too conservative) and sometimes the highlights are blownup.

Even spot metering covers a considerable area, even more so if you're metering on a distant point source... so you're always going to get some variation in metering on highlights, especially specular highlights. It's a poorly-named term, really... "localised" is perhaps a more accurate description than "spot".

When all's said and done, until Pentax cameras feature a histogram based on the raw processing pipeline rather than JPEG, there'll be practical limits and inaccuracies in your otherwise logical approach to use the histogram with a flat in-camera JPEG profile. It's still a pretty good approach, though...

[biz-engineer]

Even spot metering covers a considerable area, even more so if you're metering on a distant point source..

Yes, correct. So I used spot meter a color uniform target. On a green target, +1.3 ev of exp. comp. is the maximum before the green channel is clipped. On a white target, I can push exp. comp. to +2.7 ev without any channel clipped.
When using various WB are applied to the same Raw file, it looks like the green channel histogram is fixed with the red and blue histograms move around. Once one of the channel is truly clipped, color can't be recovered.

[BigMackCam]

I just got round to trying different JPEG profiles to see the effect it has on the histogram. My K-3 doesn't have a "Flat" profile, so I used "Natural" instead, and dialled the contrast down to its minimum setting, with all other parameters set to their defaults. Shooting in Live View with AE-L engaged and switching between "Bright" and (my edited, low-contrast) "Natural", I see no obvious difference in the live histogram. Am I doing something wrong?

[Lowell Goudge]

I just got round to trying different JPEG profiles to see the effect it has on the histogram. My K-3 doesn't have a "Flat" profile, so I used "Natural" instead, and dialled the contrast down to its minimum setting, with all other parameters set to their defaults. Shooting in Live View with AE-L engaged and switching between "Bright" and (my edited, low-contrast) "Natural", I see no obvious difference in the live histogram. Am I doing something wrong?

Yes, the way to evaluate the jpeg profiles, especially contrast, is to use a uniformly lit target, such as a paved road, or block wall, and in manual mode, take one shot, so that you are metering correctly with a shutter speed around 1/125

Without changing iso or aperture change the shutter speed in 1/3 stop increments, down to over expose and up to under expose.

If you take the full range of outputs, over a range of +/-6 stops, and plot the grey scale value vs the exposure stops based on shutter speed, you will get a curve of grey scale value vs exposure.

Repeat this test over the full range of contrast values, and you will see, if things have not changed since my *istD, the following.

Over the range of grey scale values from 30 to 225, you will have a very uniform relationship of 45 greyscale per stop, this slope changes +/-5 greyscale as a function of contrast

At the ends, from 30 down to 0 and from 225-255 the first stop below 30 or above 225 will be 15-20 greyscale, the next stop will be 7-10 and the next will be 3-5

Overall there will be 10-11 stops but the ends are highly compressed compared to the middle.

This approach can’t be used with the highlight and shadow protection because for those functions you need a graduated target and you need to measure the exposure over the graduated target to see the differences,

[biz-engineer]

I just got round to trying different JPEG profiles to see the effect it has on the histogram. My K-3 doesn't have a "Flat" profile, so I used "Natural" instead, and dialled the contrast down to its minimum setting, with all other parameters set to their defaults. Shooting in Live View with AE-L engaged and switching between "Bright" and (my edited, low-contrast) "Natural", I see no obvious difference in the live histogram. Am I doing something wrong?

Styles do make a difference in the histograms but the change may be small especially near left and right end of the scale. I was wrong to assume that a contrast curve has anything to do with exposure, as long as both end-points (0, 255) of the curve aren't changed, any curve will do. The risk of pushing the histogram to the right is that the picture may appear nor to be clipped but some of the colors are lost due to one of the primary channels being clipped. I find blinking pixels easier to read that the histogram anyway.

[Ian Stuart Forsyth]

How can uniWB work on a Pentax camera.

There is no problem using uniWB with pentax cameras, the challenge is finding the WB setting that give you uniWB.

I have an issue because the Pentax light meter looks at R, G and B channels to optimize exposure, so, depending on the actual incoming color mix, the exposure amount changes.

uniWB is not for calibrating your camera metering system, where it is useful is setting up your cameras play back feedback as to how the image is being stored within the raw file

How can I use UniWB on a Pentax camera? Does this mean I can look on the camera at G channel histogram and ignore R and B channel?

It depends on the luminate that is your light source, for the sun the red channel falls well below the other 2 channels

Here I have exposed very far to the right and as you can see the red channel is not even clipping

Here I am using an exposure that again is further to the right and in the reds of the car are not close to clipping but the tent in the back ground is clipped, this brings me to a point that I have said regularly, if you are shooting with the sun as your illuminate and you are setting your exposure base on a white card or you have a white object that is found in the frame and that white target is not showing any clipping in the greens and blues your reds are pretty safe . So for sunlight you can ignore the red channel unless you are photographing a sunset then you will want to keep track of your red channel.

Not having a RGB histogram in LV is not only a problem with pentax cameras.

Using the spot meter is one of the techniques that I use for anytime chimping is not workable. What I do is set the camera ( for Nikon +1 13) pentax +0.7 to + 1, this is not so that the camera adds that + to the image, but rather allow me to see beyond what my meter can display when I place my spot meter on where I want the white point to be. For Nikon I would need the metering display to show all the way up to +4 1/3 in the display. Basically I am applying an offset to the metering system so that any white target that falls at 2 2/3 scale closely resembles where that white falls within the raw file


Here I metered in red and with the correct offset I know how that white target falls within the raw file, without the offset and I place the spot meter over that target it would only show +3 without knowing how far beyond that 3 I am going. With the offset I have decided and metering on that target falls bellow +3 I know its not going to clip and click easy ETTR.(also take notice as to how the red channels was captured) Check the playback image with uniwb setting and the profile and see if you have any hotspots and your done.

---------- Post added 10-06-2021 at 08:10 PM ----------

When using various WB are applied to the same Raw file, it looks like the green channel histogram is fixed with the red and blue histograms move around.

If you really want to see how the channels relate to how they are captured

go into the exif and change line C628 to 1,1,1 and in your raw converter select as shot for your white balance and now you are viewing it in uniWB

---------- Post added 10-06-2021 at 08:32 PM ----------

Also if you look at the Baseline Exposure that is what is telling your raw how to handle the raw file

If your are using most raw converters and you what to see a more accurate raw file for processing take your BLE adjust your exposure value slider by -1 and then - the BLE = how far you move the slider
For the k5 as shown above -1 +0.5 = 0.5 is how far you move the slider ( the K5 showed -0.5 so that becomes +0.5 two negative make a right )
Depending on the converter the brightness, contrast and black will very



Some good reading if you have sometime
https://www.rawdigger.com/howtouse/deriving-hidden-ble-compensation

https://photographylife.com/where-are-my-mid-tones-baseline-exposure-compensation

The Optimum Digital Exposure - Luminous Landscape

If you are shooting red objects and you want the best out of them here is a good read
https://www.libraw.org/articles/magenta-filters-on-digicam.html

For the last couple of years I have been using Magenta filters for photographing red flowers and it really brings out the fine details that can be lost to noise in the reds



Here is what the raw file looks like in sunlight as you can see the red is now shifter to the right and more equal to the other 2

[biz-engineer]

uniWB is not for calibrating your camera metering system, where it is useful is setting up your cameras play back feedback as to how the image is being stored within the raw file

Thanks for showing further details. I'm still not sure about how helpful is uniWB practically. Is it correct if I say that uniWB is there to make sure the right side of the live view histogram represent the max level the sensor can record for any of the primary colors?