Highlight weighted metering

[OleBe]

Hi all,

Will digging deeper into different metering modes and understanding the lack of the live histogram in th X1D is most likely because of the hardware and not because of the software I was wondering why there is no highlight weighted metering mode in the X1D. That would effectively capture max DR and not blow out the highlights. This should be possible to do in the firmware as well? Maybe someone from Hasselblad is reading - think it would be a great addition to this great camera. :-)

What do you think?

[ShawnK]

Maybe I'm not understanding your problem. If I wanted to make sure my highlights are not blown out then I'd be taking a meter reading from the highlight area, recompose, focus & shoot. If not satisfied than I'd adjust my +/- EV compensation, that should get me where I want to be. Maybe shoot manually as I do. With high D.R for X1D this should never be a problem. IMHO

[OleBe]

Actually when time is there I am doing the same. Spot metering brightest part. Lock exposure dial in compensation to ETTR Limit and shoot.

However this is a little bit trickier if there are no blinkies in live view mode compared to other manufactures.

Therefore for example Nikon has such a highlight weighted metering mode which is basically always metering for ETTR. This is handy in shoots where there is no time to recompose, like weddings.

As this is likely easier to implement compared to live view histogram or blinkies, I would love to have such a mode also on the X1D.

[MGrayson]

A specular highlight may be many stops brighter than you want captured. If you want *nothing* blown out, you may have a mostly black frame. So where do you draw the line? You can have the camera guess (and most cameras guess pretty well), or you can adjust manually.

[OleBe]

Yesterday I was experimenting a lot further. The good thing about these crazy times is that you have more time for trials...

Seems also the RGB histogram of the X1D is ,,only" preview jpg based. I tried to expose as ETTR as possible in the camera (brightest part about 2,3 stops over middle grey) to find out in Lightroom, that there is about 1 stop more which I can increase exposure without blowing my highlights.

Understand that it is hard to have a RAW histogram in camera. But now I know I have to expose so that there are no Blinkies in my highlights and can afterwards double exposure time to get one more stop of information.

Did anybody else try this out? Funny thing, for my Leica Q it is exactly the same. Exposure warning is screaming and blinking, but in reality there is on more stop of headroom available, which is dozens of information given the physics of the sensor...

[hcubell]

I believe camera manufacturers calibrate the histograms, the overexposure warnings and the blinkies on a conservative basis. The idea is that you can always push the exposure in post if there is actually more headroom, but if you blow the highlights in the field, you can never recover them. I also would not rely too heavily on blinkies, as it can often be difficult to see a very small area of blinkies.

[OleBe]

Today there had been more tests. White paper and constant light - no ambient. From the first movement of the histogram away from black (blackpoint clipping gone away) in Lightroom to last 1/3 stop before white clipping I have measured 10 stops of real life dynamic range.

The highlight warning in camera starts to go of at 2,6 + exposure. However in reality you can go easily to +4 (!) without blowing anything. The preview image in camera looks terrible though.

[JCM-Photos]

Why should a specular highlight no be blown out ?

It should be to appear as a direct sun reflection.

[OleBe]

Why should a specular highlight no be blown out ?

It should be to appear as a direct sun reflection.

Specular highlights can be blown out. That is for sure. Regular highlights like white clouds in camera review and histogram are nearly white at about 2,6 + if spot metered whereas in Lightroom they are the same at about +4 metering in camera. This is what I mean. Camera says it is overexposed but it is not - with a great distance.
This will bring you a lot more information in shadows if the scene exceeds the dynamic range of the camera.

[Michael H. Cothran]

OleBe - Great inquiry. Here's my two cents, as I, too, always meter the highlights outdoors (when applicable, of course) to prevent them from clipping -

Take a spot meter reading of brightest highlights (NOT specular) where you want to preserve any amount of detail. The brightest part of white clouds is ideal. Open your exposure 2.5 to 3 stops. This will always prevent clipping your highlights, but still render them as highlights.
I learned this from instructions on using a hand meter with a spot meter built in (Sekonic 758-DR). The Sekonic's 1-degree spot meter is more accurate than a camera's spot meter, but the same principles prevail.

The X1D is capable of over-exposing a strong highlight 3-3.5 stops before any danger of clipping (I've done testing with mine). So, spot meter a bright highlight, then over-expose anywhere from 2.5 to 3 stops, depending on just how bright you want this highlight. This will yield the results you are after, while maintaining white values of 245-250 or even less. In post, you rarely ever want your highlights to go above 245-250. Conducting your own tests will indicate just how much you should over-expose to yield the results you're after. But anything up to +3 stops on a bright highlight should always be safe. To test, bracket, perhaps in 1/3 stop increments from +2 to +3.5 stops, then check the white values in post. Good Luck.

[JCM-Photos]

Olebe and Michael you are right and the X1D is very capable to spot meter a highlight with the required plus exposure compensation.

My own technique is
Mode A
spot metering
holded AE lock (to set in X1D menu)
rear wheel exposure compensation (to set in X1D menu)
focusing in MF

I point the brightest highlight area of te subject (not specular zone) with the spot area in the viewfinder
I dial in the required plus correction with the rear wheel
I press shortly the AE lock button for a permanent exposure lock
and it is done, I can go to back button AF focusing in MF mode

[OleBe]

I would like to come back on this one.

Tested this even further in the last couple of months.

What I do not get is, why the internal preview of the camera, including highlight warning is that off compared to the RAW file.

Would it not be much more useful if we could see the real white clipping point in the preview instead that assumed one based on the current JPEG preview?

Basically if we go by the current method we are all underexposing the sensor and are trashing a lot of valuable picture information into black.

Anybody has an idea why it is so difficult for the manufacturers to introduce a real RAW histogram and a JPEG based on the real sensor usage?

[PaulMcB]

The historical reason for a JPEG derived histogram is that it will clip the highlights and darks before they're actually unrecoverable, because the RAW shot has more information and latitude than the JPEG the camera is using to generate the histogram.

Given the ability to recover shadow and light details, it seems anachronistic.

[SrMi]

Highlight weighted metering typically underexposes an image. The idea behind ETTR is to saturate the sensor without clipping relevant highlights (at base ISO).
The best solution would be to provide histograms and highlight warnings based on raw data.
LrC and PS histograms are not very reliable. Use RawDigger or FastRawViewer to see if there is any clipping in your raw files.

[OleBe]

Exactly, they need to provide highlight warning and preview based on RAW and not on JPEG.

But it seems the industry is not interested...

To give an example of what I am talking about.

Situation bright sky with white clouds the white clouds are relevant highlights in this case.

If I spot meter on the brightest white cloud and dial in +2,4 exp comp the camera is at the edge of going to overexposure warning. That must be due to the JPEG preview and how they have adjusted the warnings.

However the sensor reaches saturation just over +4,0 stops spot measured on the same cloud.

Basically we are throwing away 1,5 stops of light, if we go for the method the camera likes.

When you import the picture to Lightroom it is 1,5 stops too bright, but since at no point the highlights were clipped in the raw, you can dial back exposure 1,5 stops and have a perfect exposure with maximum details possible.

What I have read, but cannot judge since I do not have Capture One nor a Fuji cam, they have a setting called linear response which in conjunction with Fujis natural live mode does basically import a ETTR file which reaches sensor maxed out data and imports it correctly.

I can only speak for myself, but I would love to be able to use the maximum of my cameras sensor DR and gladly accept that I need to measure relevant highlights extremely carefully to not blow them.