M240 light meter oddity

[edwardkaraa]

Advertisement (gone after registration)

I'm not trying to argue here, but I did read somewhere that modern sensors do not like overexposure, not because of clipping channels, but because there is a loss in color quality of the mid tones. My previous A900 certainly wasn't very keen on ETTR, neither does my M9.

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[theno23]

ETTR is good advice when shooting at a sensor’s native sensitivity (not so much at higher ISO levels as has been extensively discussed here). But then, ETTR means you have to keep an eye on the RGB histogram so you don’t overexpose – not even one out of the three channels. Dialing in a -1/3 EV correction as standard is sometimes recommended to prevent accidental overexposure when you are not checking the histogram. Both recommendations have their place; ETTR is generally the better advice but requires more care.

 

I don't think ETTR can be regarded as a good advice, in 2013. Using the RGB histogram won't let you see if you're preserving your colour fidelity, or your specular highlights, it won't register on any histogram you can see on a tiny camera screen. The quantity of pixels at that level is just too small to show, by definition.

 

The tiny gain in noise you would get if you theoretically managed to squeeze out that extra 1/3rd of a stop of dynamic range is massively offset by the loss of highlight curve. At base iso that extra tiny fraction of DR just isn't significant on a modern camera - especially the M-240, which has loads of DR.

 

10 years ago, when sensor DR was less than half what it is now it was worth giving up highlight detail for a bit less noise, but we don't have to do that anymore.

 

- Steve

[theno23]

I'm not trying to argue here, but I did read somewhere that modern sensors do not like overexposure, not because of clipping channels, but because there is a loss in color quality of the mid tones. My previous A900 certainly wasn't very keen on ETTR, neither does my M9.

 

The loss of colour quality, and clipping channels are due to the same thing.

 

The problem is that some channels clip out before others, and you get strange, and incredibly hard to correct colour shifts from it.

 

The classic example is pictures of red flowers, which come out a kind of pinky-magenta colour. This is caused by ERRT, and/or the camera's meter not being sensitive enough to red light, making the red channel clip out before green and blue.

 

That example is quite easy to detect on a RGB histogram, but more subtle things like specular highlights can't usually be seen on a histogram, but can be detected by more sophisticated meters - not the M-240's centre weight one, but the Liveview one probably can? I've not experimented with it that much.

 

The article I linked about why ETTR is a bad idea explains it more clearly than I can. It has ETTR and correctly exposed example shots, and shows the real, and obvious difference in image quality. It was written by one of the worlds best currently practicing photographic printers.

 

- Steve

[mjh]

I'm not trying to argue here, but I did read somewhere that modern sensors do not like overexposure, not because of clipping channels, but because there is a loss in color quality of the mid tones.

That would be news to me.

[mjh]

I don't think ETTR can be regarded as a good advice, in 2013. Using the RGB histogram won't let you see if you're preserving your colour fidelity, or your specular highlights, it won't register on any histogram you can see on a tiny camera screen. The quantity of pixels at that level is just too small to show, by definition.

Why would you care about specular highlights? These are brigher than the rest of the scene by a wide margin; it isn’t usually possible or even desirable to preserve the exact tonality or colour of specular highlights. That’s what zone X is for.

 

The tiny gain in noise you would get if you theoretically managed to squeeze out that extra 1/3rd of a stop of dynamic range is massively offset by the loss of highlight curve. At base iso that extra tiny fraction of DR just isn't significant on a modern camera - especially the M-240, which has loads of DR.

Depending on the subject matter, the gain can be considerably larger than 1/3 EV. And you still preserve all the highlights worth preserving (because that’s what ETTR means).

[120]

You need to read the ISO sheet on exposure to understand the difference between digital and film. ISO values are conceived for film and narrowly defined with densities etc. Digital "ISO" values are supposed to be equivalent to the corresponding film values without a rigorous method being prescribed...

 

shouldn't you read this document before panning it?

[jaapv]

Advertisement (gone after registration)

Strange post - how could I comment on it without reading it? You can download and read all ISO sheets if you wish. It is not a free download I might add; I read it years ago, I'm not going to shell out money to download it again.

[theno23]

Why would you care about specular highlights? These are brigher than the rest of the scene by a wide margin; it isn’t usually possible or even desirable to preserve the exact tonality or colour of specular highlights. That’s what zone X is for.

 

Not always - did you read the article I linked? It has a real-world example where ETTR makes colour noticeably less accurate.

 

Depending on the subject matter, the gain can be considerably larger than 1/3 EV. And you still preserve all the highlights worth preserving (because that’s what ETTR means).

 

I think the difference is in what "worth preserving" means - the tutorials on ETTR I've seen jam the histogram up against the right of the exposure range (as the name implies). If it preserved everything that matters there would be nothing to distinguish it from a "correct" exposure, and there would be no need of the buzzword.

 

This article (from 10 years ago note) Expose Right certainly shows an exposure compensation that clips the channels enough that it would lose colour fidelity.

 

A quick google for "expose to the right myth" shows many counter examples of when ETTR damages images, compared to a "correct" exposure.

 

I know that 10 years ago it was sometimes a good idea, I used to do it myself, but the noise / colour fidelity tradeoff is very different now. Now we can raise the shadows by 2-3 stops, enough to get an unnatural HDRish look to the image, but still be noise free, I just don't see the motivation.

 

Alternatively, a challenge for anyone that still shoots ETTR: come up with a M-240 ISO 200 image that is visibly better when shot with ETTR as opposed to a "correct" exposure.

[120]

Strange post - how could I comment on it without reading it? You can download and read all ISO sheets if you wish. It is not a free download I might add; I read it years ago, I'm not going to shell out money to download it again.

 

You spent $100 on it and then deleted it?

 

There are separate standards for film and digital sensitivity, not one "ISO sheet on exposure." There are five different standards for digital--the one you are apparently commenting on intentionally leaves things to the manufacturer's judgment. Four others you don't mention are for different applications, and are rigorously defined.

[jaapv]

Thanks for the confirmation of my remark that it is indeed left to the manufacturers' judgement.

[edwardkaraa]

Does anyone know how the iso works with the typ 240? the M9 is basically an isoless camera, meaning that shooting at iso 2500, or pushing the exposure by 4 stops in the raw converter from iso 160 basically produces similar results. In the case of the M9, ETTR doesn't really make sense as all you have to do is lower the iso for the same shutter speed/aperture combination.

[120]

Thanks for the confirmation of my remark that it is indeed left to the manufacturers' judgement.

 

Yes you got one out of five right

[MarkD]

At base ISO setting, my M240 takes about 1/2 (or more) stop exposure increase over what my M8.2 requires. Given that M240 is supposed to be ISO 200 vs 160 for the M8.2, that's pushing the difference close to a stop slower for the M240. I use manual 95% of time in sunny weather and noticed the difference right away.

 

I guess the shutter speeds would need to be checked though. The speeds above 1/1000 sec. my M8.2 were fast - I believe the 1/1000 to be correct and aligned well when I used incident meter to check. The comparison in the first paragraph was based on a shutter speed of 1/1000 sec.

 

Mark

[gt3photography]

The reason I came to your post, is because I'm having the same issue. Using a handheld light meter, I get a reading (ISO 400, Shutter 80, f-stop 4), I put those numbers in the M240 and the image is underexposed, and when I put the same numbers in my D800 the image is properly exposed. What gives? I've gotten use to overexposing what the hand held reads. Does Leica know about this?

[jbl]

I had a recent thread on this too. It seems the effective ISO is 2/3-1 stop lower than the rated ISO. For exposing from experience, I set to 640 and expose at 320 which matches my Tri-X experience. Irritating but no different than film as it turns out.

[Lagozzinojd]

My Sekonic L-758dr uses 18% as neutral gray (as most for film)-- the value in digital cameras is closer to 13%. I've calibrated my 758 by half-stop to compensate for otherwise opening up the M240 by half-stop from the 758 reading. Credit goes to my recent National Geographic Society Workshop which I found reading "Exposure Challenges" (pg 134) in their "Complete Phototgraphy" (2011 ed) book. With the calibrated 758 to correctly match a correct exposure in agreement with the M240, I haven't since used my 18% grey card except infrequently for color management.

[jrp]

For what it is worth, I quite often find that clicking "Auto" in the Lightroom exposure settings normally sets Exposure to +0.75 to +1.5. (I tend to use the M240's advanced metering mode.)