1 EV darker photographs

[jaapv]

Advertisement (gone after registration)

I explained in an earlier post. The ISO rating is less rigorously defined on digital than on film. In practice it means that a deviation of +/- 1 EV value is still within the norm.

This is nothing new. The M8 would typically produce the same exposure one ISO value below the Canon cameras of the era. So ISO 200 on the M8 was rated ISO 400 on a Canon Rebel. It was discussed extensively in this forum. Over time, Canon (and I suspect Leica too adjusted their rating gradually and the values converged.

It is quite possible that the M10 will show a bias compared to for instance Nikon. and that both comply with the ISO norm

[schattenundlicht]

The chapter „Digital camera ISO speed and exposure index“ halfway down the Wikipedia entry on „film speed“ https://en.m.wikipedia.org/wiki/Film_speed might help to clarify jaap‘s point with regard to ISO ambiguities with digital systems.

 

Revisiting the complex topic from a practical standpoint, I prefer a camera that I will use +/-0 EV most of the time to one that I permanently preset to -2/3 EV on theoretical grounds, but I never felt that to be an issue of any importance in a decade of using Nikon DSLRs as my main cameras. You have to learn each new body as you have to learn each new lens.

Edited March 22, 2018 by schattenundlicht

[Advertisement]

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

[pgk]

The ISO rating is less rigorously defined on digital ....

 

 

Its defined alright, but there are a variety of ways of both defining and assessing ISO, none of which really deal with the fundamental problem which is that we are trying to mimic an analogue system using a digital one. The whole concept of inbuilt ISO (as in specified film 'speed' based on chemical sensitisers incorporated into the emulsion and resulting in 'fixed' ISO ratings, which can be defined) versus gain applied to a digitally derived imaged is daft. We should have started again and not attempted to go down the mimicing route. Imagine having an amplification control on cameras together with uploadable custom amplification profiles (shadow boost/highlight reduction at various boost settings) to fine tune images straight out of camera - that would have been an interesting development (no pun intended).

That is not what I say. I just say that the brightness is the same. I do NOT say anything about correct exposure. I probably did not make myself clear enough.

 

The word which lacks clarity is 'brightness' - it is undefined and doesn't tell us anything other than giving an overall impression of the view of the image. I would agree that the files from my Leicas have always looked 'darker' on initial viewing of the RAW files, but the files themselves are more malleable and adjustments can be made whilst still retaining highlights and lightening shadows. 'Brightness' is not useful terminology I'm afraid.

[jaapv]

Thanks for clarifying, Paul. I think that we are saying the same. A norm which can be interpreted in a number of ways is, to me, undefined.

[pgk]

Thanks for clarifying, Paul. I think that we are saying the same. A norm which can be interpreted in a number of ways is, to me, undefined.

 

Absolutely. Comparisons are difficult unless everyone is open and manufacturers specify how they determine things and preferably all also supply multiple specification (different ISO interpretation data - which they won't/don't/can't.

[M11 for me]

 

. . . . 'Brightness' is not useful terminology I'm afraid.

 

 

 

Well, that is certainly correct and not so ideal at the same time. When we look at the equivalence theory then we compare different sensor sizes and assign the respective values that always refer to the 35mm format. We do this for ISO, and aperture. If we say now that the term brightness has no meaning (you say: "not useful terminology . .") then that theory would completely collapse. 

 

Still in practice all of that interesting discussion does not help a lot and is rather theoretical. I am extremely happy about how I can work on my M10 DNGs in Lightroom. And that is what counts for me and for most of us.

[pgk]

Advertisement (gone after registration)

Still in practice all of that interesting discussion does not help a lot and is rather theoretical.

 

Actually I have routeways through post processing that I certainly apply which are both relevant and practical. It all depends on whether we use what the camera produces or a modified version of it. My modified versions are anything but theoretical, although I base them on both theory and my particular view on how my images should look.

[jjesp]

Old thread. But found it since my M10-P has the same. I am not talking metering. If I shoot manually in sunlight as I often prefer, on iso 200, 1/500 and f/8 (both on Voigtlander and Summicron lenses), the images is about 1,5 stops darker. This is compared to Leica Q, Ricoh GRIII and all Fujis. So when going out in the streets with the M10, I have to put the ISO on 400 or 640, to have the same exposure. Maybe not a defect, but strange - and a little annoying.

I can of course just raise the exposure in post with one or two stops. But why? Isn't Leica just cheating a little here, to brag about the better lowlight on high iso settings in marketing? ISO 6400 on the M10 is clearly not ISO 6400 on any other camera.

Has anyone compared it to the M11?

Edited December 22, 2022 by jjesp

[jjesp]

22 minutes ago, Musky said:

It’s tough to compare because the M10 is always on Center Weighted meter and it measures the light bouncing off the shutter blades. The other cameras you mentioned meter off the sensor. 

You can try taking it in live view and use average meter? See what it does. 

 

Again, I am not talking metering. I am just looking at two images, shot manual - no metering. And the exact same "values" in ISO, shutter speed and aperture. The M10 image is at least one step darker. But of course, I can live with it, knowing it. But just wondering why the M10 is that much different than any other digital camera I used. Including Leica Q...

[Adam Bonn]

The M10 has quite an aggressive native tone curve that crushes shadows and mid tones

Things such as the differences in various digital ISO standards, lens T-stop and what the camera pre-bakes into the RAW file mean that sharing exposure settings between cameras won’t always produce an identical exposure.

If I set my M9 to ISO 160 and my M10 to ISO 200 and use the same lens to shoot the same thing at the same time with the same aperture and SS settings the M10 is appreciably darker (despite a ‘brighter’ ISO)

iirc (….) the hunch is that seeing as the M10 doesn’t exactly have class leading exposure latitude in the highlights and that the M10 arrived after the 240, that was often criticised for flat images, the M10 is intentionally configured to produce a darker frame to add contrast and protect the highlights 

FWIW I generally tended to adjust the global brightness in post for scenes where I’ve protected the highlights, used ISO100 to maximise the contrast on flat light days and set either +1/3 or +1/2 EV in scenes where the mid/dark tones where more important to me than the highlights.

But most often my ‘set and forget’ way of working with the M10 was ISO200 zero EV comp and meter of the part of the scene that would give me what I wanted.

Of course OMWV

[shanefking]

4 hours ago, jjesp said:

Again, I am not talking metering. I am just looking at two images, shot manual - no metering. And the exact same "values" in ISO, shutter speed and aperture. The M10 image is at least one step darker. But of course, I can live with it, knowing it. But just wondering why the M10 is that much different than any other digital camera I used. Including Leica Q...

Yup, this is known.

“ISO” was an international standard for film sensitivity rating, now its more of an abstract concept, simply to give an idea of the sensitivity settings.  Kodak Tri-x 400 was the same in a leica or a nikon, but now a sensor in a leica or a nikon will have entirely different manufacturing, and teams of engineers with different priorities.  
 

Personally I shoot by the histogram, and adjust the ISO appropriately.  Often what appears like crushed black in the preview has plenty of latitude in the actual file.

 

[jaapv]

3 hours ago, Musky said:

It’s not film so now ISO100 is whatever the manufacturer decided it was  

Basically yes, but not quite.The ISO norm suggests an equivalence between film and sensor output, but as sensor output runs through a series of digital steps,including amplification, the outcome can differ. Especially as the ISO norm assumes equivalence.

 

https://photographylife.com/equivalence-also-includes-aperture-and-iso

[Adam Bonn]

13 hours ago, Musky said:

His example is what you see. Same settings and -1 or even -2 stops off.

Stops off what though? That's the problem!

It's ultimately like saying the velocity of 100 mph is 60 increments off of the velocity of 160 kph.

Perfectly accurate statement, but the actual take away is that 100mph and 160kph is the same, also if you're traveling at either 100/160 and you wish to reduce velocity by 50% you need 50/80

Much like speed, temperature and weight digital ISO has no universal standard - just a series of values that work in conjunction with other values of the same scale (or require conversion to a different scale)

Likewise when we meter for a scene (be that light meter, external light meter, experience) we end up with ISO values that deliver what we want. The fact that these values might not swap between cameras or brands is a little annoying but experience compensates pretty quickly.

As long as +/- one stop of ISO produces +/- one stop exposure we should be able to cope..

If we review the DXO ISO test scores for the M10 (https://www.dxomark.com/Cameras/Leica/M10---Measurements)

We see that Leica's values are indeed approx. one stop higher than DXOs measurement scale (like kph to mph don't sweat it) but 400 is still a stop more than 200 and a stop less than 800 (more or less) so it all works ok, ie raising/lowering a stop of ISO has the desired effect on my photo

[Alberti]

On 3/21/2018 at 6:58 AM, Gobert said:

I have the same experience, but I’ve considered it as a learning curve.

Nothing beats the accuracy of Nikon’s light metering (no reason to return to Nikon although).

What I remember is that (also) the Nikon has a difference between auto-exposure and manual metering.

[jip]

This has nothing to do with the metering of the camera, Leica lies a little bit when it comes to ISO sensitivity on the M10. 

The M10 is always 1 stop slower than the set ISO value, of course the integrated light meter accounts for this, but when metering with sunny 16, or a external light meter the M10 is 1 stop less sensitive than it should be. Most digital cameras have about 1/3 stop less sensitivity than advertised this helps protect highlights. Leica M10 has a whole full stop less sensitivity than advertised.

ISO 400 frankly is 200. etc. 

As is also very easily visible on this sensitivity graph: https://www.dxomark.com/Cameras/Leica/M10---Measurements

The M10 ISO sensitivities:
100 = 89
200 = 105
400 = 201
800 = 402
1600 = 787
3200 = 1566
6400 = 3274

 

When correct exposure can be achieved at: ISO 400 - F/8 - 1/125 on most cameras and film on this planet earth you have to add one stop of light, be it shutter speed, aperture, or iso on the M10 to get the same result.

I've learned to use 1 ISO stop higher than written on the ISO dial with M10, then I can use sunny 16, or external meters just fine etc. 

Edited December 29, 2022 by jip

[Adam Bonn]

But there isn’t a universally recognised set of digital ISO values.

Expecting all cameras to behave the same or expecting sunny 16 to always be the same is a bit like taking an American car, a euro car and a ship, making them go along until the speedo says “25” then afterwards wondering why 25 mph/kph/knots isn’t  the same actual velocity 

See also weight, distance, temperature etc 

Even this digital rhetoric of brand x cheats at ISO is rather disingenuous I mean sure when brand x uses an ISO metric that implies the ISO can go to a very high value I’m sure it’s a great sales technique, but by the same token the same camera has a base ISO that appears to start far higher than it would with a different measurement scale, which has less appeal in the marketplace 

As long as increasing/decreasing the ISO value a stop has the desired effect then it works properly and if one is finding that one’s exposure choices from a different camera/film don’t exactly carry over to their current camera then one tweaks their exposure choices to suit.

Most cameras meter for the jpeg, the iirc >M9 M range doesn’t have a RGB histogram (and if they did it would be the jpeg), the clipping blinkies are the jpeg, LR/ACR, C1, etc do stuff in the background to the RAW file and the off the shutter blades metering of most M cameras can get confused.

So based on all of that a degree of experience with a camera, with the lenses (vignette, T-Stop), and with the RAW software is always required to get consistent and desired results.

Experience usually prevails! 

[pgk]

1 hour ago, Musky said:

It should be called GAIN. Gain is the appropriate term for digital photography. 

No. Digital ISO is defined: https://www.iso.org/obp/ui/#iso:std:iso:12232:ed-3:v1:en and you can buy fuller information here: https://www.iso.org/standard/73758.html 

Gain is simply amplification and can be varied to suit the image either incamera or using software. 

[pgk]

1 hour ago, Musky said:

Not really. It talks about “camera sensitivity” or “photographic sensitivity” that is sometimes referred to as ISO “for historical reasons only”. That’s what your document says. basically it says that calling it ISO is nonsense. 

ISO is a misnomer at the best of times (it simply refers to the International Organisation of Standardisation) but there needs to be some standardised way of comparing 'base sensitivity' of digital sensors, even if gain or amplification is then used on this afterwards (I don't think that it is quite that simple either). But once proprietory image processing is added into the equation it quickly becomes rather trickier to compare between models and makers. Add in CCD and CMOS and more. Complexity.

I still like my M9 files because I am very used to them, they are very malleable and I'm generally a low ISO shooter. Experience helps no end with digital files.

[charlesphoto99]

It's a bit like audio amplifiers. Two may have the same specs of 50W a channel, but play very differently. It's all about the implementation of those watts - i.e in large part the volume will come from the amount of gain in the preamp, and power regulation in the amp itself. An Onkyo AV amp will play at volumes differently than a Naim, at the same specs of watts (and the Onkyo almost guaranteed of peaking and distorting long before a Naim). 

[Advertisement]

Hallo,

Look here: Leica M

[Adam Bonn]

10 hours ago, Musky said:

Everyone’s camera base is 0dB. All you’re doing is increasing the sensitivity after that (by turning up the gain)  

Yes like an amplifier, but worth noting that the sensor has some degree of amplification to even be base ISO and that the RGB channels are not always amplified equally.

If you take a base Sony chip, let’s say the one used in the Fuji, the NEX5, the Nikon D5100 (iirc they’re all the same) then compare the ISO numbers and base ISOs and DR we see that SO MUCH of what is thought of as ‘base’ is actually all to do with implementation and not anything more physical than code

10 hours ago, Musky said:

The confusion comes from using film iso numbers on digital sensors

Maybe it wasn’t worth quoting me as I’m clearly not confused 😅 but like I wrote, the problem comes from expecting something that works on camera or film stock XYZ to perfectly carry over to something different, without really understanding that just because the numbers are the same it doesn’t mean that the numbers work the same (eg KPH/MPH Ilbs/KG kelvin/celsius etc)