ISO 100 on M10

[Printmaker]

Advertisement (gone after registration)

I'm still waiting for mine to ship but in the mean time I was wondering is the ISO 100 setting a pull setting? If so, is there a loss of quality?

 

I'd go try it for myself but it is a 2,500 mile journey to my dealer.

 

Thanks in advance.

[Mr.Q]

No, ISO 100 is the native base ISO.

[Advertisement]

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

[adan]

Actually, my experiments today with my new M10 indicate that ISO 100 behaves like pull processing.

 

Whether Leica labels it as that or not. (And they do not).

 

Examples - same shot, with widely different WB settings in Camera Raw. Note that the light grays above about 240 do not change color with different WB - they are "locked" as a fish-belly gray. Especially the white parts of the "Mini-Market" sign, and the brightest parts of the door, and signs on the lamppost.

 

As compared to the slightly darker (white/gray) area under the "Marlboro" sign, which changes color with WB noticeably and happily.

 

Which looks pretty "signature" for pull processing. Same as the M9 at "Pull 80." Likely one or more color channels blew out, and Leica's in-camera processing has simply tamped down all the colors to be an equal gray.

 

I didn't notice this while shooting, only this evening, so I did not try ISO 200 as a comparison. I will tomorrow. But shots I made outdoors in shade at ISO 800 (with some sunny blown highlights in the background) showed a more normal behavior - the color only "locked" to equal amounts of RGB above levels of 250-252. And of course, shots made with the 21mm that picked up a lot of sky in metering and were dark enough that no part of the picture topped 240 in levels, also looked fine. As did shots with flatter lighting and no large area exposed to brighter than 220-238.

 

Not surprising, given the great high-ISO performance.

 

Used ISO dial for ISO setting, FW is the original 1.0.2.0

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[adan]

Similarly, this girl's arm were actually pale pink - but once exposed at a level of 247-250, it retained tonal variations, but the color was lost (it's all equal RGB) and no logner responds to WB changes.

 

100% pixel crop (NICE detail, 90 'Cron pre-APO!)

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[01af]

ISO 100/21° is the native base ISO.

 

No, it isn't. ISO 200/24° is.

 

 

Actually, my experiments today with my new M10 indicate that ISO 100/21° behaves like pull processing.

Yeah, same here (with firmware v1.3.4.0). At ISO 200/24°, exposure range is wider than at ISO 100/21°. So ISO 200/24° is the M10's sensor's native base ISO.

[adan]

I will say this - IF the lighting, metering, and general scene arrangement are such that the clipped highlights at ISO 100 are in small areas, or areas where "going gray" is not especially noticeable (white clouds, silvery steel building), then the ISO 100 "pulling" on the M10 can give the impression of very nice dynamic range. Those overexposed shadows "open up" nicely in post-processing.

 

This was with no exposure compensation at ISO 100 - between the bright sky in the metering area (non-LV metering), and the backlighting, the clipped gray areas are somewhat camouflaged. Just a hint of "steel gray" in the rimlighting of the skin, hair and green hat. But I will stick with 200 unless there is a compelling reason to use 100 (which for me, there usually is not).

 

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[UliWer]

Advertisement (gone after registration)

...

 

But I will stick with 200 unless there is a compelling reason to use 100 (which for me, there usually is not).

 ...

 

Sounds reasonable to me.

 

How does Auto-Iso deal with ISO 100?

 

Since one cannot set a "minimal" ISO-value I think Auto-Iso will regularly choose ISO 100 as long as the light and your maximal shutter time allow for it.

 

A camera regularly choosing a "pull" value which gives inferior file quality might be called faulty, or is this too harsh?

[adan]

For me, Auto-ISO is something that happens to other people.

 

But it seems like something (almost) trivially easy to fix in a firmware upgrade. Just add a line for choosing minimum ISO allowed. Lots of room on the Auto-ISO Setup submenu.

[Printmaker]

Thanks Andy. That is what I thought. It was puzzling when Leica listed the base ISO at 200 but then offered ISO 100 without labeling it as a "Pull." There is almost always a quality loss whenever there is a pull setting. Even back in the days of film when we would overexpose and underdevelop to gain shadow detail while controlling contrast, the results were far from ideal.

[01af]

A camera regularly choosing a "pull" value which gives inferior file quality might be called faulty, or is this too harsh?

 

Harsh indeed—but appropriate (unfortunately).

[digitalfx]

 

 

But it seems like something (almost) trivially easy to fix in a firmware upgrade. Just add a line for choosing minimum ISO allowed. Lots of room on the Auto-ISO Setup submenu.

 

 

agree, should be an easy fix, or at the very least the minimum ISO should be the base ISO. 

[Aktenschrank]

...

 

Since one cannot set a "minimal" ISO-value I think Auto-Iso will regularly choose ISO 100 as long as the light and your maximal shutter time allow for it.

 

A camera regularly choosing a "pull" value which gives inferior file quality might be called faulty, or is this too harsh?

 

Totally agree, i really hope Leica will allow us to remove the pull setting from Auto-ISO in one or the other way in the future!

[UliWer]

... It was puzzling when Leica listed the base ISO at 200 but then offered ISO 100 without labeling it as a "Pull." ...

Is there any Statement by Leica saying base ISO for the M10 was ISO 200?

 

I only found descriptions which don't make any difference for the M10  between the ISO value 100 and the higher values - as they did for the M9 or the M (Typ 240) calling ISO 80 or ISO 100 "Pull".

 

There was a long discussion here about ISO 50 for the SL. Descriptions by Leica don't call this "Pull". Some testers called it "Pull" others didn't. There were even different statements by Leica if it was "Pull" or not. I think in the end there was common consense, that it didn't matter, since you could not see any degradation of the SL files taken with ISO 50 compared to ISO 100 or 200. Some even said latitude or "dynamic" was slightly better with  ISO 50.

 

As far as I know Leica didn't change  Auto-ISO using ISO 50 as "base" by a firmware upgrade. I don't see anybody complaining.

 

For the M10 the situation might be different.

[Steve Spencer]

Is there any Statement by Leica saying base ISO for the M10 was ISO 200?

 

I only found descriptions which don't make any difference for the M10  between the ISO value 100 and the higher values - as they did for the M9 or the M (Typ 240) calling ISO 80 or ISO 100 "Pull".

 

There was a long discussion here about ISO 50 for the SL. Descriptions by Leica don't call this "Pull". Some testers called it "Pull" others didn't. There were even different statements by Leica if it was "Pull" or not. I think in the end there was common consense, that it didn't matter, since you could not see any degradation of the SL files taken with ISO 50 compared to ISO 100 or 200. Some even said latitude or "dynamic" was slightly better with  ISO 50.

 

As far as I know Leica didn't change  Auto-ISO using ISO 50 as "base" by a firmware upgrade. I don't see anybody complaining.

 

For the M10 the situation might be different.

 

I think this makes sense. We also know that DXO Mark rates the 50 ISO on the SL as measuring a having a lower ISO than ISO 100 and as having better dynamic range than ISO 100. Even if ISO 100 is "base" ISO on the SL, if anything there is a bit better performance of the SL at ISO 50. The same may or may not be the case for ISO 100 vs. 200 on the M10, but let's wait for some more testing before we draw any firm conclusions.

[edw]

This whole thread of discussion vaguely irritates me for two reasons. First, there are the conspiratorial post(s) that suggest the ISO 100 is some sort of scam. And then there are the post(s) indicating that we should wait around for Saint Sean Reid to enlighten us with Truth.

 

The reality is that three photographs can shed a lot of light on this question. So I took them. And I am sharing them with you; go forth and go all Zapruder film on them.

 

https://www.dropbox.com/s/6dezeypg7clvs5h/ISO%20100%20Test.zip?dl=0

(approximately 83 MB)

 

The photos:  A.) M10 photo taken at ISO 100 with "proper" exposure. This is at f4, 1/360 second. Latest firmware. B.) M10 photo taken at ISO 200 with same exposure. Also at f4, 1/360 second. Same camera, same firmware. C.) M-D photo taken at ISO 200, also 1/360 second. Latest firmware.

 

Suggested methodology: If the "ISO 100 is a pull" theory is correct, you should be able to process photo B in Lightroom in such a way that it lacks none of the highlight detail contained in photo A. Photo C was taken so that we can assess whether the M10 (photo  contains additional highlight detail than the M-D (photo C) at the same ISO.

 

Possible outcomes: I.) There is more highlight detail in A than B. IIa.) There is equal (or less) highlight detail in A than B, *and* B contains equal (or less) highlight detail to C. IIb.) There is equal (or less) highlight detail in A than B, *and* B contains more highlight detail than C.

 

Interpretations: I.) The M10's ISO 100 is correctly labelled. IIa.) The M10's ISO 100 setting is baseless marketing and should be labelled PULL. IIb.) I think reasonable people will be able to disagree here.

 

I will withhold offering my take, but I will say that this masturbatory pixel peeping exercise has led me have a deeper understanding of both the M10's and M-262's capabilities.

 

Edited February 28, 2017 by edw

[adan]

Thanks for the samples. They behave exactly like my M10 exposures (having now done some shooting at ISO 200 to compare)

 

There is a loss of COLOR information in the brightest highlights. The "gray" building panels are slightly blue outside of the specular reflections (e.g. 172/173/181 - blue net plus of about 8-10 levels, "As shot" WB), but pure gray (e.g. 248/248/248) in parts of the brightest highlights, that are not all the way to 255/255/255.

 

With what I would call "normal" or "accurate" in-camera AD conversion and creation of the .dng, that color difference should remain visible right up until the blue channel hits 255 - it should be possible to find an extreme highlight that is 248R/249G/255B. The color should remain slightly blue (less red and green than blue) right up until one or more channels actually clips to 255.

 

Instead, at any brightness above 240 (in the M10 ISO 100) and 250 (in the M10 ISO 200), the color variation is lost. The M10 (and the M262, for that matter) is artificially equalizing the color values to neutral gray, sooner than I'd like. I don't want a blue sky, for example, to turn gray before clipping actually sets in. A gray ring around a blown-out sun, for example.

 

There is "detail" - in that there are luminance variations  - 253/253/253 vs. 250/250/250. But it is crummy detail if it renders actual blues (or skin tones, or green hats, or any color) as grays. Normal for an M Monochrom - not good in a full-color M.

 

The fact that we lose 10 additional upper levels of color information at ISO 100 vs. ISO 200 (240 vs. 250) shows that ISO 100 is losing color dynamic range over ISO 200. It can only produce 240 levels with full color information - at ISO 200 it can produce 250 levels with full color information. Generally speaking, if you lose dynamic range at a lower ISO, that means you've entered the realm of pull-processing. The ISO with the largest DR is the "base" or natural ISO of a sensor.

 

EDIT - but that does not mean ISO 100 is a "scam" - most cameras offer at least one lower-than-base ISO, that cuts a bit closer to the edge than ideal, but offers creative options if handled carefully. Larger apertures for "bokeh" or longer exposures for blurred waterfalls. No more of a scam than putting on a 2X ND filter. Don't read in cynicism where there is none.

______________________

 

Now, in addition, I've noted that the M10 (and the 262) still clips color a bit at ISO 200. You still cannot find highlights above 250 net brightness, with color information or variation. They are also grayed out. Better than ISO 100, but not where I want to be.

 

There is a workaround, which is to permanently set an exposure compensation that guarantees that no part of the picture (as shot) is ever brighter than 250 (ISO 200) or 240 (ISO 100). Then - you get full color information in even the absolute brightest areas, and can brighten the highlights back up to as close to 255 as you want, while retaining the different RGB values - the color. My M9s spent their lives at -0.7, to achieve that. And thanks to the otherwise-better DR of the M10, there's a huge amount of shadow detail that can be recovered, without significant increase in noise, or loss of color fidelity.

 

I'm gonna try ISO 100 again, with exp. comp. of -0.3 or -0.7 dialed in all the time (effectively, a pull-125 or pull-160 exposure) and see which - if either - consistently keeps the brightest tones (except the sun, can't expect to hold any detail there) dark enough to retain the true colors. The juicier shadow exposure at 100 (or "125" or "160") is rather nice.

 

Personally, I would ask Leica (if they are listening) - what exactly is the deal with equalizing the color values in the brightest highlights? At any ISO? I can't think of any rational reason to do that.

Edited February 28, 2017 by adan

[edw]

Thanks for the samples. They behave exactly like my M10 exposures (having now done some shooting at ISO 200 to compare)

 

[...]

 

EDIT - but that does not mean ISO 100 is a "scam" - most cameras offer at least one lower-than-base ISO, that cuts a bit closer to the edge than ideal, but offers creative options if handled carefully. Larger apertures for "bokeh" or longer exposures for blurred waterfalls. No more of a scam than putting on a 2X ND filter. Don't read in cynicism where there is none.

 

Thanks for sharing your thoughts. I'm interested on your thoughts on the M-262 at ISO 200 vs the M10 at ISO 100. For what it's worth, I've never counted you among the "wake up, sheeple!" conspiracy theorists.

[adan]

Based on your one picture, the M10/M262 at 200 seem virtually identical - predictable color flat-lining at/above 250, occasionally at 247-248. Just poking around in those brighter highlights with an eyedropper.

 

The fact you gave ISO 100 and ISO 200 the same exposure makes it a bit tricky doing a direct comparison. The M10 was "expecting" more light when it processed the ISO 100 picture, and thus "underdeveloped" it - it is a stop darker overall. Although the most extreme highlights are still blown (such was the lighting). Plus of course, the M10 uses the M10 EMBEDDED calibration, while the M262's is different - although that should only affect overall color rendition, not exposure curves.

 

In that shot, the color equalization is still reliably present around and above level 240 - but even occasionally in the 230s and even 220s. Although I don't put much weight on those, due to the exposure, and the fact that, well, it's a weathered building - there may just be "yellower" spots that register as 228/228/228 when the normal original color of the panels registers 107/108/117 (blue) under the same WB. And the different calibrations - the variables start to stack up.....

 

But overall - more highlight color remains with the 262 @ 200 than with the M10 @ 100.

[Steve Spencer]

Maybe it is just me, bit it seems that the approach here i only looking at retaining highlight detail, but you have to look at retaining highlight detail and the ability to recover shadow detail at the same time, it seems to me, to know the full dynamic range at each ISO. I think it is totally possible for Leica to have built the camera sot that there is not better or even a bit worse retaining of highlights at ISO 100, but much better ability to recover shadow detail at ISO 100 and in this way for there to be more useable dynamic range at ISO 100. Am I missing something? Don't you need to consider retaining highlights and recovery of shadow detail when considering useable dynamic range? To me pull ISO is defined when the camera has equally sensitivity to light at the 2 levels of ISO, and the "lower" ISO has less useable dynamic range. These tests which seem to focus only on retaining highlights don't seem to be able to address that issue.

[Advertisement]

Hallo,

Look here: Leica M

[adan]

Fair point - but you do have to start somewhere, and with digital, the highlights are always the critical areas. Especially if the color detail gets truncated before the brightness detail. It's obvious the M10 does something different with the highlights at 100 vs. 200 or higher.

 

Once it is determined what exposure will not produce gray highlights - then we can see what is left over for shadow recovery. In the M10, ISO 100 starts out a little behind in that regard - but may catch up in the shadows. Thus my plan to try exp. comp with 100, and see what is still available for recovery.