If you want something done right....

[adan]

Advertisement (gone after registration)

...do it yourself.

I got to borrow an M11, and compare its very-low-light, high-ISO noise to my original 2017 M10.

The experimental question was simple - "Will the M11 allow me to shoot in dim light, at motion-stopping shutter speeds, in color, with less noise than the M10?"

This test is not controlled for, or intended to demonstrate, any other camera characteristic, except high-ISO noise.

Set-up - took both cameras into a camera-store storeroom, and adjusted the light (or darkness ) to match many of the venues I shoot in.

At ISO 10000 and f/4.0 (the aperture of my 135mm Tele-Elmar - although the test shots were made with the M11 owner's 50mm f/1.0 Noctilux, stopped down to f/4.0), that light level requires a shutter speed of 1/90th sec. Being able to use ISO 20000 or higher would be solve that, but I have generally limited my use of the M10 to 10000, for noise and banding reasons. That gets to be borderline for stopping subject motion.

"Image stabilization" will not be a help - my subjects in those venues (or similar) are people in motion (rock performers, and similar). I have to have a higher ISO to get the shutter speed up to at least 1/180-1/250 to minimize subject blur.

For this test I therefore used ISO 20000 with both the M10 and M11. Exposure was shutter speed 1/180, and f/4.0. Images are straight from the camea with no noise-processing or sharpening. Just  white-balanced to the same temperature (4050°K, tint +48 (magenta)). I shot the M11 as both18.4 Mpixel small DNG and 60 Mpixel large DNG. The M10 was, of course, a 24 Mpixel image.

The images below are similar crops at 100% view.

The answer appears to be "No." The M11 does NOT improve over the M10 in terms of noise at ISO 20000. If anything, the M11, tp my eye, performs a tiny bit worse,. Although with some processing it can probably come close enough to equal the M10 for practical purposes.

That doesn't mean the M11 is bad - simply that it is not a significant upgrade for my needs. It won't move the noise needle over the M10.

Pix are split into the M10 sample, and then M11 samples as shot in the two file sizes, so as to minimize any compression artifacts.

M10 output. Click for best view. Again, ISO 20000.

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]

Now the M11 examples at ISO 20000.

Again, click image for best reproduction.

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!

[Advertisement]

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

[Jeff S]

Well, if you want to shoot b&w, the M10 Monochrom is the king of high ISO/noise performance in the current Leica portfolio, as David Farkas has so far shown in his ‘showdowns’.

Jeff

[anickpick]

If you compare M10 to M11 files at ISO 10'000 or even at ISO 20'000 and there is no need for post processing, the M10 files look cleaner. Imo with todays noise reduction AI software, noise is not such a big deal anymore.

However, if one shoots at ISO 10'000 and has to increase exposure or lift shadows in post, the M10 files fall apart, while the M11 files are holding up well.

So I actually agree with you: It really depends on your specific needs.

Edited May 4, 2022 by anickpick

[adan]

If I wanted B&W, I would have made the test that way. And wouldn't have carefully specified "color" in my experimental question.

The M11 has upped the game in many dimensions - just not the particular one where I have the greatest "felt need" at the moment, and hoped it might help. Especially with the 18-Mpixel "pseudo-M11-S" option. So it goes.

I will look into other "options in the current Leica rangefinder portfolio," such as the 135 APO-Telyt (half-stop gain in light over my TE) or the goggled 135 Elmarit-Ms (full stop increase; massive - but also quite inexpensive). Ahh, for a 135mm APO-Summicron-M that a rangefinder can focus reliably.

Although not the point of the test, I was interested to see that the M11's smaller pixels do what it says on the box regarding moiré. In that sense they do "outresolve" some lenses, in a good way.

I was also interested to see that the M10 - with "honest" ISO 20000 correctly exposed in the first place - does a little better than I'd thought. When necessary (and with care) I can try that route more often.

[Jeff S]

2 hours ago, adan said:

If I wanted B&W, I would have made the test that way. And wouldn't have carefully specified "color" in my experimental question.

 

I’m well aware. If I thought you WANTED b&w, I would have started my comment with “SINCE you wanted b&w…”, not IF you wanted b&w”, or maybe I should have more carefully wrote “If you HAD wanted b&w”. In any case, the point of my comment was that the M10 Monochrom shows the least noise at high ISO of any current Leica, whether in the M, SL, CL, Q or S line…it’s the king…but obviously one needs to want to shoot b&w.

Jeff

Edited May 4, 2022 by Jeff S

[hdmesa]

Advertisement (gone after registration)

I don't think noise is the biggest problem here with the M10 shot:

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]

Already noted.

Quote

Although not the point of the test, I was interested to see that the M11's smaller pixels do what it says on the box regarding moiré. In that sense they do "outresolve" some lenses, in a good way.

 

[hdmesa]

3 hours ago, adan said:

Already noted.

 

13 hours ago, adan said:

...Although not the point of the test, I was interested to see that the M11's smaller pixels do what it says on the box regarding moiré...

I couldn't make sense of "the M11's smaller pixels do what it says on the box" and kept reading it wrong – as if somewhere on the Leica M10 box there's a warning about moire. As for in the photos, it looks more like the basket has the moire and the box just has noise 🙃

 

[Tailwagger]

What I'm scratching my head over a bit, and I find quite interesting, is the utter loss of resolution in the 18MPx shot from the 60 MPx one.  At normal ISOs, I've seen nothing like the loss of detail found in the sticker.  One might have suspected, as I assume you were hoping, that the binning would reduce noise, but instead it appears not only to have amplified it, but killed any detail in the process. I'd guess they really didn't spend much time, if any, trying to utilize binning for reducing noise in high iso scenarios. Perhaps future firmware updates might improve things, but who knows. I doubt they really care very much about this sort of thing.

[Kwesi]

21 hours ago, adan said:

...do it yourself.

I got to borrow an M11, and compare its very-low-light, high-ISO noise to my original 2017 M10.

The experimental question was simple - "Will the M11 allow me to shoot in dim light, at motion-stopping shutter speeds, in color, with less noise than the M10?"

This test is not controlled for, or intended to demonstrate, any other camera characteristic, except high-ISO noise.

Set-up - took both cameras into a camera-store storeroom, and adjusted the light (or darkness ) to match many of the venues I shoot in.

At ISO 10000 and f/4.0 (the aperture of my 135mm Tele-Elmar - although the test shots were made with the M11 owner's 50mm f/1.0 Noctilux, stopped down to f/4.0), that light level requires a shutter speed of 1/90th sec. Being able to use ISO 20000 or higher would be solve that, but I have generally limited my use of the M10 to 10000, for noise and banding reasons. That gets to be borderline for stopping subject motion.

"Image stabilization" will not be a help - my subjects in those venues (or similar) are people in motion (rock performers, and similar). I have to have a higher ISO to get the shutter speed up to at least 1/180-1/250 to minimize subject blur.

For this test I therefore used ISO 20000 with both the M10 and M11. Exposure was shutter speed 1/180, and f/4.0. Images are straight from the camea with no noise-processing or sharpening. Just  white-balanced to the same temperature (4050°K, tint +48 (magenta)). I shot the M11 as both18.4 Mpixel small DNG and 60 Mpixel large DNG. The M10 was, of course, a 24 Mpixel image.

The images below are similar crops at 100% view.

The answer appears to be "No." The M11 does NOT improve over the M10 in terms of noise at ISO 20000. If anything, the M11, tp my eye, performs a tiny bit worse,. Although with some processing it can probably come close enough to equal the M10 for practical purposes.

That doesn't mean the M11 is bad - simply that it is not a significant upgrade for my needs. It won't move the noise needle over the M10.

Pix are split into the M10 sample, and then M11 samples as shot in the two file sizes, so as to minimize any compression artifacts.

M10 output. Click for best view. Again, ISO 20000.

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!

 

 

Hi Adan,

Do you remember if noise reduction was turned off in the M11 menu?

[adan]

At the risk of repeating myself....

21 hours ago, adan said:

This test is not controlled for, or intended to demonstrate, any other camera characteristic, except high-ISO noise.

That 18Mpixel image is probably slightly front-focused - I was not testing for resolution in these low-light shots.

I was kinda juggling my M10 plus the $14000 worth of 50mm Nocti and M11 belonging to someone else in the relative dark, so I'll take the blame for the blur part.

Although that won't change the noise, which exists only in the plane of focus (the sensor) anyway. I could have photographed a detail-free gray card, out of focus.

Here is an 18 Mpixel "bracket" from the test, which is sharper for the stitching and printed number (but not the right ISO for comparison across cameras (it is ISO 40000)). But identically processed (zero noise reduction or sharpening).

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!

For Kwesi - I don't know how the M11 was set for NR.

I hope it was "off" because as I also mentioned in post #1, this was a test of each sensor with "no outside help." No NR or sharpening, only what comes off the sensor raw.

I can go back and look at the images with noise processing added.

 

[Kwesi]

@adanI just dug a little deeper in the manual pp76-77 and actually based on your shutter speed and ISO it wouldn’t have made a difference. 

[adan]

A partial update - won't have examples ready for prime time until tomorrow (but they are saved and ready for cropping).

Checked the M10 and both resolutions of the M11 at ISO 40000 - AND applied color and luminance noise reductions (standard Adobe Camera Raw tools, not additional software).

Good news for both - all that blobby glowing color noise goes away very easily. But you knew that.....

Good news for the M11

- the additional DR is clearly visible in the deepest shadows compared to the M10, in either file size.

- the M10 can do ISO 40000 better than I thought - but shows some occasional banding that the M11 does not.

Bad news for the M11 - the tendency to exaggerate or reveal CA at 60 Mpixels (as already described by many elsewhere) shows up

Good news for the M10

- absent the occasional banding and weaker shadow separation, the ISO 40000 noise speckles per se are much tamer than I thought possible and the pix are pretty usable, for journalistic work. And that gains another stop over ISO 20000 (and two stops over my previous limit of 10000 - getting a quick graduate course in "noise avoidance and correction"). Pixel for pixel, actually about the same for all three sensors (but then the pixel count will kick in, for prints of the same size - advantage M11 60Mpixels).

- that nasty white-suitcase moiré we noticed in the first round disappeared (??!!)

Bad news for all three - None of them is going to win the "medium-format Velvia 50" award this year. But in ISO 40000 lighting, I guess that isn't the point.

The color rendering difference between the two cameras is kind of a pain to balance fairly - the M11's "M9 colors" include the M9's rather noticeable magenta shadows**. And fiddling too much with that then also influences the shadow noise. But that is the kind of thing it usually takes me a couple of weeks and 100 pictures to sort out with any new sensor, at any ISO. Can't be done in two days with a half-dozen samples.
__________________
** To steal from Biff Tannen (Back to the Future movies): "Magenta! I HATE magenta!!"

Edited May 5, 2022 by adan

[chris_tribble]

Adan - interesting. Thanks. Really interesting.

I think a lot of this shows the problem all makers face at the moment where there’s a marginal return on investment in innovation or more MP / higher ISO etc… For me, the M10 design is at the apex of digital RF, and we could all stick with this and make award winning images with the lenses we’ve got - and never need to buy another camera until we either dropped it once too often or it suffered catastrophic mechanical failure. However, as capitalism doesn’t work that way,  desire has to be generated and we all get sucked in.  I bought the M10-R for the improved highlights and quiet shutter - but I probably didn’t need to.  I made other decisions for what I think have been good photographic reasons.  I decided to run M10-R + Q2 and 2 SL2 bodies with 24-90 and 90-280 because this set up is helping me meet clients’ and personal needs better than 2 M10 bodies + Canon 5 mk 4 + zooms.  

Will I ever buy another camera? At the moment it really think not (apart from for the reasons set out above). I was involved in testing the M11 and got to know it very well.  I bought the M10-R.

I’ll go out and take some photos…. There’s a project I’m working on 🙂

[darylgo]

21 hours ago, adan said:

Ahh, for a 135mm APO-Summicron-M that a rangefinder can focus reliably.

Never say never but the above is an example of 100% highly unlikely.  I’ve been tempted to try the 135/2 Zeiss, a well regarded lens. I know it has no rangefinder coupling but the evf solves reliability.   

[hdmesa]

In casual shooting at high ISO in a very dark room with lamp light sources, the M11 files can get noisy, which not very shocking given they are 60mp color files. They look way better than my Q2 images in the same lighting condition.

When shooting the M11 in very low light at ISO over 3200, I've simply started shooting in 36mp and 18mp modes – not because they look better than downsized 60mp files but because this way I don't have to look at so much noise on the rear LCD and be annoyed by it 🤫🙃

[adan]

6 hours ago, darylgo said:

Never say never but the above [135 f/2.0] is an example of 100% highly unlikely.

Of course.

Which is why I'm trying to squeeze another (ideally) 2 f/stops worth of image (10000 > 40000 = f/4 > f/2 ) out of the sensors.

[anickpick]

vor 1 Stunde schrieb adan:

Of course.

Which is why I'm trying to squeeze another (ideally) 2 f/stops worth of image (10000 > 40000 = f/4 > f/2 ) out of the sensors.

I have a question for you, please don't take it as criticism - I am honestly curious: You wrote in another thread that 135mm is your preferred focal length - why are you using a Leica M camera? There are fantastic, faster, easier to use 135mm lenses, for example the ones by Sigma or Sony (both f1.8)? 

Edited May 5, 2022 by anickpick

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[adan]

Twenty-one years ago I dropped every camera system I had in favor of the sole remaining system based exclusively on split-image manual focusing - Leica M rangefinders.

In the process I discovered and fell in love with a certain color and contrast rendering I can only get consistently from Mandler-era Leitz lenses, including the 135mm Tele-Elmar.

I'm simply not interested in any other approach to photography if it can possibly - through ingenuity - be avoided. (This thread is me applying ingenuity and brain-grease to the M10 or M11 to see how far I can push them. )

Unless it is something completely outside the realm of what the M rangefinders do - i.e. lenses longer than 135mm (for which I have a Canon 5D2 and 300mm - last actually used maybe seven years ago).

Or 6x6 medium-format film work. Although even there, I stick to split-image manual focusing (Mamiya 6 rangefinder, Rolleicord with a classic split-image ground glass), or a Hassy SWC with optical window finder, and no focusing except guessing the distance and setting it on the focus ring.

It is rather like - why do people still paint with oils, when there are "fantastic, faster, easier" acrylics, or even digital sketch pads. It is the way I want to work (and have always worked, and am skilled at) to get the images I seek.