Really “Is it the end of M road”?

[IkarusJohn]

Advertisement (gone after registration)

True.

 

PS - response to Andy.

Edited August 5, 2018 by IkarusJohn

[Advertisement]

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

[pgk]

Correct me if I'm wrong but I believe that 24MPs is approximately equivalent to the resolution of 35mm film stock

 

Depends . Argument's still raging but I reckon 10MPixels is equivalent to 35mm Velvia 50 myself based on my own testing for scientific purposes.

Edited August 5, 2018 by pgk

[IkarusJohn]

... If you print very large prints, especially where fine detail is needed like landcapes, then 24mp is pretty useless ...

 

 

Not even remotely my experience.  I have a Monochrom print on my wall (a mere 18MP), printed 105cm on the long side, and the detail is astonishing.  No resolution issues whatsoever.

 

It's all very well using a large format camera and enjoying the resolution you get - that's fine.  But to say "24MP is pretty useless" is rather sweeping.  Might not work for you, but for a lot of us the resolution is just fine, and balanced against dealing with large files and getting a sharp image, Leica has made a good choice setting the M and SL cameras at 24MP.

 

Now, if your normal landscape process is a tripod, ground glass screen, loop, multiple zone exposure readings and assistants, that's also  fine.  But some of us work to simpler procedures and get more than acceptable results.  What begs the question is why would you want more MP in a compact, relatively limited camera like an M when you  use a large format camera for your work?

Edited August 5, 2018 by IkarusJohn

[dkmoore]

In think this post :

 

https://www.l-camera-forum.com/topic/287222-solo-exhibition-coming-up/

 

proves Mr. DigiL 100% wrong. (And the MP brigade on this forum too, I would suggest) The same 24 MP produces a great exhibition.

The proof of the pudding is in the eating - this one appears to be rather tasty.

 

 

 

Wow, this is some good work. Thanks for re-posting as I missed that.

[Jon Warwick]

Not even remotely my experience. I have a Monochrom print on my wall (a mere 18MP), printed 105cm on the long side, and the detail is astonishing. No resolution issues whatsoever.

 

It's all very well using a large format camera and enjoying the resolution you get - that's fine. But to say "24MP is pretty useless" is rather sweeping. Might not work for you, but for a lot of us the resolution is just fine, and balanced against dealing with large files and getting a sharp image, Leica has made a good choice setting the M and SL cameras at 24MP.

 

Now, if your normal landscape process is a tripod, ground glass screen, loop, multiple zone exposure readings and assistants, that's also fine. But some of us work to simpler procedures and get more than acceptable results. What begs the question is why would you want more MP in a compact, relatively limited camera like an M when you use a large format camera for your work?

For the first paragraph - I agree, and I think the Monochroms are in a different league of resolution for very large prints compared to the “colour” sensor M or SL cameras. It is astonishing just how much detail the M 246 and 50 APO produces (I’ve printed off the combo to 60”x40”, and the size / weight / resolution is highly, highly compelling. The “effective” resolution of a Monochrom is probably twice that of the colour M/SL though .....)

 

For your final paragraph, I guess I like printing big and the Bayer sensor needs a lot more oomph in megapixels compared to the Monochroms ....so when I see the image quality for large prints off the Monochrom, I’d love to see the equivalent in such a small package from a colour M/SL.

 

Whilst I use the 5x4 for ultimate image quality (and especially for movements, mainly front rise), maybe my desire for a very high megapixel colour Leica full frame is due to wanting to have a cake and eat it (ie, Leica glass’ rendering, high resolution - but all in a small package).

 

Given the Leica CL is 24mp on a small sensor, that technology scales to 40-50mp when extrapolated to 35mm full frame, so I’d guess we are within a year of Leica producing that via the next SL ....so my hope might not be far off, hopefully!

Edited August 5, 2018 by Jon Warwick

[adan]

I admit this is coming from someone who is currently in the French Alps and carrying a 5x4 large format camera around, with negs that I will get drum scanned to 750mb so that the image opens at 60”x50” at 300dpi.

 

So maybe I have a quite different perspective of “resolution”.... what i do think is that images off my M240 are very unsatisfying in comparison!

 

Leave MP aside for a moment. How much resolution do you think Leica M lenses have?

 

Hint 1: lenses do not have "infinite" resolution

 

Hint 2: a 24Mp "FF 35mm" format sensor (24mm x 36mm, 4000 x 6000 pixels) can capture 166.7 lines per mm, or 83 line pairs per mm (avoiding most "aliasing"). That is already "off the charts" for most MTF graphs, which generally only display down to 40 lppm.

 

Which is just to say that unless your "35mm-format" lens can consistently produce 80-100 lppm, across the whole picture, in conditions outside of a 1-tonne optical bench (compared to which any man-portable tripod is so much aluminium foil) and the aperture welded at the "best" aperture, 24Mp is already overkill. You won't get details 1 pixel wide anyway, let alone details 1 pixel wide on finer, smaller, more numerous pixels.

 

The 50mm APO-Summicron, at an eye-watering $7000, just about gets to 160 lppm. Alledgedly - I don't put much faith in "internet claims," just actual measurements. as shown in imagery. Any "lesser" lens won't, even from Leica. They fall into the 80-100 region. The modern APO/ASPH lenses spread that high-res coverage to more area, or at wider apertures, compared to older lenses that only hit that target in the image center. But don't increase the peak resolution much above that range.

 

Diffraction limit @ f/10 (want DoF for landscapes?) for 35mm-format sensors/lenses is about 2000 cycles (line pairs) across 24mm image height - which coincidentally is exactly what 4000 pixels should capture cleanly (better if it is a Monochrom, of course). Beyond that, more MP won't reveal any more detail than simply up-rezzing the image.

[farnz]

Advertisement (gone after registration)

Andy,

 

For a digital sensor, wouldn't we need to take Nyquist frequency into account so the resolution is likely to be half that at best?

 

Pete.

[jdlaing]

No.

[adan]

Andy,

 

For a digital sensor, wouldn't we need to take Nyquist frequency into account so the resolution is likely to be half that at best?

 

Pete.

 

Yes. And I did. The Nyquist frequency is half the sampling frequency. The sampling frequency of the M10 sensor is 166.666...7 samples per mm, thus its Nyquist frequency is a signal of 83.333 line pairs (one white, one black) per mm. I did mention 80-100 lpmm.

 

Although that is not a "hard" limit - you start to get increasing artifacts with a signal frequency above the Nyquist frequency, but things don't immediately fall to pieces. And it depends on the subject matter - a lens test chart, or bar code, or clothing weave, that has its own repeating texture or pattern or "frequency" completely independent of the imaging process, will show artifacts strongly, when a more organic signal (random blades of grass at different spacings and angles, the random pimples and lumps in a stucco wall) will not.

 

Which was what Leica (and some MF back makers) figured out when they chose to eliminate AA filtering - in 99% of the textures in 99% of pictures, we can "cheat" the Nyquist limit by quite a bit before getting into serious trouble. Fashion photographers being one exception (their pictures are ALL clothing weave).

Edited August 6, 2018 by adan

[pgk]

Yes - but once again, I think the number of MPs needs to be connected to print size.

 

Often on this forum, people say “24mp is enough for me”, but without any info to support why that is the case. All I can assume is many people don’t print, or simply print in small sizes.

 

If you print very large prints, especially where fine detail is needed like landcapes, then 24mp is pretty useless.

 

I admit this is coming from someone who is currently in the French Alps and carrying a 5x4 large format camera around, with negs that I will get drum scanned to 750mb so that the image opens at 60”x50” at 300dpi.

 

So maybe I have a quite different perspective of “resolution”.... what i do think is that images off my M240 are very unsatisfying in comparison!

 

A 5" x 4" camera isn't about resolution though is it? Because generally speaking its used where its diffraction limited. So it relies on 'poorer' lenses with large coverage. And lens options and technique are limiting too. Large prints from a 5" x 4" won't show the micro-detail available from 35mm lenses used at larger apertures.

 

10, 18 or 24MPixels are fine for print sizes which don't overblow their files - which is why I'm constantly baffled by people suggesting that M cameras can't be used for landscapes - even the ltm cameras were carried by landscape photographers (think mountaineers such as W A Poucher) and used to great effect. A restaurant not far from me has some of Poucher's original B&W 1940s prints up - they are sizeable but not huge and, despite being shot on Leica ltm cameras with 1930's lenses, have been enjoyed for 7 decades so far. If you need a very large print then you can use appropriate equipment, but saying that because an M can't produce a very large print from a single exposure so is not suitable for landscape photography is to apply a caveat which is quite untrue for most photographers.

 

And FWIW its not even as simple as MPixels and print size because subject matter and lighting affect print size and acceptability too. I talked to another photographer recently about an approximately 28" high vertical print from my old M8 which he liked and as he said, its not simply the technical specifications because the subject matter, lighting, lens and technique all coalesced to allow a larger than usual print from the M8 to stand up against prints of the same size from much higher MPixel cameras.

 

Cameras are what they are. We use them as our personal biases dictate and deciding if one is unsuitable for a task is made harder if we over think technical specs which we all too often do.

Edited August 6, 2018 by pgk

[jaapv]

I'm not sure that's right, Jaap.

 

The electronics are made to specification.  Leica makes the case, top and bottom plates, rangefinder and shutter.  In future models, it is not as impossible as many argue to standardise this specification, rather than changing it for every new model.  It is not so hard to require the electronic suppliers to work to the standardised Leica body specification.  it would require forethought and heat and battery management would be a challenge, but not as hard as some suggest.

 

The issue, as Andy points out, would be one of CS managing sensor and electronic upgrades and one of volume.  $20,000.00 for an upgrade might be somewhat pessimistic, but it would be more than the unit cost of the sensor, processor and related electronics.  I seem to recall that those components cost somewhere around $2,000.00 (not sure if this was guesswork).  Fitting the new upgrade should not be too much more than $1,000.00 per camera, provided the volumes are right.

 

I love the idea that we can match the longevity of the Leica hardware with the externally sourced electronics. I don't think it will happen (not because of Mr Lee) - more because of volumes and because standardising the body specification would remove some of the excitement many owners feel about a new M camera.

 

Cheers

John

It depends on the design parameters, John. Leica took one year longer to design the M10 because of the complication of fitting all the hardware into such a small body - the body shells are highly specific.

FWIIW - I like the idea of upgradabliity - I just don't think that it is economically, or even technically, feasible.

[pgk]

FWIIW - I like the idea of upgradabliity - I just don't think that it is economically, or even technically, feasible.

 

I suspect it is technically feasible - providing of course that things don't change and that technology doesn't develop in ways we haven't considered (but it almost certainly will ).

[farnz]

Yes. And I did. The Nyquist frequency is half the sampling frequency. ...

Ah thanks; I'd assumed - obviously incorrectly - that you'd halved the lp/mm to account for the effect of the Bayer Matrix.

 

Pete.

[wlaidlaw]

For what it is worth, I suspect Leica may use a custom version of Sony's 40mp full frame sensor with angled microlenses, in effect a full frame version of the CL sensor, in the next M camera. Now whether this will add anything to the equation, with the resolving power of the current range of Leica's M lenses is a totally different question. I have printed as large as 24" x  36" from M8 10mp images on an Epson 7880, where I think the quality of the Epson large format RIP engine used, seemed to have done a remarkably good job and the sharpness of the image is just fine at the distance you view an image of this size. Similarly I have printed from my SL to 44" x 66" on an Epson P10000 and again the resolution seemed fine, so the only benefit of extra pixels would seem to be for extreme cropping. A further increase in high ISO performance would probably be more useful to me than more pixels. Using my M240 recently, its low light performance was a noticeable step back from the SL and CL. 

 

My own personal view is that the manual focus and methodology of the M cameras seems more suited to the more considered process of film than digital. I now use my SL and CL for nearly all my digital photography and film M's (M3, M4, M4-P and M7) for most of the rest plus just occasionally my IIIg and Reid and Sigrist III. 

 

Wilson

[jaapv]

Now I wonder. Is the resolving power of the lens really a decisive factor? I rather doubt that we have a "weakest link" situation here.

Just consider the way old to ancient Leica lenses with low resolution shine on modern 24 MP sensor - and on modern, high-resolving film.

I think a better sensor will lead to better results with all lenses, and a better lens will lead to better results with all sensors.

 

There are studies by Zeiss - on film-  that  lens resolutions over 80 LP/mm do not result in higher image quality. Erwin Puts addresses this in the old version of his compendium.

[Ouroboros]

A 5" x 4" camera isn't about resolution though is it? Because generally speaking its used where its diffraction limited. So it relies on 'poorer' lenses with large coverage. And lens options and technique are limiting too. Large prints from a 5" x 4" won't show the micro-detail available from 35mm lenses used at larger apertures.

 

10, 18 or 24MPixels are fine for print sizes which don't overblow their files - which is why I'm constantly baffled by people suggesting that M cameras can't be used for landscapes - even the ltm cameras were carried by landscape photographers (think mountaineers such as W A Poucher) and used to great effect. A restaurant not far from me has some of Poucher's original B&W 1940s prints up - they are sizeable but not huge and, despite being shot on Leica ltm cameras with 1930's lenses, have been enjoyed for 7 decades so far. If you need a very large print then you can use appropriate equipment, but saying that because an M can't produce a very large print from a single exposure so is not suitable for landscape photography is to apply a caveat which is quite untrue for most photographers.

 

And FWIW its not even as simple as MPixels and print size because subject matter and lighting affect print size and acceptability too. I talked to another photographer recently about an approximately 28" high vertical print from my old M8 which he liked and as he said, its not simply the technical specifications because the subject matter, lighting, lens and technique all coalesced to allow a larger than usual print from the M8 to stand up against prints of the same size from much higher MPixel cameras.

 

Cameras are what they are. We use them as our personal biases dictate and deciding if one is unsuitable for a task is made harder if we over think technical specs which we all too often do.

 

Much of WA Poucher's photography was intended to illustrate his guidebooks and large print sizes were not really a consideration.  Lugging plate cameras of the day around Snowdonia would have been onerous, even John Clow restricted his cameras to mainly Hasselblad V series.  I agree with your general point, though.  You can find many examples of 'low resolution' landscape prints at reasonable sizes on the walls of any of the tea rooms, cafes, pubs and restaurants that are favoured by walkers and climbers around the National Parks.  The last thing anyone would consider is 'resolution'. 

 

I think a little context is needed here, this thread is becoming tediously bogged down with various definitions of 'resolution' and in danger of sinking under the influence of geeks.  You don't use a large format camera for any other sensible reason than personal choice, and that's good enough. 

 

It's perfectly possible to make large prints from 24mp full frame and aps-c cameras.  The printed image is not constrained by native pixel dimension at 300dpi and image quality would frequently suffer with more megapixels.

[dantemi]

Just to underline the concept that more MPs do no lead to better photos, I traded my X1D, the resolution and the sensor of are awesome, for the M10 and this move really improved my photography and my landscape attitude didn’t suffer this change at all...

[01af]

Leave megapixels aside for a moment. How much resolution do you think Leica M lenses have?

This question is irrelevant. More pixels will always improve the resolution of the final image, behind any lens.

 

Whether we actually need more pixels is another question. When the first 35-mm sensors with 24 MP came up, people dropped their jaws. They marveled at the image quality and quickly started asking questions like, who actually needs this kind of bloated resolution, and can current lenses keep up? (By the way, they asked exactly the same questions when the first 11 MP sensor was introduced in 2002.) Now, after people got used to 24 MP and we got sensors with even more MP, people start saying things like, 24 MP is useless. It's so ridiculous.

 

24 MP is enough for a high-quality 1 × 1.5 m (40 × 60 inch) print. You can step up to it at reading distance and still won't see any pixels.

 

 

Beyond that, more MP won't reveal any more detail than simply up-rezzing the image.

That's the usual mental short-circuit when people confuse lens resolution, sensor resolution, and image resolution. The latter is not equal to the minimum of the other two, but always a product of both the lens' and the sensor's resolution.

[Advertisement]

Hallo,

Look here: Leica M

[jaapv]

As I already indicated in my post, Olaf has it 100% right. This particular misconception is widespread and it is an uphill struggle to combat it - especially when marketing departments like Nikon's tout "lenses optimized for our new high-resolution sensors", implying that everybody should trade in all their grotty old ones as they will suddenly be junk. Blathering nonsense.

[pgk]

This question is irrelevant. More pixels will always improve the resolution of the final image, behind any lens.

 

But only if the final output requires it. Simply having MPixels for the sake of is pointless. Which is where these threads always go wrong. IF you must print extremely large for close viewing then everything has to be optimised for this. But few genuinely, actually do print large for close viewing .....