New LEICA M vs M9 – Daylight picture RAW files comparison

[AlanG]

Advertisement (gone after registration)

FWIW they both are good and could be adjusted to get pretty much whatever look you want. I see a few more areas of moire in the M9 image. C1 lets you paint a mask and apply moire removal just to a local area. Otherwise I see that some colors are desaturated by overall application of the moire removal tool. Also, I looked at the M9 image in DXO 8 and it suppresses almost all of the moire without needing any adjustment. (The M is not supported in DXO.) However it also gets some weird checked patterns in some of the roof tile where there is moire in the M image. A number of tests and careful study would be required to decide which program is "better."

 

I used C1 and saw that each image was using a different profile. So I set them both to "Linear Response" and saw much less difference than using the profiles that applied curves. I darkened the M9 file 1/2 stop which made the exposure curves of the two linear response versions match. I clicked on a neutral and gave the two images a very similar white balance in the mid gray areas. So the rest I think represents variation of the cameras. Although it would have been better to start with an M9 image that was a half stop darker.

 

So in order here is the M image using C1's DNG Workspace Film standard, followed by the M image using DNG File neutral - Linear Response, and the last one is using the Leica M9 Generic - Linear Response.

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!

Edited March 12, 2013 by AlanG

[lct]

Thanks for those DNGs.

The light is not the same but after adjusting exposure and white balance, i don't see any significant difference in color rendtion.

Moiré is visible on both files but seems more pronounced from the M9.

As far as sharpness is concerned, files must be upsized (M9) or downsized (M240) to avoid unfair comparos. Doing this, the downsized M240 file looks slightly softer than the original M9 file, whilst the upsized M9 file looks very close to the original M240. This is true as long as i don't sharpen the files but with little sharpening on, the M240 looks clearly the winner.

Seems like the M240 accepts sharpening much better than the M9.

[Advertisement]

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

[BerndReini]

Which one is which?

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!

[rramesh]

One thing we will know for sure. It's not the MM.

[Rick]

Left is the M or it is shot slightly defocused.

[Manolo Laguillo]

Thanks for those DNGs.

...

As far as sharpness is concerned, files must be upsized (M9) or downsized (M240) to avoid unfair comparos.

...

 

I can't really understand why this must be done, please explain it.

We could compare Plus X against Tri X, or Agfapan 25 against Plus X, or Plus X 135 against Plus X 120, it was an interesting exercise.

If we want to know what the M delivers compared to the M9, it's ok to let the files be, isn't it?

The final comparison will be of course the printed images, on paper. And when doing those prints it will be ok to print the files with their original information, without up- or downsampling, IMO.

[lct]

Advertisement (gone after registration)

No problem if comparing a big image to a small one is your cup of tea but it's like comparing apples to oranges to me. YMMV. Now with or w/o up or downsizing, the results look very close anyway but the most interesting thing to me is what i suspected previously i.e. that M240 files look easier to sharpen if need be.

[skinnfell]

First of, a big thanks for making these beautiful DNGs available.

 

Then my observations, based on processing on LR4,3 (without proper M profile) and viewed on my Eizo ColorEdge. And my personal veridict in the end.

 

- M has slightly calmer highlights (visible in the sunlit building to the right) and shadows (visible in the buttresses of the cathedral). I attribute this to the slightly larger dynamic range that has been reported. I am not going to estimate amounts, because the difference is VERY slight to my eyes.

More DR is desirable, so a slight win to M here.

 

- M9 picture appears a bit better defined, even at lower resolution. This does of course result in some moire. This becomes particularly evident when some sharpening is applied.

Moire aside, I prefer crispness, so a win to M9 here.

 

- M produces a bit warmer, less magenta colors, with all color settings being the same. This is in tune with the general "look" in todays magazines and blogs. Anyhow, this is easy to correct in post.

I suppose I would prefer the M here, but only very slight.

 

M image is slightly better overall, but less detail even with larger resolution is noteworthy.

[skinnfell]

Which one is which?

 

The right one is slighly bluer and sharper, so IF the settings are otherwise all the same, then the right is M9.

[kdriceman]

I can't really understand why this must be done, please explain it.

We could compare Plus X against Tri X, or Agfapan 25 against Plus X, or Plus X 135 against Plus X 120, it was an interesting exercise.

If we want to know what the M delivers compared to the M9, it's ok to let the files be, isn't it?

The final comparison will be of course the printed images, on paper. And when doing those prints it will be ok to print the files with their original information, without up- or downsampling, IMO.

 

It is not a "must" that it be done, but if, as a reviewer, you are going to penalize or reward the images for things like sharpness, noise, resolution, etc you should at least see how they stack up when normalized for size. I compared these images both ways because I first wanted to know that any differences were not a function of size differences, then I wanted to compare based on how I would use the files in practice (I'm not going to resize all my files).

 

Ultimately, I will probably print portions of each image both reducing the size of the M file and without resizing and compare them side by side because I just want to know. I doubt I will be able to tell any difference at normal viewing distance either way after printed (possible exception may be the very slight difference in color rendering).

[AlanG]

I tried to make the images match and then compared sections before and after moire removal in C-1. Overall moire removal kills some colors in other parts of the image. Especially vibrant reds.

 

I applied a bit more sharpening to the M image. I think once you start examining for fine detail you also see areas where the image starts to break down. So it is a question of balance. The shutters show this problem and I think this affect is what some see as "micro-contrast" in landscape and other pictures where fine detail is at its resolution limit and is picking up this kind of edge effect.

 

The M9 image is a half stop brighter so that might be a factor. I think a key point is that there is no specific profile in C1 for the M. And the profile for the M9 is very blue but once balanced has a cleaner look on this image for me without doing fine tuning on the colors. (Richer yellows, bluer blues.)

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!

Edited March 12, 2013 by AlanG

[rscheffler]

I tried to make the images match and then compared sections before and after moire removal in C-1. Overall moire removal kills some colors in other parts of the image. Especially vibrant reds.

 

I applied a bit more sharpening to the M image. I think once you start examining for fine detail you also see areas where the image starts to break down. So it is a question of balance. The shutters show this problem and I thin this affect is what some see as "microcontrast" in landscape pictures where fine detail is picking up this kind of edge effect.

 

The M9 image is a half stop brighter so that might be a factor. I think a key point is that there is no specific profile in C1 for the M. And the profile for the M9 is very blue but once balanced has a cleaner look on this image for me without doing fine tuning on the colors. (Richer yellows, bluer blues.)

 

That's very interesting. I tried LR 4.3 because that's what I typically use and got much different results from the M9. Very, very interesting to see how much better C1 apparently is with the M9's files and the roof tiles. Here's what I got with LR; M is on the left, M9 on the right (please note this is at 200% view):

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!

[AlanG]

That's very interesting. I tried LR 4.3 because that's what I typically use and got much different results from the M9. Very, very interesting to see how much better C1 apparently is with the M9's files and the roof tiles.

 

I don't have LR so I can't compare. C1 is made for high end digital backs that don't have AA filters so maybe they have worked on the moire tool more than others because their MF back users really need it. As I said before, it is the kind of feature you may want to apply to a masked section otherwise it can destroy some colors.

 

Clearly at 200% you can see how parts of the image break up and get replaced by that checked pattern that in some cases may look like enhanced micro-contrast, until you examine it closely. DXO does not do a very good job on this section either - despite having a moire reduction slider. It stays checked like the sample you posted from LR. The role of the raw processors is clearly what we are down to at this point. Standardizing on LR is probably not a good idea.

Edited March 12, 2013 by AlanG

[Manolo Laguillo]

Perhaps I should clarify myself a bit more:

 

if we agree that the developer's role is essential, and if it's clear that there are yet no LR profiles for the M available, why are we insisting in developing both files in LR?

 

Because of this I used "Raw Photo Processor 64" for developing both files, and the results are convincing.

 

I include 2 screenshots where both windows can be seen side by side. The files were first developed in "Raw Photo Processor 64" (no sharpening, only K64 profile applied), and then both opened in Photoshop. The more blueish and bigger ones belong to the M.

 

Look for instance at the snow spots on the mountains between the 2 cathedral towers. And that animal sitting on top of the roof...

And look at the "Zurich" sign on top of the tall building.

 

The M has more detail. Not much more, but in any case a quantity that is clearly visible.

 

And suddenly the difference in crispness between the M9 and the M when developing in LR (where the M9 clearly wins) has disappeared, and as I said before, things are again as they should: the image with more pixels delivers more detail. :-)

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!

Edited March 12, 2013 by Manolo Laguillo

[rscheffler]

I suppose it depends on personal preference, but despite the M lacking profile support in LR 4.3, I prefer the M output to that of the M9 primarily because of the slight amount more resolution that at least with LR appears to result in better, truer definition of fine details.

 

I don't use RPP much and therefore am unfamiliar with most of it's adjustment settings. I ran the two files with pretty much default settings other than selecting K64, which is why the M file looks comparatively dark. I should note this is again at 200% and I forgot to mention in my previous post that I've sized the M file down to the M9's dimensions. As you can see, RPP also struggles with the M9's version of the roof tiles.

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!

[lovelyleica]

I made a Lightroom profile for the new Leica M, you can find it here :

http://www.l-camera-forum.com/leica-forum/2342335-post52.html

 

Try it and you will get what the Lightroom Standard Profile for Leica M should be ;-)

[haroldp]

I can't really understand why this must be done, please explain it.

We could compare Plus X against Tri X, or Agfapan 25 against Plus X, or Plus X 135 against Plus X 120, it was an interesting exercise.

If we want to know what the M delivers compared to the M9, it's ok to let the files be, isn't it?

The final comparison will be of course the printed images, on paper. And when doing those prints it will be ok to print the files with their original information, without up- or downsampling, IMO.

 

but you would print them at the same size.

 

for digital viewing, the most convenient same size comparison is by resampling

 

... H

[svenjosh]

They use CCD because they have no choice.

 

Most MFD users - including myself - would kill for a CMOS based MF system with all the inherent advantages.

 

 

[Richard.no]

I'm certainly no expert, but if there isn't any difference whatsoever between CCD and CMOS, why do all the MF manufacturers use CCDs? Why would they deliberately provide systems that can't do the higher ISOs? Why would Phase One, for example, just introduce new MF backs with the highest ISO being 800? Why wouldn't they just drop a CMOS sensor into their backs and tout ISOs north of 6400?

 

I think this blog post by Truesense Imaging says quite a lot about the (historic?) usage of CCD versus CMOS for high quality applications:

 

Who said, “You Can’t Have the Best of Both Worlds?” - Truesense Blog

 

Sorry if someone already have pointed to this one. It's not very thorough, more like a sales pitch, but anyway.

 

So, did Leica go for the "best of both worlds" part, or maybe the ROI and flexible read-out part?

[Advertisement]

Hallo,

Look here: Leica M

[Elderin]

I loaded both samples into LR, did only some WB and brightness correction to match the images. put them both into Nik pre sharpener (same settings) and saw only minor differences.

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!