Erwin Puts' M9 Part 3 review

[patrick parker]

Advertisement (gone after registration)

M9, part3

[georg]

The girls face with the M9 is not sharp, the uprezzed M8-shot lost all detail, the last M8-picture is entirely worthless...

Why is it so difficult to take some reasonable comparison-shots in the studio for noise and resolution evalution? Mr. Puts is on par with Mr. Rockwell...

[microview]

M9, part3

 

Hmm. But what an ugly thing to work with: the +£200 or so Nikon:(.

But those m9 colours and the moiré?

[kalina]

I love my Nikons, but I feel that Puts' M9 pictures (the Asian girl and the combs) look better than the D3X pictures because of the color. I'm glad the M8 makes the same choices in color. I'm sure the D3X can be programmed to look like the M9 pictures if desired.

[Dan States]

What a mess. The M8 is clearly not focussed properly, the M9 is not white balanced properly and his D3x image of the combs shows jpeg compression artifacts. The man writes a lot of words but is just a lousy photographer. A totally useless comparison. If the M8 made images as bad as the last one of the combs they would have sold ZERO.

 

At this point his "reviews" are pretty much useless when compared to what Dpreview puts out.

[arthury]

...but the absence of the low pass filter does not bring any truly competitive advantages. The M9 pictures are not better than the comparable Nikon pictures, but presumably at this stage of the evolution one could not ask for more.

 

 

I think it is pretty safe to say that the M9 is probably the best digital rangefinder in the world and it is probably one of the best compact FF large sensor cameras in the world.

But, claiming that it is the best camera in the world with no qualifications is ludicrous and being disillusioned. And, I agree with Erwin.

[diogenis]

Advertisement (gone after registration)

This is no review, it's a mess.

I prefer 100x Sean to review these cameras, even though it's just silly, there is no comparison: it's like comparing eggs with apples they are both edible but that is all. And these cameras are also very different, but indeed eventually they both take pics.

Since when Chinese girls get the skin complexity of a European? Also you guys see the supposedly correct m8 girl pix the same as Nikon's? And the combs photo shows more natural on the m9, but who can tell? based on what?

[wookchang]

Mr. Puts failed me miserably this time.

It's a painful experience to examine sorry pictures he presumably took.

Maybe he should rewrite Part III with fresh pictures.

[pascal_meheut]

But, claiming that it is the best camera in the world with no qualifications is ludicrous and being disillusioned. And, I agree with Erwin.

 

I saw nobody claiming such a thing. They object to the test methodology, meaning the conclusion is irrelevant. It could be true as well as wrong bu† we cannot know because it is tainted by Mr Puts errors.

[diogenis]

I am not complaining about his methodology, but he presents no reference pics to compare.

Also, m9 is just under a month old, and I would expect someone to come up with real world calibrated profiles to process its Raws. Then it is also a bit more study from Leica to introduce better firmware much like how they did on m8.

But from what I see here, the m9 can easily be compared with the best of the best 24x36 sensor cameras the world have to offer. And when you stretch its lenses to limits then I guess it should be much better.

But we need time for this

[stunsworth]

If the M8 made images as bad as the last one of the combs they would have sold ZERO.

 

I have a lot of time for Erwin - though I think he does lenses a lot better than he does digital. However that M8 image of the combs is plainly out of focus. I have no idea why he decided to include it in the sequence.

[sandymc]

Erwin's confused me on this one - I've looked at the results several times today, and every time I look, I get more confused!

 

The graphs of the cropped M9 seem to show the M9 as having lower effective resolution than M8, the M9 sitting at 50% at around 700 LP/PH, while the M8 is 800. On the edges of the crop, M9 400 to 600, M8 500-700, depending on segment. Given the pixel pitch is exactly the same, and the crop means the images are exactly the same size, I'd have expected closer results.

 

I also don't understand how the D3x is doing that much better than the M9 in the center. Edges, yes, we know edges are a problem for digital M's, but the center? And the D3x curves look heavily sharpened....or are we just seeing differences in lenses, rather than cameras? Or have the JPEGs for Imatest been processed differently? If that's a Zeiss lens on the D3x, I doubt that there would be any significant image enhancement happening to the raw data in-camera.

 

 

Sandy

[Michael Hiles]

Erwin Puts’ reviews are of almost no value to me. I understand the graphs and I don’t care. If I buy an M9 (not likely), it will be because it is a Leica M with pretty much the same sense and feel as all the M cameras. It will take all my current lenses without a 1.33 factor. And it will produce more than satisfactory results – sharp pictures that utilize the basic qualities of all (almost) the lenses Leitz ever made. That is what will be useful to me.

 

I am trying to take good pictures, and in the current context, the technical stuff at the margins doesn’t matter. The quality of my pictures does not depend on anything Mr. Puts deals with in his reviews. And if anyone wants to judge my pictures based on their relation to an MTF graph, I know the value of their opinion.

[tgm]

I am also puzzled by Erwin Puts results and also by his conclusion that the omission of the AA filter has no positive effect at all.

 

1. The d3x MFT curves show what Erwin calls a digital bump, even without sharpening. It indicates that the software either in the camera or in the capture one profile of the d3x tries to recover some details which are lost as a result of the AA filter.

 

2. At the resolution limit there must be always Moire, there is an interesting comparison of different raw converters for the M8 in the dpreview review of the M8. They claim a resolution limit of 2400-2600 lines (!) for the 10 Mpixel chip of the M8, which is not so different of the d3x (2700 Lines) using the same method. For the M9 we should expect 1.3 times more lines, something like about 3300, or in other words more details than with the d3x. The M8 review also nicely shows that there is Moire at the resolution limit. If the frequeny of the resolution chart matches the pixel spacing there must be Moire, unless you smooth either with AA or with a software. But with a AA filter you always loose, you have no choice.

 

3. The two recent reviews at lens.tip and on luminous landscape both came to the conclusion that the M9 images show slightly more details than that of the d3x, at least at low ISO.

How does that fit with Erwin Puts results?

 

All in all, the whole discussion is somewhat academic, in most prints you won't see the differences. But it is also a question of concept, whether omitting the AA filter has an advantage.

 

all the best

 

Thomas

[luigi bertolotti]

Sorry, but this is one of the less appreciable tests from Puts: no disputation on graphs, but the 3 pictures of the woman are an almost worthless comparision : the white balancing of M9 and Nikon are clearly so different that speaking of differences in highlights rendition is unuseful... and the M8 image is simply out of focus, I think.

[Marty]

Love his optics reviews, but why does he always make it seem as if film-to-digital has just happened yesterday, and we all need time and guidance to adjust to this new world devoid of grain, emulsions & chemical developer.

...the current attitude of pixel peeking is detrimental...

but pp is exactly why M9 users would want to see a good technical review. If I want to read about the pleasures I go to Steve Huff.

[ljclark]

Well.

 

I'm certainly relieved.

 

Fairly soon after the link was posted I went to the Erwin's site and skimmed through the new section of the M9 review. When I came to the photos I had to slow down a bit. I went back up into the text and tried to figure out what he was trying to show and couldn't find any coherent explanation. (The charts, btw, don't concern me too much.)

 

So I decided to do some stuff around the house, run some errands, etc.

 

Upon return and reading some of you postings I realize I wasn't the only one a bit mystified.

 

Moire: So if I rack an M9 back and forth shooting an open weave textile, I should be able to hit the Moire sweet spot. Then I just shoot a D3x at the same distance and my point is proven. Have I got that right? Oh...And I also include an out-of-focus M8 shot just to make the test complete.

 

And those color pictures of a picture: What is that all about? I wouldn't mind seeing some static shots of colored 3D objects so we have some sense of control here.

 

But I'm glad to see I'm not the only one who is a bit puzzled by this part of the review.

 

(Even his critics are civil and restrained -- so far. I hate to think of the mob's reaction had one of the mere mortals here posted such a dodgy set of images.)

[lct]

According to Puts, "The pictures are processed by Capture One with very moderate sharpness parameters" so that "the comparison with the D3X images shows that even with a low pass filter and a thick system of cover glass in front of the sensor Nikon engineers can deliver first class imagery at least comparable with the M9 performance".

Wonder if this guy is serious really. Applying "very moderate sharpness parameters" (which ones BTW?) to an M8 and, i guess, M9 file hardly changes anything in the apparent sharpness of the pic, due to the absence of AA filter.

On the contrary, applying the same sharpening to a D3 and, i guess, D3x file does change the apparent sharpness due to the AA filter.

No wonder how final results can look very close this way.

What he should have said is that with a good raw converter like C1 the advantage given by the absence of AA filter tends to disappear compared to other systems.

But we all know this already don't we.

[hsittel]

...I also don't understand how the D3x is doing that much better than the M9 in the center. Edges, yes, we know edges are a problem for digital M's, but the center? And the D3x curves look heavily sharpened....or are we just seeing differences in lenses, rather than cameras? Or have the JPEGs for Imatest been processed differently? If that's a Zeiss lens on the D3x, I doubt that there would be any significant image enhancement happening to the raw data in-camera.

 

 

Sandy

Hi Sandy,

the D3x uses a (Sony?) CMOS 24.5 MP sensor, which results in 6 µm pixels - instead of 6.8 µm of the Kodak sensor used in the M8/9. This gives a 12% better sensor resolution over the Leica chips.

[Advertisement]

Hallo,

Look here: Leica M

[pascal_meheut]

Well.

(Even his critics are civil and restrained -- so far. I hate to think of the mob's reaction had one of the mere mortals here posted such a dodgy set of images.)

 

At first, the critics would have been civil and restrained. But the "tester" itself or some other people would have respond that the critics were irrelevant, somehow implying that they were coming from Leica fanatics, die-hard conservative, geeks obsessed with equipment but not pictures or any other easy way to avoid a real point and a dialogue.