Guy, was your flower shot taken with an IR filter or without? I always felt that the DMR shots everyone was showing had a slightly deformed gamut, but it is a hard thing to judge if you weren't there when the shot was being taken. Greens always seemed a bit yellowy, reds a bit browny, and skies were too green. In your M8 flower picture the greens in the leaves seem way off. My girlfriend described them as 'sludgy'. This could be consistent with an overall yellow cast, but it seems to affect different colours to different degrees. Difficult and time-consuming to correct. I remember Kodak film used to render slightly yellowish greens, at least compared to Fuji.
Interestingly, Irwin Puts' naughty M8 sneak preview showed dramatic colour cast variation at different densities of grey in the same image. The 5D sample beside it showed quite neutral greys.
It seems to me that colour accuracy is an area where Leica/Kodak have a _lot_ of catching up to do, and I don't believe that it was adequately tested and appraised in the DMR bible thread.
talked to a Leica guy today at a presentation and he said his info from this morning was that besides the filters Leica will correct a part of the problem with a firmware upgrade available in december. He said that he didnt understand himself how but that the behaveour would improve and filter would be needed only for few situations.
He did not think that Leica will do any hardware change.
I have experienced similar IR issues in the past starting with one of the first pro 6 megapixel cameras in 1995 which I paid almost $30,000 for. That camera produced an unusable photo with an obvious red cast and guess what the solution from the manufacturer was, use hot mirror filters. These first cameras had many other imaging issues as SLR digital was in its infancy. A few years later I bought a 14 megapixel camera and the IR problem with this was a red spot that would appear in the center of photos under certain lighting conditions and with some lenses and not others. Even the normal 50mm standard lens created this problem. The solution from the manufacturer was replace the image sensor for $1500 at the customers expense. It is hard to understand how a manufacturer can produce a product and not test it properly in real world situations before selling it to photographers. We have become the research and quality testing group for digital camera manufacturers and are paying premium prices for high end digital products at the same time. If this is acceptable to photographers in order to get products faster and fix problems as they arise there is no complaint. On the other hand if we deserve a product that has been fully tested and researched for this price than we should not encourage manufacturers to work with this manufacturing model. This comment is not directed at any one manufacturer, just a general industry observation. Manufacturers that may have been the major creator of SLR cameras are now dependend on partnerships for sensors and components which they no longer have total control over and a buying public which demands better technology now.
I love the shot of the barn, Kurt. Lovely rich colours. To me though, the blue sky is distractingly greener than I would expect it to be. The small areas of living grass look khaki. I would say that the many DMR shots I have seen have a similar romantic palette, which often is not an unpleasant effect. However, I would prefer to apply colour effects myself rather than start with them. I don't think it is a general cast, since the clouds do not appear to be shifted towards yellow. Although I am experienced with hue-saturation-brightness controls which allow shifting of ranges in the colour wheel, I would not want to have to apply such changes to every picture. In short, I think a camera should aim for realistic colour.
In the shot of the band there is quite ugly chroma noise on the wall between the 2 guitarists. If it is distracting at this reduced resolution and 640 iso, then I would say that this picture would have better been taken on a 5D. Presumably a quiet shutter wasn't essential :)
I have a 1920x1200 screen and well calibrated, and I just don't see the noise. Als with the green and blue (etc??) lighting you are going to get specular reflection. Thisis not an image I would use to judge chroma noise (acoustic, maybe )
Also, look like frost bitten vegetation we have here as well, kahki-green (touch of green-black
But Kurt, can you say? Is this pretty close to the scene. Maybe show a M8 -DMR comparison.
on other subject, from what I hear, the filters may have a potentially POSITIVE effect on definition.
In fact, and not intending to cause offense, I think the band picture demonstrates perfectly the lack of versatility of the M8 in the following ways:
1) It is often useful to freeze motion with this kind of shot, and the ability to do this depends on the light levels, iso, and aperture. A Canon 5D at 1600 iso with the excellent Canon 85 f1.2 or a cheap Zeiss 85 f1.4, or even a Leica R f1.4 would far better accomplish this.
2) matching the very narrow focal plane of a wide aperture lens with the subject is hard. You have failed to do that in this case. The best way to achieve crispness in this kind of shot, in my experience, is to constantly seek focus with the manual focus ring while holding down the shutter release. Digital makes this method affordable, and statistically the shorter the frame interval, the more shots you will get where crispness and a good moment collide. The M8's slower frame rate means that you are significantly less likely to get that crisp magic moment.
3) neither of your shots displays, to my eye, excellence in composition, in that they almost look as though you couldn't precisely see where the border of the frame was. When seeking the best image quality, pixels aren't a luxury we can discard so casually.
I am sorry to have to cast a critical eye, but I do think it is important that none of us brushes the M8's shortcomings under the carpet. $5000 is a lot of money to spend and we must all think carefully about what we get for it.
you really need to sort out your monitor. I am using a bog standard old G4 Apple powerbook and the streaks and noise are glaringly clear. If I didn't know you were for real, and respect your views and contributions to this great forum, I would worry that you were a Leica astroturfer! I am seeing green/red blotches on the wall between the two guitars. Many of the blotches are as big as the guitar knobs. They spoil the graduation of tone from the highlight upwards to the darkness. I would estimate that it is about the same level of noise as I would get on a 5D at 3200 iso. Am I going mad or do others see it?
That is y point, by filtering OUT the IR and far visible red, the restof the image that resultswill apear sharper.
Victor, OK with me. I think we are using different terms to describe the same thing. The longer wave lengths are focusing at a different point to make the image fuzzy. An APO is not going to be an APO anymore!
I never really thought about it before, but it goes both ways. When you are intentionally taking an infrared photo - if your filter allows too much visible light through, it will result in a less sharp image. I have to pay some more atention to the cut off points of my infrared filters.
brainiac wrote:
In the shot of the band there is quite ugly chroma noise on the wall between the 2 guitarists. If it is distracting at this reduced resolution and 640 iso, then I would say that this picture would have better been taken on a 5D. Presumably a quiet shutter wasn't essential
I have a well calibrated montor and I do not find that level of noise in that size photo to be ugly or distracting. Perhaps you are just much more sensitive to that problem that others? If the photo were enlarged significantly, the noise would probably be a problem.
It is a matter of taste and if the photographer is happy with the image and it suits the purpose, all is well.
If I were using the image to decide whether to buy the camera or not, I would want to make sure it suited MY needs, which might be different than that of others.
Kurt, have you done any comparison shooting with the 75 Lux and Cron? This is one of the crucial decisions I still have to make. The 80 Lux is probably my favorite R lens at this point, but I am still really torn on this one.
Well, I confess I may not look that close, but my monitor otherwise seems pretty accurate (and seems accuracy on the red line isnt the point.)
and on the noise, you are correct - the focus is a bit in front of the boys. Means that the back wall is OOF. In general this will reduce noise.
I am really curuios what you see. Can you hypersnap, rez up and post? your certainly not talking about the stuff near the spot on the wall which is clearly reflected color bands from in front.
My optometrist didnt call you by any chance, did he?
But I do have a rule in my office - No nose prints on the screen
brainiac wrote:
3) neither of your shots displays, to my eye, excellence in composition, in that they almost look as though you couldn't precisely see where the border of the frame was. When seeking the best image quality, pixels aren't a luxury we can discard so casually..
uuum...can we hold on here for a moment? The technical aspects of the camera are of course fair game. But I'd like to suggest we draw the line at composition. Yes, a rangefinder won't be as exact as a SLR, but that isn't a sign of deficiency in the M8.
The Leica M8 has a real problem with infrared. Read for details.
I agree...I'd be pissed, but that's just me, can't speak for anyone else.
Enclosed is a link to DPR Leica talk...I found this very honest and straight forward...The link mostly covers reviews...they make sense. http://forums.dpreview.com/forums/read.asp?forum=1038&message=20832562
Yes, yes, Brainiac, the various color problems you describe are there - although fra,kly the green sky in the first image is more disturbing than the noise in the second.
Edmund
brainiac wrote:
Gogopix, like Slade said, "cum on - [you have to] feel the noize" ;-)
you really need to sort out your monitor. I am using a bog standard old G4 Apple powerbook and the streaks and noise are glaringly clear. If I didn't know you were for real, and respect your views and contributions to this great forum, I would worry that you were a Leica astroturfer! I am seeing green/red blotches on the wall between the two guitars. Many of the blotches are as big as the guitar knobs. They spoil the graduation of tone from the highlight upwards to the darkness. I would estimate that it is about the same level of noise as I would get on a 5D at 3200 iso. Am I going mad or do others see it?...Show more →
uuum...can we hold on here for a moment? The technical aspects of the camera are of course fair game. But I'd like to suggest we draw the line at composition. Yes, a rangefinder won't be as exact as a SLR, but that isn't a sign of deficiency in the M8.
I just want to say, thanks kidigital for posting in the first place, since the performance of this camera is of interest to me, and I am sure, lots of other pros. It's good of you to make the effort and share your initial tests here.
I don't mean to offend when talking critically about your pictures. I do think the composition of the band picture might have been better if you had been able to focus on your subject, the central figure, while placing him further to the right. That's why I think it is an example of the _lack_ of versatility of the rangefinder, since you can _only_ focus accurately on the centre of the frame. Even if you focus centrally and then re-compose, you are unlikely to get correct focus, especially with moving figures at f1.4, since the plane of focus is, well, a plane: rotating the camera makes your focal setting wrong. This is a criticism of a technical aspect of the camera in comparison to its rivals the 5D or DMR, which allow accurate focussing of any part of the image. I don't think it's irrelevant to point out what appears to be an example of the focus system hurting composition.
As regards my noise complaint, to my eye, the most noticeable instance is exactly half way between the two guitars. If you can't see it, fine. I don't agree that noise is least intrusive on flat or graduated areas, I think that is where it catches the eye most easily. The question of whether this level of noise is acceptable must be informed by the availability of a camera with less than half the noise problem at less than half the price. All analysis of camera performance is relative. If the 5D didn't exist, I might buy an M8. But looking at the shortcomings visible in these shots, I won't, and I think that these shortcomings are worth discussing so we all know what to expect.
Of course it's possible to hue shift the sky and the greens, and switch into lab colour to apply median to the colour channels so reducing chroma noise. But for that price?
Brainiac, you are making some strong assumptions here, and I don't think they hold up.
To address your points 1 and 2, the unsharpness is almost certainly not due to focus in the wrong plane, but due to a slow-ish shutterspeed and hand-holding. If you look closely, the guitarist and his guitar are as close to sharp as anything in the image. Finally, both the 50 Noctilux and 75 Lux are slightly soft wide-open. In fact, if the Noct was used at f/1.0, then it would give the same shutterspeed (okay 1/3-stop less) as the 5D with a f/1.4 lens.
Also, the Leica rangefinder system is more accurate than an SLR at these kinds of distances, not less. Focus-recompose at those distances does not invoke any significant error, and anyway, I don't think a recompose was necessary in this picture.
And to address your point 3, you criticised the cropping as having to do with the rangefinder system. Again, I find it far more likely that the lens was simply a little long for the shot. I don't know why you are so certain that the rangefinder is to blame here. Perhaps it would have been better to single out the two guys on the right, or the two guys on the left, but that is an operator choice, not a rangefinder-induced problem.
Finally, with my 5D I found that the focusing on the off-center points was error-prone and difficult to verify, so (before I switched to manual lenses), I was strictly using the center one, which was faster, more accurate and worked better in dim situations.
I agree that the rangefinder has certain strengths and certain weaknesses. I just don't think we are seeing any weaknesses displayed in this particular picture.
)