carstenw wrote:
The reason the M8 has it so strong has to do with the very thin filter. A thicker filter would have caused more internal reflections due to the sharp angle of incidence with some of the lenses. It had to be this thin. I wonder if they can improve it. On the other hand, Victor said that even his P25 had shown such things, and that filters cured it.
I think that Leica should have noticed it. I think that the reviewers should have noticed it. On the other hand, I work in software, and I know how fallible such processes can be, so I am not angry, just disappointed. Leica is a good company. They will fix it. The M is their cornerstone....Show more →
There's a guy named Wisniewski over on DPR who says the effectiveness of an IR filter has nothing to do with the thickness of the glass, unless they tried to use a cheap generic filter instead of a custom design. According to him if Leica does not want to replace the sensor assembly then the only fix is to use filters on the lens.
However, the replace color is quite specific it doesn NOT reduce or replace all magenta or other color.
It will only replace the magenta specifically where it is superimposed on a neutral and low area, that is black or dark grey (read up n it) since this is a low light area values are low and the red level is on top of an x,x,x RGB
Note the example that almost all the other red areas are untouched. This means ANY green or blue above neutral will keep the item untouched
You can dismiss if you want butit is quite specific and is due to red leaking not any other spectral values
try it before you pooh pooh it
In any case, my only point is that an action or post process fix is likely, and is FAR less destructive than noise reduction, and likely to make fewer value changes than are within bounds of most images (cool to warm eg)
try it, you will like it, and for sloppy people, yes, it could literaly 'clean the image'
my graduate work was in physics nd while primarily plasma physics it was inclusive of optics etc.
hot spots are not caused by the ir filter, rather the lens not being designed to operate in the IR spectrum and are actually internal reflections within the lens. this is inclusive of the entire optical assembly, glass formulae and coatings.
glass, in and of itself, does block IR to a degree. but not the only component to IR blocking, obviously.
the thickness of a filter and its impact on IQ are not entirely applicable in a potential fix that could be as simple (actually not) a designed and appropriate deposition coating to the existing cover filter. look at the impact of coatings, single and multi on lens performance.
based on my experience with e-beam vacuum deposition, it is likely that a coating fix can be developed with no impact on the remainder of the system, no change in filter thickness or waveform interference (the blur in AA that MUST destroy absolute resolution to work at all, particularly at high frequency, nyquist, which is approached al all hard edges)
myself, i'm still in the game waiting for my M8 but believe a fix is absolutley feasible with no deleterious effects, but will likely require a return to leica for implementation.
Carsten
if you do weddings with tux in some lightings, then you will get. But not i most shots at all
Also, the solution works with selecting only ONE area in the image, since all black PLUS IR leak will be roughly same TONE withing a narrow band (use fuzziness)
at fuzziness=130 most images fix with one eye drop and just zero saturation. That is it.
Guy,
save your money. I do not think the IR filters are needed and do not work as well as the post fix.
What filters did you get Guy? I looked at the B+W 486, but the only size I could find that I need is 46 (or was it 49mm). I have lenses with 36, 39, 43, 46, and 49mm filter threads. The 489 is available in more smaller sizes but is cyan tinted since it attenuates visible red. It seems unsuitable for photography and more designed for consumer video where WB like much else is auto-everything. That's probably why it's found in small sizes. I could use 46 and 52mm filters with step-up rings, but then the hoods won't fit.
Personally, I'm not going to buy the filters. If Leica won't fix the IR block filter I don't want the camera.
Marco wrote:
My M8, though maybe not so badly, does the same thing with blacks in artificial light.
Weird and not very nice...
Although I don't know if I will send back my camera...
I'm sure at Leica they are working hard to find a solution.
And some shots I made in outdoor light DID impress me... detail and dynamic range are amazing.
Marco, how did you do to get the M8 in Italy ? and where ?
I've been told that delivery in Italy will start on the 29th of november and I will not get my before the 15th of December !
However, after reading all the messages about the IRfilter/magenta cast issue I'm inclined to not buying it anymore. I know I might be wrong but really cannot accept it from a 4.200€ camera.
Ciao
Massimo
Ma come diavolo hai fatto ad averla cosi in anticipo ?
The dynamic range is very good and there's little noise in the shadows.
Compared to my 1Ds the M8 gives much cleaner shadows, which can be recovered quite well.
This is the same DNG after some shadow recovering (double conversion, DRI action and slight highlight/shadow adjustment). The shadows was overall increased by over 2 stops.
Compared to the R-D1 details are amazing.
It's like shooting with a MF camera compared to a 35mm one.
I shot the same scene with the R-D1 + 21 asph and the M8 + 24 asph, same effective field of view, both at f/8 to minimize dof difference.
I converted both raw in CaptureOne, but the R-D1 got a slight sharpening, while the M8 none.
Then in PS I upsampled the R-D1 image to the same pixel width of the M8, to have an idea of differences in prints of the same size.
I made a gif animation of a 100% crop from the resulting files.
Guess which is the M8 ?
Forget the colors, which btw are more accurate in the M8 image...
Seeing this great detail and dynamic range just makes me even more distraught over the IR filter issue! By the way, is it me or are some darker areas in those bushes a bit magenta?
Guys, I was going to stay out of this discussion until I actually get my camera next week and time to test it, but I really don't think this "magenta issue" is that big a deal. Light is very different in different parts of the world, so I alway expect images not to be perfect without adjustment right from any new camera I buy. Shoot a grey card, and neutralize, or profile your RAW converter to the actual images your camera produces, and the magenta cast should vanish like it never existed. We run into this kind of thing all the time in MF digital backs. You just need to neutralize the settings for the sensor, then everything falls into place.
I will post more though after I get my camera, and can verify this procedure.
In case you're not a member of that site and therefore can't see the images, what they show is a shot of two cameras resting on a black cloth. Everything in the scene shot with the M8 renders as it should, including other areas of black (e.g., the leather on the camera bodies), but the black cloth renders as magenta. So, the scene is "neutralized", but that still doesn't get rid of the magenta cast on the black cloth. Moreover, a WhiteBal card was included in the scene and used to neutralize it. The blacks on the cameras pictured are indeed black, and gray is indeed gray, but the black cloth -- and only the black cloth -- is magenta.
I should note that I got my M8 last week and have not yet been able to reproduce the magenta issue. On a thread on LL, someone suggested that it only appears when shooting with strobes or tungsten hotlights, which I don't plan to do with the M8. OTOH, there are many threads floating around of outdoor shots under natural light with the magenta cast (eg., a black raincoat turning purple), so I don't think the problem is limited ot studio/artificial light settings. EDIT: here's an example, photographer says no flash was used: http://homepage.mac.com/billh96007/PhotoAlbum196.html
So, I don't think it's a simple white balance issue, and I don't think it's an artificial light issue. I think it is what Leica admits it is: the increased IR sensitivity means that certain artificial fabrics with show as purple, regardless of whether the scene is properly white balanced, and regardless of whether it was shot with natural or artificial light. Of course, bad WB and artificial light may increase the effect, but they're not responsible for it.
I'm 99% certain that, with great reluctance, I'll be returning my M8. The only thing that's stopping me is that I personally haven't been able to reproduce the problem, but I assume it's just a matter of time. I seriously doubt that I have the one magic M8 that doens't have this problem.
Someone please correct me if my analysis is incorrect, because I'd love to keep the M8.
