Guest guy_mancuso Posted November 9, 2006 Share #1 Posted November 9, 2006 Advertisement (gone after registration) As I sit here waiting for my M8 today and trying to figure some of this out. Obviously we have issues here but let's think of solutions . My opinion is right now Leica should stop production and rework the IR glass and take the 2000 customers or what ever that number is and rework our camera's through Leica repair or replace camera's and put in better IR and sell them as refurbished. But they should stop production and get a better IR filter in the box. They just have to accept the loss here. Okay but let's look at the root cause. We have a IR filter that is still letting to much IR light into the sensor. I am not a engineer but cut you not in firmware tweak the light transmission coming in or block that wavelength out in the sensor. i guess I am wondering how much can you eliminate that wavelength before it records and not through external filters. We all have seen the large color gamut from the LL site and I would imagine that is the amount of visable light but also the IR light . Can that color gamut be shortened on the IR wavelegth as not to record. Am I grasping here. I know in raw processing we can tweek some of this out but black is black and it does not know which black to attack in a image. There has to be a way or hopefully a way to eliminate the IR light that actually records through the firmware/sensor side. It maybe just a slight change in the color gamut or a slight tweek Okay engineers slap me silly on this one Okay i need Starbucks to start my day but let's work on the fix instead of the why. It's there it's a mistake now let's put our heads together and figure it out and make it workable . Some of us are not giving up and that includes me Link to post Share on other sites More sharing options...
Advertisement Posted November 9, 2006 Posted November 9, 2006 Hi Guest guy_mancuso, Take a look here Some Thoughts. I'm sure you'll find what you were looking for!
MP3 Posted November 9, 2006 Share #2 Posted November 9, 2006 Guy I agree with you with all my heart. Don't know how much they have produced so far, but your suggestion will win customers, existing and future ones. There are surely more of us who put such great faith on Leica, and we are not giving up. Enjoy your new M. Best Matthew Link to post Share on other sites More sharing options...
tom0511 Posted November 9, 2006 Share #3 Posted November 9, 2006 Even though I dont like filters I would think using the filters would be the easier "fix". Link to post Share on other sites More sharing options...
brucek Posted November 9, 2006 Share #4 Posted November 9, 2006 I just did some IR and UV tests with my M8 and I'm trying to process the photos as quickly as I can to post them. I remember hoping that the M8 would be sensitive to IR and it looks like I got my wish. I've more testing to do but the M8 may prove to be one of the most flexible digital cameras available. Bruce Link to post Share on other sites More sharing options...
jrgeoffrion Posted November 9, 2006 Share #5 Posted November 9, 2006 i guess I am wondering how much can you eliminate that wavelength before it records and not through external filters. Okay engineers slap me silly on this one In addition to making an ok living making images, I'm also a Professional Engineer by training (among other things). Unfortunately, the sensor is totally blind to color / wavelength within its designed response and the response can't be altered via software/firmware. It would be great if it was since the Bayer pattern could be totally removed and we would gain a go stop or two of sensitivity. Link to post Share on other sites More sharing options...
rosuna Posted November 9, 2006 Share #6 Posted November 9, 2006 Leica is probably using the best technology at hand. Maybe there is not a better solution for the IR filter than that provided by Kyocera. The whole design process is too long. Leica cannot stop the production and redesign the sensor filter and software. It would take months, maybe more than a year. It would be a disaster. In my opinion, Leica should consider how to reduce the incidence of the problem by means of a sotware tweak. The camera has a second light sensor (external), and this combination of two sensors could help in finding a solution. Anyway, Leica can report the problem to the future users and they will decide if this problem is important for them or not. A revised M8 camera will bring improved technologies, for sure. But this M8-2 is at least two years from now, and the development will cost millions of euros. The question is if these problems affect you or not, how severe they are. I have not a clear idea of it yet. Does the problem affects all cameras, with all lenses, in many ordinary circunstances? Can I live with these limitations? Link to post Share on other sites More sharing options...
brunom Posted November 9, 2006 Share #7 Posted November 9, 2006 Advertisement (gone after registration) Reminds me of when Mercedes introduced the A class with all its 'elk test' failures - they took it on the chin, dealt with it, and it went on to become a great success. Bruno Link to post Share on other sites More sharing options...
john_r_smith Posted November 9, 2006 Share #8 Posted November 9, 2006 Guy and everyone take a look at Michael R's test of the M8 on the LL site again. Look at the colour gamut for the M8 with that huge IR spike, like nothing else you have seen. Effectively there is no IR filtering. This is not something which can be fixed by firmware or software profiles or post-processing. Only a change to physical IR filtering will work, either with a filter on front of the lens or a redesign of the internal sensor + filter unit. John Link to post Share on other sites More sharing options...
bradhusick Posted November 9, 2006 Share #9 Posted November 9, 2006 Okay, while we are waiting for our latte drinks, here's another idea... Why not take the existing 2000 or so M8s and issue a statement that these are meant to be BLACK AND WHITE cameras only. In a few months, introduce a new camera called the M8c that also works for color (with a revised sensor). These M8s will produce stunning B&W photos and become collectors items in no time. Leica could say that the "M" in "M8" stand for "monochrome" and call the fixed camera the C8 and start a new legacy. Link to post Share on other sites More sharing options...
dritz Posted November 9, 2006 Share #10 Posted November 9, 2006 More thought on that thought, please. Mine just arrived at the dealer, and my intention is to use it purely for black and white. Is the IR problem really a plus for this? Still, the banding is a problem. Advice please! Thanks. Link to post Share on other sites More sharing options...
sean_reid Posted November 9, 2006 Share #11 Posted November 9, 2006 More thought on that thought, please. Mine just arrived at the dealer, and my intention is to use it purely for black and white. Is the IR problem really a plus for this? Still, the banding is a problem. Advice please! Thanks. You, Dean, may be in good shape. I don't think the IR sensitivity harms BW work at all. It may change some tones in some subjects but not necessarily for the worse. I think BW conversions from the M8 RAW files are gorgeous. The light streaking, I hope, will be fixed in firmware. Cheers, Sean (who also works largely in BW) Link to post Share on other sites More sharing options...
steve_l Posted November 9, 2006 Share #12 Posted November 9, 2006 Anybody know what IR sensitivity does to skies - i.e., direct, not reflected, exposure to IR? Is it like UV, seeing the IR will wash out clouds and we'll need to filter to see them, even in B&W? Link to post Share on other sites More sharing options...
brucek Posted November 9, 2006 Share #13 Posted November 9, 2006 Anybody know what IR sensitivity does to skies - i.e., direct, not reflected, exposure to IR? Is it like UV, seeing the IR will wash out clouds and we'll need to filter to see them, even in B&W? Usually the contrast is increased between the skies (which get darker) and clouds (which get lighter). I posted some samples in another thread (http://www.leica-camera-user.com/digital-forum/8991-m8-ir-uv-photos.html) that show a couple of clouds in the sky when using an IR filter. Bruce Link to post Share on other sites More sharing options...
Guest rubidium Posted November 9, 2006 Share #14 Posted November 9, 2006 As a physicist, I feel compelled to offer a few comments here. Guy's wishful thinking of desensitizing pixels to IR in firmware cannot be realized. Each pixel in the basic semiconductor sensor array is sensitive to ALL wavelengths of the visible spectrum, in addition to wavelengths in the near infrared and likely some of the ultraviolet. The sensitivity is not exactly uniform throughout this wavelength regime (the variation with wavelength is depicted in the so-called quantum efficiency distribution), but for the purpose of this discussion it can be considered uniform. What distinguishes pixels as corresponding to "blue", "green", and "red" in the familiar Bayer pattern is not an intrinsic difference among the pixels themselves (which are all identical as described above), but rather the distinction is made by the individual bandpass filters that reside above each pixel. Evidently, for reason(s) presumably known to the engineers at Kodak, the "red" bandpass filters in the Bayer pattern do not themselves provide sufficient cutoff in the near-IR to prevent IR energy from exciting the pixels beneath them. Evidently, the choice was made to instead provide IR cutoff with a filter that covered the entire sensor plane, which in the case of the M8 sensor appears to have insufficient cutoff as well. I would guess that the same is true for ultraviolet leakage through the blue Bayer filters. However, with the exception of quartz, most of the glasses used in optics strongly attenuate UV so this is not a problem. Lenses do, however, pass IR more readily, with APO lenses doing so especially well at it. Actually, the "magenta" cast that everyone is talking about is not magenta at all but actually a slight overabundance of red. It has to be, because the pixels beneath the red Bayer filters are getting more than their share of visible light due to the spillage of IR energy. True magenta would instead be a deficiency of green, and could not be perceived in a black subject since all colors are deficient by definition. Go into PowerPoint and create a large black rectangle on the screen. Double-click on it and edit the color to a custom one. Change the black you start with (red=0, green=0, blue=0) to something like (red=70, green=0, blue=0), and you will start to see the tone regime that the M8 is producing with black synthetic fabric or plastic subjects. To reduce this tone in firmware by down-weighting the "red" pixel responses would unfortunately introduce a red deficiency (or yellow cast) to "normal" subjects. Sorry Guy! BTW, I'm not an M8 owner, but rather a DMR owner, so I don't have a dog in this fight. However, as a Leica owner, I have a vested interest in the success of this camera and the company that made it, and so I'm hoping that they come up with a satisfactory solution to what I have no doubt is a sensor hardware issue. Jim Link to post Share on other sites More sharing options...
gogopix Posted November 9, 2006 Share #15 Posted November 9, 2006 Doesnt need a hardware change it is a lot simpler see the "Magenta ?Treatlike moire" thread Link to post Share on other sites More sharing options...
Guest guy_mancuso Posted November 9, 2006 Share #16 Posted November 9, 2006 As a physicist, I feel compelled to offer a few comments here. Guy's wishful thinking of desensitizing pixels to IR in firmware cannot be realized. Each pixel in the basic semiconductor sensor array is sensitive to ALL wavelengths of the visible spectrum, in addition to wavelengths in the near infrared and likely some of the ultraviolet. The sensitivity is not exactly uniform throughout this wavelength regime (the variation with wavelength is depicted in the so-called quantum efficiency distribution), but for the purpose of this discussion it can be considered uniform. What distinguishes pixels as corresponding to "blue", "green", and "red" in the familiar Bayer pattern is not an intrinsic difference among the pixels themselves (which are all identical as described above), but rather the distinction is made by the individual bandpass filters that reside above each pixel. Evidently, for reason(s) presumably known to the engineers at Kodak, the "red" bandpass filters in the Bayer pattern do not themselves provide sufficient cutoff in the near-IR to prevent IR energy from exciting the pixels beneath them. Evidently, the choice was made to instead provide IR cutoff with a filter that covered the entire sensor plane, which in the case of the M8 sensor appears to have insufficient cutoff as well. I would guess that the same is true for ultraviolet leakage through the blue Bayer filters. However, with the exception of quartz, most of the glasses used in optics strongly attenuate UV so this is not a problem. Lenses do, however, pass IR more readily, with APO lenses doing so especially well at it. Actually, the "magenta" cast that everyone is talking about is not magenta at all but actually a slight overabundance of red. It has to be, because the pixels beneath the red Bayer filters are getting more than their share of visible light due to the spillage of IR energy. True magenta would instead be a deficiency of green, and could not be perceived in a black subject since all colors are deficient by definition. Go into PowerPoint and create a large black rectangle on the screen. Double-click on it and edit the color to a custom one. Change the black you start with (red=0, green=0, blue=0) to something like (red=70, green=0, blue=0), and you will start to see the tone regime that the M8 is producing with black synthetic fabric or plastic subjects. To reduce this tone in firmware by down-weighting the "red" pixel responses would unfortunately introduce a red deficiency (or yellow cast) to "normal" subjects. Sorry Guy! BTW, I'm not an M8 owner, but rather a DMR owner, so I don't have a dog in this fight. However, as a Leica owner, I have a vested interest in the success of this camera and the company that made it, and so I'm hoping that they come up with a satisfactory solution to what I have no doubt is a sensor hardware issue. Jim Jim you bursted my bubble but love ya anyway:D Seriously thanks for jumping in here and help us understand this stuff. Not being a engineer I know when to get off the bus and ask for directions Link to post Share on other sites More sharing options...
mhoutman Posted November 9, 2006 Share #17 Posted November 9, 2006 Guy, You are receiving the bugger tomorrow right ?: what will do you with it ? Just curious Michiel Link to post Share on other sites More sharing options...
alexr Posted November 9, 2006 Share #18 Posted November 9, 2006 I have a question for JR, if he still is here, cause he sounds like he knows his stuff. Do you think an improved sensor with thicker, more powerful IR filter is actually fittable into an M8. I mean is the current filter this thin because there is no alternative or out of choice? Because if there is no alternative we are really up against it. Also,if a thicker filter can go in, how will it affect general image quality? I like the idea of the Monochromatic 8. I would never buy one but it's an elegant solution and the idea of changing to C8 and starting anew i LOVE. I am so fed up with this bloody tradition thing. It is beginning to feel like a ball and chain.I want the best of the best of NOW not of yesteryear. Finally going back to the original request by Guy I think they must stop and redisign, no question. Would it spell disaster? Maybe but it is already a disaster with a capital D with knobs on. Either go for it and redisign for perfection (IF it can be done and however long it takes) or give up, sell it as a BW digital (which wont make them any money) and then perhaps give up. I am serious. Leica are great lens makers, let them do just that. Become the (modern!) Rolls Royce version of Sygma and make superb manual and auto-focus glass to fit bodies made by others. Ok with me. I love the jewel like feel of the M bodies, but the single one and only reason i am still a Leica user is the glass. Let me put it onto something else and i will still pay for it, top dollar. Link to post Share on other sites More sharing options...
Guest stevenrk Posted November 9, 2006 Share #19 Posted November 9, 2006 As a physicist, I feel compelled to offer a few comments here. Actually, the "magenta" cast that everyone is talking about is not magenta at all but actually a slight overabundance of red. It has to be, because the pixels beneath the red Bayer filters are getting more than their share of visible light due to the spillage of IR energy. True magenta would instead be a deficiency of green, and could not be perceived in a black subject since all colors are deficient by definition. Go into PowerPoint and create a large black rectangle on the screen. Double-click on it and edit the color to a custom one. Change the black you start with (red=0, green=0, blue=0) to something like (red=70, green=0, blue=0), and you will start to see the tone regime that the M8 is producing with black synthetic fabric or plastic subjects. To reduce this tone in firmware by down-weighting the "red" pixel responses would unfortunately introduce a red deficiency (or yellow cast) to "normal" subjects. Sorry Guy! Jim Jim, that's very interesting. Does that mean that one could take a DNG and simply work the red channel on import to C1 and solve the issue -- ie., hold your real reds while removing them from where they are appearing because of the lack if an IR filter on the sensor? Link to post Share on other sites More sharing options...
Guest guy_mancuso Posted November 9, 2006 Share #20 Posted November 9, 2006 I just got the M8 check out the stop and breath thread Link to post Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.