rosuna Posted January 3, 2007 Share #21  Posted January 3, 2007 Advertisement (gone after registration) When originally announced, the M8 when used with coded lenses was to calculate the taking aperture within +/-1 stop and place that information in the EXIF data. Is this still a planned feature, or has its introduction been suspended?  That is a very good question. Link to post Share on other sites More sharing options...
Advertisement Posted January 3, 2007 Posted January 3, 2007 Hi rosuna, Take a look here New Band is Firmware Bug. I'm sure you'll find what you were looking for!
Guest guy_mancuso Posted January 3, 2007 Share #22 Â Posted January 3, 2007 Very good question, just not sure it can be done. I'll ask Link to post Share on other sites More sharing options...
sean_reid Posted January 3, 2007 Author Share #23  Posted January 3, 2007 Hi Sean, This thread could easily turn into a laundry list of M8 firmware bugs. ;-)  I've had a couple of quirky, but relatively minor things happen: When switching between JPEG and DNG, the frame counter won't reset to the proper number of remaining frames until the camera has been switched off for a few minutes.  After I entered and saved user profiles, and then selected one (for outdoor shooting), the viewfinder display indicated a very long exposure although I was shooting in bright daylight. I went ahead and shot, and the exposure was fine. This display problem persisted until I turned off the camera and then back on.  Certainly not big issues, but worth investigating when debugging the code. Despite this, the M8 is still tops in my book. Thanks for passing this information along to Leica.  Larry  Hi Larry,  What version firmware do you have?  S Link to post Share on other sites More sharing options...
larry Posted January 3, 2007 Share #24  Posted January 3, 2007 Hi Larry, What version firmware do you have?  S  Sean,  It's version 1.09. I just picked up the camera last Wednesday.  Larry Link to post Share on other sites More sharing options...
FastFashnReloaded Posted January 3, 2007 Share #25 Â Posted January 3, 2007 I agree the DMR situation is very poor, hardly inspires confidence in the future of the R. Â Please consider that Leica is a small company and I have no doubt that all available resources have gone to fixing M8 bugs. Â The DMR works well from what I have seen, so you can live with a delay. The M8 needs the fixes now. Link to post Share on other sites More sharing options...
marknorton Posted January 3, 2007 Share #26 Â Posted January 3, 2007 Dana, I can certainly live with a delay forever because I do not have a DMR. I was commenting that the release of a bad firmware version which is immediately retracted and still not fixed after - what? - 6 months is hardly a good example of Leica's support for its customers. Â The delay of the fixed version of the DMR firmware is more likely down to a failed commercial relationship between Leica and Imacon, the acquisition of Imacon by Hasselblad, the competitive jitters which the announcement of the acquisition of Sinar by Leica will have caused (now cancelled of course) and the deployment of Imacon people who might otherwise have been fixing the DMR firmware on to higher priority Hasselblad projects. Â It's a classic error to rely on a third party for the provision of core technology and Leica are learning the lesson. For now, their relationship with Jenoptik seems to be holding up to get the M8 done but Leica really need to start building their own electronics and software expertise. Link to post Share on other sites More sharing options...
robsteve Posted January 3, 2007 Share #27 Â Posted January 3, 2007 Advertisement (gone after registration) Mark: Â They never did release 1.3 or retract 1.2. Link to post Share on other sites More sharing options...
marknorton Posted January 3, 2007 Share #28 Â Posted January 3, 2007 Robert, you know better than I but my understanding was that people tried 1.2, found a bug in it and went back to 1.1 and have been waiting since then for 1.3. No? Is anyone using 1.2? Link to post Share on other sites More sharing options...
robsteve Posted January 3, 2007 Share #29 Â Posted January 3, 2007 Robert, you know better than I but my understanding was that people tried 1.2, found a bug in it and went back to 1.1 and have been waiting since then for 1.3. No? Is anyone using 1.2? Â The DMR are shipping with 1.2. The main problems with 1.2 was the clipping indicator was set too low and the preview display is too warm and maybe of an odd gamma too. With 1.1, the clipping indicator was very accurate, as was the colour on the preview display. With 1.1 what you saw on the preview was pretty accurate to what you will get from the RAW file. Â I go back to 1.2 if I am going to shoot sports, as the camera seems to write quicker with 1.2. Link to post Share on other sites More sharing options...
sean_reid Posted January 4, 2007 Author Share #30 Â Posted January 4, 2007 Hi All, Â Still gathering info. for the next M8 article. Here's a clarification with the latest on the banding issues: Â 1. The vertical band we've been discussing (appears just to right of center) is indeed being fixed via firmware 1.10. So that's one less thing to be concerned with (I hope <G>) Â 2. The strange horizontal rectangle band (often green, discussed by Furrukh and Rex, et. al.) that only appears when the camera is at exactly a certain angle to a light source, etc. has been replicated by Leica and they are working on a firmware solution. Apparently this phenomena is related to what they called a "reference pixel" and I won't pretend to know yet what that is or how it works. I'm sure there may be many theories put forward on the forum but I'll await further information from Leica on that. However it might work, they're trying to banish the problem in firmware. I know that it exists but I've yet to encounter it in my own work with the M8s I have here. Â Sorry that my first post suggested the inverse, Leica and I needed to clarify which banding we were talking about. Â Cheers, Â Sean Link to post Share on other sites More sharing options...
khanosu Posted January 4, 2007 Share #31  Posted January 4, 2007 Thanks for the update, Sean!  Reference pixels are the peripheral/edge pixels on the sensor. Our very own Rex indeed theorized that exactly those pixels were the culprit. After that theory was proposed we all agreed that it made loads of sense and since then no other counter theory was put forward  Furrukh Link to post Share on other sites More sharing options...
marknorton Posted January 4, 2007 Share #32 Â Posted January 4, 2007 Reference pixels are pxiels around the periphery of the sensor which are shielded from the light and are there to provide a black reference for the electronics. Without them, the charge coming off the CCD cannot be related to an absolute black. Â From what Sean says, it sounds like, somehow, these pixels are being illuminated when they shouldn't be and messing up the black level. Almost like the pixels are only black up to a certain angle of incidence or else they are being illuminated by internal reflections. Â Who knows, lots of theories will abound. I'm a bit worried. Link to post Share on other sites More sharing options...
scott kirkpatrick Posted January 4, 2007 Share #33 Â Posted January 4, 2007 Have you noticed that the background noise with structure, the runaway streak from a reference pixel, and AWB without an external sensor are all in the same part of the firmware? The area that is now (I hope) finally getting beaten into shape is the blackpoint, whitepoint, and color cast determination that occurs just before the raw file is output, and before jpeg encoding gets underway. I think the vertical bar in the background noise can be removed at that stage, too. Or was the result of some silly software glitch occurring at that stage. Â I wouldn't point too much blame on this one. Except for the white balance inaccuracy (and worse, erratic variations), this is stuff that is impossible to find without extensive beta testing, which we have been doing. Â scott Link to post Share on other sites More sharing options...
sean_reid Posted January 4, 2007 Author Share #34  Posted January 4, 2007 Thanks for the update, Sean! Reference pixels are the peripheral/edge pixels on the sensor. Our very own Rex indeed theorized that exactly those pixels were the culprit. After that theory was proposed we all agreed that it made loads of sense and since then no other counter theory was put forward  Furrukh  Thank you Furrukh. One of my operating priniciples is to try to give help in those areas where I do have knowledge and not to pretend knowledge in those areas where I wouldn't know what I was talking about. I'm hoping that idea catches on over the web some day but I'm not holding my breath. <G>!  BTW, Scott is a professor of engineering (I believe), Furrukh a professor of computer science (I believe) and Mark a business person who deals extensively with technology (I believe). I am a mere photographer and don't count myself as an expert with technology or engineering. It's useful to have these three (and many other) knowledgeable people on this forum who do know high-tech, engineeering, etc.  So I assume then that the reference pixel will not replace the encyclopedia? <G>  Cheers,  Sean Link to post Share on other sites More sharing options...
sean_reid Posted January 4, 2007 Author Share #35 Â Posted January 4, 2007 Â Who knows, lots of theories will abound. I'm a bit worried. Â Why? Â Cheers, Â Sean Link to post Share on other sites More sharing options...
sean_reid Posted January 4, 2007 Author Share #36  Posted January 4, 2007 Reference pixels are pxiels around the periphery of the sensor which are shielded from the light and are there to provide a black reference for the electronics. Without them, the charge coming off the CCD cannot be related to an absolute black. From what Sean says, it sounds like, somehow, these pixels are being illuminated when they shouldn't be and messing up the black level. Almost like the pixels are only black up to a certain angle of incidence or else they are being illuminated by internal reflections.  Who knows, lots of theories will abound. I'm a bit worried.  I'm told (by Leica) that the light source needs to be coming in at exactly two degrees to trigger this. That's my understanding, at least.  Cheers,  Sean Link to post Share on other sites More sharing options...
mike prevette Posted January 4, 2007 Share #37 Â Posted January 4, 2007 I'm guessing because it could mean another hardware revision. Adjusting baffles etc. But what bothers me about that theory is why it wasn't apearing in 1.06. I suppose it's possible they reallocated some o their reference pixels. Â Two degrees to what? two degrees angle of attack? two degrees under the baffle? Â _mike Link to post Share on other sites More sharing options...
khanosu Posted January 4, 2007 Share #38 Â Posted January 4, 2007 Hi Mike, Â There were also examples of this banding from earlier versions (1.06). This particular banding was being masked by the older variety of banding so people lumped it all as horizontal banding. Once the older banding was fixed by the hardware upgrade the new banding stood out more clearly. Â Furrukh Link to post Share on other sites More sharing options...
marknorton Posted January 4, 2007 Share #39 Â Posted January 4, 2007 I tried to reproduce this problem on my unmodified/1.09 camera and could not and need to go back to Furrukh's original posts and think about it. As I recall, as a bright light source moved off the edge of the image, there was a sweet-spot (I'd call it more a sour-spot) which induced horizontal banding into the centre. Â Sean has mentioned that it occurs at a particular angle, which sounds like some sort of optical effect contaminating those reference pixels. If so, how are they going to be able to tell when it's happening if your reference is being messed up? It's a bit like trying to measure something when you don't know whether your rule is accurate or not. Â That's why I'm worried, the nightmare scenario of more hardware changes. Â I'm also a bit worried about the firmware-jockey attitude - let's hack around in the firmware to try to identify these odd situations and correct them. The firmware is meant to extend and enhance the basic imaging system, not patch it up. Link to post Share on other sites More sharing options...
sean_reid Posted January 5, 2007 Author Share #40  Posted January 5, 2007 I'm guessing because it could mean another hardware revision. Adjusting baffles etc. But what bothers me about that theory is why it wasn't apearing in 1.06. I suppose it's possible they reallocated some o their reference pixels.  Two degrees to what? two degrees angle of attack? two degrees under the baffle?  _mike  I've wondered the same. Will get clarification before I write about this.  Cheers,  Sean Link to post Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.