marknorton Posted September 15, 2009 Share #21 Posted September 15, 2009 Advertisement (gone after registration) I know from my own experience that programming tightly coupled processors each doing their own thing to recover gracefully from a resource bottleneck somewhere in the system in all circumstances and timing conditions is pretty tough. The RAW data buffer is there as a sort of parking lot to hold image data which has been captured but not yet processed or processed and not yet written to the data card and the aim is almost certainly to try to safeguard and retain the last capture rather than dump it which would be the easier approach to follow. I expect there's a timing condition in there which Leica/Jenoptik do not yet handle correctly but by their nature, it's tough to know whether you've handled every possible scenario. Link to post Share on other sites More sharing options...
Advertisement Posted September 15, 2009 Posted September 15, 2009 Hi marknorton, Take a look here My M9 is eating SD cards [MERGED}. I'm sure you'll find what you were looking for!
jrc Posted September 15, 2009 Share #22 Posted September 15, 2009 I know from my own experience that programming tightly coupled processors each doing their own thing to recover gracefully from a resource bottleneck somewhere in the system in all circumstances and timing conditions is pretty tough. The RAW data buffer is there as a sort of parking lot to hold image data which has been captured but not yet processed or processed and not yet written to the data card and the aim is almost certainly to try to safeguard and retain the last capture rather than dump it which would be the easier approach to follow. I expect there's a timing condition in there which Leica/Jenoptik do not yet handle correctly but by their nature, it's tough to know whether you've handled every possible scenario. Can it be fixed? Link to post Share on other sites More sharing options...
sean_reid Posted September 15, 2009 Share #23 Posted September 15, 2009 Do tell Sean! I've developed a pretty quick technique on my M8 classic of removing the baseplate, battery, reinserting, deleting the last frame on the card (as its usually corrupt from the freezeup) and going back to shooting..... Any series of frames leading to the crash, I usually consider lost and try to reshoot. Not much to tell. Sometimes it would loose the pictures in the buffer and sometimes not. Cheers, Sean Link to post Share on other sites More sharing options...
andrewrossphotographs Posted September 15, 2009 Share #24 Posted September 15, 2009 Can it be fixed? If you look at my last post in the thread, there may not be the same problem with the M9 as the video shows a M8 being cleaned and not a M9. Link to post Share on other sites More sharing options...
marknorton Posted September 15, 2009 Share #25 Posted September 15, 2009 In the video of the assembly on the M9 when a tech was going to clean the sensor, @ 14m 40s +/-, when she first tried to enter Sensor Cleaning Mode the camera froze making her take out the battery and reinsert it. This same thing has happen to every M8 and M8.2 ever made at one time or another. Yes, I saw that too and the dummy battery pack with the wires coming out looks just like the one I put together for the power consumption stuff I did. Makes you wonder whether the technician has ever mentioned it to anyone or else it's like taking the car in to be fixed, "normal - they all do that"... Interesting to note too that far more use is made of the USB port in the camera calibration and setup than we ever do in the real world. Link to post Share on other sites More sharing options...
dalippe Posted September 15, 2009 Author Share #26 Posted September 15, 2009 If you look at my last post in the thread, there may not be the same problem with the M9 as the video shows a M8 being cleaned and not a M9. Please read the very first message in this thread. Link to post Share on other sites More sharing options...
marknorton Posted September 15, 2009 Share #27 Posted September 15, 2009 Advertisement (gone after registration) Can it be fixed? Don't see why not. A camera like this has a number of asynchronous activities running at once - shutter firing and winding, exposure assessment, image readout, processing, display, card writing, even the flashing of the red LED is an asynchronous activity which will be handled by a software timer somewhere. All is fine providing no-one messes up and the problem comes the more you try to overlap the activities to increase throughput/burst capacity. The accountants only want to put, say, 256Mb of RAM in there but the marketing people want a burst of 10 shots - immediately conflicting requirements. Things will get worse if you try to predict along the following lines: "the RAW data buffer is currently full, but by the time the next exposure and readout has happened, there should be space, so it's OK to fire the shutter and capture the image". Assuming there is no straightforward bug in the code, one way to address this might be to be more conservative with their memory management. Link to post Share on other sites More sharing options...
GlennB Posted September 15, 2009 Share #28 Posted September 15, 2009 This may be a little off topic, and may just be a coincidence, but I've only had this problem a couple of times with my M8. I have two 4 gig Sandisk Ultra cards and one 8 gig Sandisk Extreme III card. Both times that the camera locked-up, I was shooting several (5-7) photos in a row in single mode with the 8 gig card. Put a 4 gig card in, and it hasn't happened again yet (knock-on-wood). Could it be that the processor Leica is using is not happy with the larger OR faster cards ? The Extreme III card I have is the 30 mb/s rated card. Ultras are rated at 15 mb/s. I seem to remember that the M8 could only use the larger/faster SDHC cards after firmware update 2.004. Like I said, maybe just a coincidence, but it makes me wonder. Link to post Share on other sites More sharing options...
35mmSummicron Posted September 15, 2009 Share #29 Posted September 15, 2009 From my experience of owning an M8 for close to 3 years now, cards don't matter (I've used many different SD and SDHC cards) and all firmware upgrades since Feb 2007 have not rectified the problem. This may be a little off topic, and may just be a coincidence, but I've only had this problem a couple of times with my M8. I have two 4 gig Sandisk Ultra cards and one 8 gig Sandisk Extreme III card. Both times that the camera locked-up, I was shooting several (5-7) photos in a row in single mode with the 8 gig card. Put a 4 gig card in, and it hasn't happened again yet (knock-on-wood). Could it be that the processor Leica is using is not happy with the larger OR faster cards ? The Extreme III card I have is the 30 mb/s rated card. Ultras are rated at 15 mb/s. I seem to remember that the M8 could only use the larger/faster SDHC cards after firmware update 2.004. Like I said, maybe just a coincidence, but it makes me wonder. Link to post Share on other sites More sharing options...
spylaw4 Posted September 15, 2009 Share #30 Posted September 15, 2009 Do tell Sean! I've developed a pretty quick technique on my M8 classic of removing the baseplate, battery, reinserting, deleting the last frame on the card (as its usually corrupt from the freezeup) and going back to shooting..... Any series of frames leading to the crash, I usually consider lost and try to reshoot. Regrettably, as I have recently experienced, this can sometimes not be an instant fix, although it is the first thing to do. The fevered processors or whatever may need a night's rest to "cool down" from their hissy fit. It was also suggested that recovery was more certain with a really full battery and fresh card. Certainly worked for me. It is disappointing that this still appears to be a glitch in the M9. But forewarned is forearmed as they say. Link to post Share on other sites More sharing options...
scott kirkpatrick Posted September 15, 2009 Share #31 Posted September 15, 2009 As I recall, somewhere in the firmware evolution for the M8, around the time that upgrades and the M8.2 appeared. we went from timing lockups that battery removal wouldn't fix to a situation in which most lockups were recoverable by battery removal. This felt like progress to me and maybe to most of us, so it didn't become a big forum issue. The most common circumstance was when folks who liked the discreet setting (like me) would override it with C mode when it became necessary to take a few shots in rapid succession. This is probably particularly hard to track down because i suspect that the race conditions occur in the overlap between fiunctions programmed by Jenoptik and by Leica. It's simply time to dig into those issues again. Since the working relationship between Leica and Jenoptik on the M9 sounds strong at the moment (remember the talk about Leica bringing all of this in house previously?), I am hopeful that we will see some progress on both lockups and card writing speed. scott Link to post Share on other sites More sharing options...
wlaidlaw Posted September 15, 2009 Share #32 Posted September 15, 2009 My personal view is that a short term battery voltage drop over a burst could be a factor in some of these lock ups. I have noticed that virtually every time a lock up (burst or single) has happened to to me the battery on my M8 has been at one or two bars and not fully charged. I now always swap out batteries at one bar and twice a year, run them all the way down by turning off the auto switch off and running all three down to empty. The other thing I always do if I have time, is to reload the firmware after a lock up. I carry an otherwise useless 16mb card in my bag with the firmware on it and the lock switch activated. Finally I clean the battery terminals and the in camera terminals from time to time with switch cleaner fluid. It is getting on for a year and about 6000 shots since I last had a lock up. I also think that the latest 2.004 firmware may have improved things. I think that Sean said that he was using the pre-production firmware when these lock ups happened. I hope so. The more intensive processing required by the M9 for a uncompressed DNG burst, would I think cause an even bigger short term battery drain than the M8. Wilson Link to post Share on other sites More sharing options...
adan Posted September 15, 2009 Share #33 Posted September 15, 2009 Wilson - could be. I only just recently had my first ever two M8 freezes due to batteries that suddenly went from 2 bars to zero. Both occurred at the end of rapid (but not "C" mode) sequence shooting. Partly my fault because I had not 'calibrated' them by full discharges recently (been using the cameras a lot). Link to post Share on other sites More sharing options...
jaapv Posted September 15, 2009 Share #34 Posted September 15, 2009 The magnifying effect of the Internet - Sean has discovered and reported a problem during what can only be described as a torture test, not during normal shooting. I am sure that Leica will adress this, if possible. In the meantime, let's wait for a report of this ocurrence during actual use of the camera, instead of getting all excited at having discovered the first "fatal flaw". Link to post Share on other sites More sharing options...
michali Posted September 15, 2009 Share #35 Posted September 15, 2009 Second day out with the M9 I was playing with the exposure bracketing and after the 3rd shot the camera froze up. I removed the battery and re-inserted, all back to normal. I've since taken another 40 to 50 shots and tried to replicate the problem, it hasn't happened again, touch wood. I had this happen to me a few times with both the M8 & M8.2 when both cameras were relatively new as well. Link to post Share on other sites More sharing options...
TimF Posted September 15, 2009 Share #36 Posted September 15, 2009 Sure ... it's your camera, you are free to do whatever you want. At 2-3 frames per sec, it still sounded a little silly to use it to capture the decisive moment versus 8 frames per sec on a D3. Henri Cartier-Bresson. How many frames per second had he? (or should that be how many seconds per frame). Link to post Share on other sites More sharing options...
UliWer Posted September 15, 2009 Share #37 Posted September 15, 2009 From my experience of owning an M8 for close to 3 years now, cards don't matter (I've used many different SD and SDHC cards) and all firmware upgrades since Feb 2007 have not rectified the problem. Strange. I had the "freezing" problem several times in the beginning. At the same time it was often reported here. After a firmware update in 2008 - not quite sure which one - i havn't noticed this isssue any more and also the forum turned silent about it. I think there has been a change of firmware which cared for the problem, so if there is anything like this on the M9 they should be able to do the same. Link to post Share on other sites More sharing options...
bjmike Posted September 15, 2009 Share #38 Posted September 15, 2009 If this is a genuine issue, why should Leica fix it if people are willing to shell out $7000.00 for the M9 "as is"? Ah, geez! These things should work. As mentioned above, if Leica provides a given setting, it should work -- not cause the camera to fail. And: that the M9 is US$7,000 is irrelevant. It shouldn't happen w/ any camera whatever the cost. Buyers aren't buying a 15 year old used camera in a garage sale, for Pete's sake. Under U.S. law (at least) there is something called a "Warrant of Merchantability", meaning that something works for the purpose its intended for. If you use a new M9 as a boat anchor on your bass boat, its not covered and (failing to work) you have no claim against the seller. If a camera fails to take pictures, you have a claim under U.S. law (I believe its in the Uniform Commercial Code, adopted by every state except Louisiana). Link to post Share on other sites More sharing options...
scott kirkpatrick Posted September 15, 2009 Share #39 Posted September 15, 2009 It doesn't take a "torture test" to lock up an M8.2. When I was trying rapid shooting to see how Ultras and Extremes compare, it took me only my first 3-4 rapid single (S) shots with a fresh battery to lockup and lose the shots in the buffer. I did the rest of the tests in C mode, with no problems. As I said above, it seems time to get this looked into and fixed. I don't think that this discussion is going to spoil in any way the excellent first impression that the M9 has made. scott Link to post Share on other sites More sharing options...
wlaidlaw Posted September 15, 2009 Share #40 Posted September 15, 2009 Henri Cartier-Bresson. How many frames per second had he? (or should that be how many seconds per frame). ....and Louis Daguerre took about 1 shot every 20 minutes. If C mode had been available to HCB, I am sure he would have been using it like crazy to catch that moment. Unless of course, it locked up the camera every time. If my theory about battery drain has any basis in fact, then maybe a new camera where everything is stiff and the battery not quite fully formed, might be more susceptible. I just ran a quick torture test on my M8 (which has had about 15000 shutter activations in total) of 10 sequences of 11 continuous shots on DNG only and it was fine. Hardly a representative sample I know but this might be a problem a) that gets better as the camera beds in and is improved by later firmware. It is not tempting me to phone up and cancel my M9. On the contrary I am waiting on tenterhooks for the phone to ring to say it has arrived at the dealer (now overdue by one whole day - impatient or what!) Wilson Link to post Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.