It is to be expected that the components that go into any production camera from Canon or any other camera manufacturer be fully characterized through extensive testing. If there is a software work around, Canon can save face. If this is a QA issue, someones career is over. There could also be a component that drifts with temperature. That might require a redesign and then a recall. I have worked in the semiconductor industry for 40 years and have spent considerable time in Japan. You can bet your bottom dollar, Canon hears us. What they probably won't do is admit to the problem until their is a solution. There is substantial computational power in this beast and even if there is a drifting component, there might be a a firmware work around. If Canon stonewalls this issue, they will fall from their present position of market leadership. It is obvious that this problem is getting more and more press and I fully expect to hear from Canon in the near future.
Jeff
Jeff Winters wrote:
It is to be expected that the components that go into any production camera from Canon or any other camera manufacturer be fully characterized through extensive testing. If there is a software work around, Canon can save face. If this is a QA issue, someones career is over. There could also be a component that drifts with temperature. That might require a redesign and then a recall. I have worked in the semiconductor industry for 40 years and have spent considerable time in Japan. You can bet your bottom dollar, Canon hears us. What they probably won't do is admit to the problem until their is a solution. There is substantial computational power in this beast and even if there is a drifting component, there might be a a firmware work around. If Canon stonewalls this issue, they will fall from their present position of market leadership. It is obvious that this problem is getting more and more press and I fully expect to hear from Canon in the near future.
Jeff...Show more →
"If" a component or function is temperature sensitive there is no chance of a firmware only fix unless (1) hardware hooks were designed in to measure temperature and (2) hardware hooks were designed in to control the the output of the temperature sensitive component/function.
There would be no fix, but there might be a workaround. You are, however, correct that unless there is a mechanism via a hardware hook, there would be no option for a firmware workaround. Even if there is a firmware workaround, I would expect Canon to either replace the faulty part or do a redesign along with an ECN for future shipments at some point.
Jeff
jkurkjia wrote:
"If" a component or function is temperature sensitive there is no chance of a firmware only fix unless (1) hardware hooks were designed in to measure temperature and (2) hardware hooks were designed in to control the the output of the temperature sensitive component/function.
Regards,
Joe Kurkjian
in digital circuits, higher temperature lengthens "setup" and "hold" times. that's fine for hold time, but lengthening setup times can cause logic to not have enough time to capture a signal.
software/firmware can often control clock skew to parts of digital logic, which can give more setup or hold time at the expense of the other. in some cases, what you say is true in that you might not be able to find a clock skew setting that makes the circuit work under all temperatures. but often times, there could be, say, plenty of hold time margin, so skewing the clocks to give more setup time under higher temperatures would work even under lower temperature cases where hold time shrinks (due to the circuit being faster).
i hope that made sense.
bottom line: a firmware solution is still possible, but not guaranteed. it depends on the particular layout/design of the circuit.
I still shoot a MKIIN, and a MKIII. My MKIII has the same keeper rate as the MKIIN during daylight field sports mostly LL baseball so far. At night, (football under the lights), the MKIII has a better keeper rate than the MKIIN. I don't believe temperature has anything to do with the issue. Some of the LL baseball games were played at 98 degrees, with the camera in direct sun. The MKIII got so hot in those conditions, I thought the battery had failed. It still worked. I believe very high contrast situations are the issue not temperature. Here in Mississippi, a night time football game, the temperature at game time will be 90 degrees and 84 degrees at the end of the game. The difference is the light and contrast.
My personal opinion; there are two MarkIII's out there. One group has a calibration or component issue that causes the cameras to act erratically. I do not believe a firmware change or CF changes will fix these units. The owners will ultimately have to send these back to Canon for repair. The other group work correctly 98% of the time and CF changes have an effect but will not completely cure the problem for that other 2%. I believe firmware will fix these units. My camera falls into the latter group, and as I am shooting mostly night football, my camera is not a $4500 paper weight.
I also believe anyone with MKII experience, will not have a steep learning curve with the MKIII. A MKII user will have the III figured out in a matter of a few hours, not a few weeks as some have implied. Tweaking the CF-III setting can improve your keeper rate by a few percent, but it will not go from 35% to 85% just by changing a few setting.
For the record; I shoot the center AF point with expansion and the AF sensitivity set to the default. I set focus priority over drive speed for all frames. I haven't made up my mind about about CFIII-4, that is a new CF and I can see that as being sport dependant.
alfieri wrote:
in digital circuits, higher temperature lengthens "setup" and "hold" times. that's fine for hold time, but lengthening setup times can cause logic to not have enough time to capture a signal.
software/firmware can often control clock skew to parts of digital logic, which can give more setup or hold time at the expense of the other. in some cases, what you say is true in that you might not be able to find a clock skew setting that makes the circuit work under all temperatures. but often times, there could be, say, plenty of hold time margin, so skewing the clocks to give more setup time under higher temperatures would work even under lower temperature cases where hold time shrinks (due to the circuit being faster).
i hope that made sense.
bottom line: a firmware solution is still possible, but not guaranteed. it depends on the particular layout/design of the circuit....Show more →
Good point and it makes perfect sense but doesn't track with the low illumination (works good) and high illumination (works bad) comments that have accompanied many (but not all) of the "temperature problem" posts. , I shoot in the Phoenix, Arizona area and don't know what to think when somebody from Canada complains about high temperatures affecting AF; I'm sure you get my point. :-)
Desert_Watcher wrote:
i wonder guys, do you have a recommendation for setting custom function for photographing football at 12:00 pm where light is harsh and it's hot ?? with 1d mark iii, my focus didn't work well using 500mm F/4 !! serial number 5223XXXX firmware ver 1.1.0, please help is needed the temperature had reached 38C or 109F it was super hot in California
Was it a natural field or Field Turf (fake grass)? If it's not natural turf, there will be a lot of heat radiating from the field which makes shooting with long lenses frustrating. The lower (closer) you are to the field, the worse it gets. So, if you prefer to sit, stand. If possible shoot with a shorter lens, like a 300 or wait for the action to come closer.
For noon/1pm starts I prefer to shoot backlit.
I photographed a 1pm game under those conditions today, though somewhat cooler ambient temperature. With the 600, results were pretty bad, which I expected because the field is Field Turf. Images were generally quite soft.
rscheffler wrote:
Was it a natural field or Field Turf (fake grass)? If it's not natural turf, there will be a lot of heat radiating from the field which makes shooting with long lenses frustrating. The lower (closer) you are to the field, the worse it gets. So, if you prefer to sit, stand. If possible shoot with a shorter lens, like a 300 or wait for the action to come closer.
For noon/1pm starts I prefer to shoot backlit.
I photographed a 1pm game under those conditions today, though somewhat cooler ambient temperature. With the 600, results were pretty bad, which I expected because the field is Field Turf. Images were generally quite soft.
I am with Richard K for this. I think it is the harsh and brght light especially backlit situation but I don't discount the the heat effects because I am facing 30-36C almost everyday.
That crack about the weather was a bad one. We here in the Nations capital have had a hot summer. The thing with being in the Arizona you have no idea what humidity is! We had 3 weeks of over 40 degrees celcius this summer and it sucked big time!
I really want a mark 3 and I have been debating getting one for the longest time. I figure I will have one by the end of the month and I hope that the 1.1.1 firmware will do the trick of fixing this puppy!!
I've seen lots of theories about the issue, most of which are pretty easy to dispel, but I am deeply puzzled by the whole good/bad issue. Specifically, if there really are good bodies and bad bodies, how did RG get six bad ones? If you're a number dork like me, when you do the math, the chance of getting six random bad bodies from a group of some good some bad is almost zilch.
I really wonder if they aren't all bad, but that isn't something that's politically correct to wonder, since it would imply that those with good bodies don't know what they are talking about, and that is nearly as insulting as the fans insisting that all bodies are good and the issue is purely user error. However, what if the problem only manifested itself under certain conditions? We've already heard that heat and bright sunlight seem to make things worse, but what else might be a common factor among the bad ones?
One thing I really wonder about is the lens being used by those reporting problems. We know that the AF sensors behave somewhat differently on a lens f2.8 or faster than they do on a lens f4.0 or slower. Has anyone with a bad body had problems using a lens f4.0 or slower? Has anyone with a good body tried a lens f2.8 or faster (wide open with AI servo in bright light)? Birders seem to have good luck, but most of them are using the 500/4, often with the 1.4x. It seems like most of the trouble makers are using the 300/2.8.
Sorry about the speculation, but as an engineer that spends most of his time tracking down obscure problems, I can't help but ponder this mystery. As for my own body, it works great with short lenses in one-shot, but falls down when I put big fast glass on it (200/1.8, 300/2.8, etc.) and try AI servo in the sun. It would be fun to try the same tests with a 300/4 and a 300/2.8 (at f4) to see if that makes any difference, but I don't have a 300/4. I might be able to scare up a 400/2.8 and a 400/5.6, if anyone thinks that experiment might yield interesting results. Gotta have something to do while we wait for 1.1.1.
Actually if the good/bad ratio was 50% the chance of getting six consecutive bad copies would be 1 in 64.
The test done by pro photo ( http://www.prophotohome.com/forum/pro-photo-reviews-articles/73689-canon-1d-mkiii-autofocus-interactive-review.html )
explains a lot of the possible scientific errors that can occur in testing (see day 15/16 of their test).
One of the telling remarks made by pro photo was that they found it impossible to achieve the keeper ratio claimed by RG for the MkII (96%). What they found was keeper rates almost the same for both cameras.
One final remark I would make is that how many sports shooters actually shoot very long bursts? Personally I usually shoot only a small controlled number, less than 10, often only 5 or 6 to get the shot, otherwise CF cards fill too quickly using RAW. (yes even 8gb can be used up fast).
Jim Gilley wrote:
I've seen lots of theories about the issue, most of which are pretty easy to dispel, but I am deeply puzzled by the whole good/bad issue. Specifically, if there really are good bodies and bad bodies, how did RG get six bad ones? If you're a number dork like me, when you do the math, the chance of getting six random bad bodies from a group of some good some bad is almost zilch.
I really wonder if they aren't all bad, but that isn't something that's politically correct to wonder, since it would imply that those with good bodies don't know what they are talking about, and that is nearly as insulting as the fans insisting that all bodies are good and the issue is purely user error. However, what if the problem only manifested itself under certain conditions? We've already heard that heat and bright sunlight seem to make things worse, but what else might be a common factor among the bad ones?
One thing I really wonder about is the lens being used by those reporting problems. We know that the AF sensors behave somewhat differently on a lens f2.8 or faster than they do on a lens f4.0 or slower. Has anyone with a bad body had problems using a lens f4.0 or slower? Has anyone with a good body tried a lens f2.8 or faster (wide open with AI servo in bright light)? Birders seem to have good luck, but most of them are using the 500/4, often with the 1.4x. It seems like most of the trouble makers are using the 300/2.8.
Sorry about the speculation, but as an engineer that spends most of his time tracking down obscure problems, I can't help but ponder this mystery. As for my own body, it works great with short lenses in one-shot, but falls down when I put big fast glass on it (200/1.8, 300/2.8, etc.) and try AI servo in the sun. It would be fun to try the same tests with a 300/4 and a 300/2.8 (at f4) to see if that makes any difference, but I don't have a 300/4. I might be able to scare up a 400/2.8 and a 400/5.6, if anyone thinks that experiment might yield interesting results. Gotta have something to do while we wait for 1.1.1....Show more →
, I shoot in the Phoenix, Arizona area and don't know what to think when somebody from Canada complains about high temperatures affecting AF; I'm sure you get my point. :-)
