Assuming you are serious, please explain why this happens:
All about 1/4000s, ISO 100, using live view (the picture I had few posts above).
Lens Carl Zeiss Sonnar T* 2.8/85:
- adapter happypage with chip: OK
- adapter no-brand ebay without chip: vignettes
- adapter happypage with chip but contacts taped: vignettes
- adapter happypage turned 30 degrees so that aperture displays "0": vignettes
Lens Canon EOS 85mm f/1.2 L USM mkII:
- lens normally in camera: OK
- lens turned 30 degrees so that aperture displays "0": vignettes
- lens normally in camera but taped connections: vignettes
Samuli Vahonen wrote:
Some of the adapters I also have wrong max value, for example Zeiss Planar 1.7/50 is coded to 1.8 and Zeiss 3.4/35-70 had to be coded to 3.5, since it's not possible to send 1.8 or 3.5 from lens to camera.
Call it a partial joke. It is the identical effect as seen on mirror graze where the mirror is getting caught during the exposure.
I agree with what you're seeing implies a software 'gotcha' with the new bodies and the use of un-chipped lens's. In your case the test would seem to clearly indicate that. But a software gotcha in what regard?
What in the light path would be causing such heavy vignetting in the same spot when using LV? Forget what triggers it to happen, but what actually physically causes the light blockage? It's not some subtle M8-like angled micro-lens vignetting correction that's failing to happen.
Since it appears consistently in one location, my bet - start looking at how the mirror(s) behaves in LV and then how chipped vs. un- chipped lenses could effect it's behavior. Could the body failing to sense a lens (e.g. an EPROM at it's contacts) they throw the timing off between the shutter trip mechanism and the mirror mechanics for LV use? That or something wonky with the shutter, but I can't see how/why they would change the shutter hardware from say the 5D to 5DII to accommodate LV.
Unfortunately, given as only happens with non-EF (or non-ZE) lenses, don't expect a resolution from Canon any time in the near future.
Conner, I don't think mirror has anything to do with this, wouldn't the vignetting then be affected to whole picture and not more on bottom right corner in the image?
I shoot some photos without lens = no connection and could not see the mirror hanging at any shutter speed, I even touched the backside of mirror when taking photos and it didn't move even slightest bit.
In the example I gave above the mirror travels as normal but just contacts the element shroud on the back of the lens. This slows the mirror down JUST enough in it's travel upward (bottom of frame in landscape mode) so that when the shutter trips, it catches the front edge of the mirror in the frame. Image captured, mirror returns to rest position.
This 'catch' produces a shadow across the bottom of frame in landscape. The effect I see on this thread is identical to that. Heavy strip of vignetting always at bottom of frame (landscape).
Turn the body 90 CCW for portrait and it's along the RHS, turn body 90 deg CW for a portrait and it's along LHS.
My guess is that firmware knows it takes shutter X ms to clear sensor FoV (+/- a deviation), it waits that long at shutter button hit, then trips - assuming mirror is physically clear. If not 100% so, you get a shadow captured.
Mirror is tight and secure, looks and feels normal. There is essentially nothing wrong with it, juts under those circumstances it isn't clearing the sensors field of view in time. How chipped and un-chipped could cause the same effect with LV - no idea, but I'd start by looking at how the mirror(s) is moved in LV mode.
The fact that folks like Lloyd are not seeing with his 1DS3 and assuming that LV implementation is the same between bodies (no reason why wouldn't be), the only real variation known to the 5D is the #$%^ mirror and variations in it's placement, etc.
Again, just a guess on my part based on prior experience. Will be interested to see what actual resolution/issue turns out to be.
Conner999 wrote:
In the example I gave above the mirror travels as normal but just contacts the element shroud on the back of the lens. This slows the mirror down JUST enough in it's travel upward (bottom of frame in landscape mode) so that when the shutter trips, it catches the front edge of the mirror in the frame. Image captured, mirror returns to rest position.
This 'catch' produces a shadow across the bottom of frame in landscape. Turn the body 90 CCW and it's the RHS, turn body 90 deg CW and it's LHS. My guess is that firmware knows it takes shutter X ms to clear sensor FoV (+/- a deviation), it waits that long at shutter button hit, then trips - assuming mirror is physically clear.
Mirror is tight and secure, it just doesn't clear the sensors field of view in time. I'll see if I can dig up a sample. ...Show more →
I know this phenomenon there was something about this in few years ago in the forum (this forum). However in live view the mirror does not move, so it cannot be slowness of the mirror causing this.
In live view the process of photo taking is:
1. the mirror is raised up when user press the live view button
2. shutter opened
3. user is able to focus and compose on the back LCD screen
4. when shutter is pressed the sensor is cleared (on old cameras shutter is closed for this e.g. 1DmkIII but new cameras like 5DmkII can do this while shutter is open)
5. shutter blades move over the sensor exposing it to light for correct time
6. shutter is closed and depending shutter speed and settings dark frame is shoot with closed shutter and finally image is written to card, then process goes back to step 2
When using camera normally you have to add mirror swinging actions to process, but mirror is all time pressed against the pentaprism when shooting with live view. So even end results are similar this has nothing to do with this "slow mirror" problem.
Resolution to the problem: use chipped adapters or buy chip and clue it to your chiples adapter.
I wonder if it is a result of a mistiming on the autoaperture and shutter opening sequence. If the body reconfirms its aperture and doesn't get a signal it could be engaging the shutter before it clears/activates the sensor
Daniel, this is also the reason why I assume this happens. However without knowing how the camera really works one can generate many theories, which may or may not be the cause.
You may very well be correct (on cause), but effect is eerily similar to non-LV mirror 'brush' and is always in same location and only with 5DIIs (so far anyway)...
There is possibility is shutter-related, but how non-chipped lenses (where body say would be set to 1.8 and just assume dark scene if lens is stopped own to say F4) would effect physical shutter timing I have no idea.
Some screw-up in the auto-aperture timing isn't a bad thought. But it is odd how the problem would vary between the 1Ds3 and 5DII given they probably share a lot of the same LV firmware code.
The solution is easy as you say (while limiting adapter choices or requiring some epoxy), but it would be interesting to see what the final determination is.
Conner999 wrote:
You may very well be correct (on cause), but effect is eerily similar to non-LV mirror 'brush' and given that it's always in same location and only with 5DIIs (so far)...
There is possibility is shutter-related, but how non-chipped lenses would effect physical shutter timing I have no idea. I'd bet if it's a firmware issue - it's related to something mechanical in the LV mechanism. Now how that would vary between the 1Ds3 and 5DII given they probably share a lot of the same LV firmware code, you've got me.
Be interesting to see what the final determination is.
Conner, live view works differently in 1Ds mkIII and 1D mkIII - they have to close the shutter in order to clear the sensor. 5DmkII in addition to 50D are the only Canons which can clear the sensor without closing the shutter. So there may not be that much common in firmwares of 1D(s) mkIII and 5DmkII.
For this reason I have been laffing when Canon says "you don't need mirror lock-up,use live view". This is bullsh#t for those older cameras, the shutter goes back and forth just before the exposure happens and with focal lengths greater than 150mm better results were achieved, even gigantic pixels size 1DmkIII, by using the real mirror-lock-up functionality. For 100mm lens it still was OK to use live view.
Sorry to come in late. I use CZ ZE 50mm and 85mm on a 5D MkII almost on a daily basis. I have been totally unable to reproduce this problem, despite all kinds of trials. This goes to confirm that, indeed, chipped lenses or adaptors work fine.
Samuli - Ah. That would seem, as you say, to lend credence to a theory of aperture-centric mis-timing re: sensor clearing of some form.
What of possibility in LV with some 5DIIs that shutter is not opening all the way or is easing back a bit after LV engaged, is not being noticed in LV and edge is getting captured when image captured? I can't recall if LV is 100% view.
That or simply on some 5DII units the mirror is not retracting as far as required. How a chipped adapter would change that scenario....would imply a firmware-related timing problem somehow triggered when the body doesn't have an eprom to read resulting in something getting either partially captured in frame - or not being cleared quickly enough.
As you say, a lot of variables and no resolution likely to be forthcoming from Canon given lenses in question.
Conner999 wrote:
That or simply on some 5DII units the mirror is not retracting as far as required. How a chipped adapter would change that scenario....would imply a firmware-related timing problem somehow triggered when the body doesn't have an eprom to read.
As you say, a lot of variables and no resolution likely to be forthcoming from Canon given lenses in question.
If I would have to bet, I would put my money on timing error on sensor clearing and exposure start when not having electronical connection to lens.
However this isn't as bad problem as it did first seem to be, it's limited to shutter speeds above 1/2000 and non-chipped adapters. Personally for me this has no effect at all, even if I would not have chipped adapters I would rarely notice; I rarely shoot at such high shutter speeds with alternative lenses, due to light being so harsh at middle of day and I don't shoot moving stuff with alternative glass.
What I don't like is that this might be just first Canon showing symptom, when used with non-Canon lenses. Will they in future disable completely the use without Canon lenses and half of the Sigma lenses stop working again etc.? If that is the direction, I will move to medium format for "art" photography and have one Canon/Nikon/whateverSLR just for family snapshots with one autofocus zoom (yuck!)...
Re-reading all the posts, my guess is you're correct re: cause. However, it would be interesting to see if 50D users see same issue. May help pin point if shutter or mirror-centric LV timing issue (smaller sensor & mirror). BTW- thanks for education on 5DII/50D vs 1Ds3 LV implementation- never bothered to explore it.
It is possible you're correct on direction Canon taking - not a pretty scenario. That said, I suspect the clueless wonders just don't realize or care that it's an issue for many folks and just plodded along. It is an AF zoom world after all. Makes me glad I stuck with my 1Ds2 as my last Canon body (replacement will be a Nikon). Nice old fashioned cheap-ass LCD that's not much good for more than checking histo and changing settings ;>
Conner999 wrote:
Re-reading all the posts, my guess is you're correct re: cause. However, it would be interesting to see if 50D users see same issue. BTW- thanks fo education on 5DII/50D vs 1Ds3 LV implemenation- never bothered to explore it.
Found it funny way (I had read it before from juzaphoto.com but somehow didn't understood it) shooting some night photos at long shutter speeds and resting my hand on tripod leg while shooting, then found out that there is vibration in tripod when shutter is opened, then compared it to vibration using mirro lock-up and I could not detect the shutter opening vibration. Then read Juza's page few times and realized what happens in camera. Then in practice learned through experimenting that live view is OK untill 100mm and above better use mirror lock-up.
Conner999 wrote:
It is possible you're correct on direction Canon taking - not a pretty scenario. That said, I suspect the clueless wonders just don't realize or care that it's an issue for many folks and just plodded along. It is an AF zoom world after all. Makes me glad I stuck with my 1Ds2 as my last Canon body (replacement will be a Nikon). Nice old fashioned cheap-ass LCD that's not any good for more than checking histo and change settings ;>
I'm stuck with Canon since I like the Zeiss so much, at the moment only possible choice is medium format for Zeiss look (Sony/Minolta doesn't yet have enough Zeiss primes available). Based on D3x I'm sure D700x will be interesting as a camera when it eventually comes out, if it ever does, for me not an option since only some Leica lenses could be adapted.
Conner999 wrote:
BTW - get that 35-70/4 yet?
Not yet, almost every day I go to ebay and almost click the "buy it now"-button, but then on other hand want to first see what Zeiss 3.4/35-70 is capable of. I promised to myself that tomorrow I go and shoot with the 3.4/35-70, which is really hard for me since I'm "short tele shooter", loving to shoot around 100mm but have difficulties to shoot at shorter focal lengths, specially at winter, but I try tomorrow.
Ah, that classic 'fatal hesitation' at the BIN button...
Have started moving my lenses over to only those that are Nikon mount (primarily CV SLs and SL2s), convertible to Nikon (my Leica Rs that will work via leitax) and adaptable to Nikon such as my Mamiyas and maybe some future Hassy/Zeiss glass. Initial plan is to use 1Ds2 with same and Nikon-EoS adapters and say D700 or used D3, then replace 1DS2 with D700x.
None of the new EoS bodies interest me (variety of reasons) an I want a couple of choice AF zooms and can't bring myself to play the Canon Lens Lottery anymore ;>
MFDB is the eventual target for slower-paced high-res shooting.
Conner999 wrote:
That or simply on some 5DII units the mirror is not retracting as far as required. How a chipped adapter would change that scenario....would imply a firmware-related timing problem .
I wonder if shaving a few mm off of the mirror would solve this problem? Perhaps it would be just enough mechanical compensation to negate the electronic problem.
If you shoot alt glass, mirror-shaving is something you'd want to or have to consider, anyway. So, this would be a win-win, sort of.
Samuli Vahonen wrote:
What I don't like is that this might be just first Canon showing symptom, when used with non-Canon lenses. Will they in future disable completely the use without Canon lenses and half of the Sigma lenses stop working again etc.?
A tempting theory. However, if Canon actually did that, the word would surely get out, and I'd imaging that would destroy their reputation in ways that... well, I wouldn't want to own Canon stock at that point.
I doubt it has ANYTHING to do with killing true alt lenses (perhaps an attempt to frustrate lens/body communication stuff with 3rd party lenses) and everything to do with tighter integration between lens and camera. We're such a small group all in all that it's not worth going after us.
My bet is the sensor reset timing when no aperture signal is detected. Sensor is reset while the shutter is already in operation instead of beforehand.
