5D Mark iii focusing issues

  • Thread starter Thread starter naterz
  • Start date Start date
Status
Not open for further replies.
One more example of my test. The dog was shot at ISO 800 70-200 f2.8, 1/8000, center focus point and Al Servo. I think it is more than except-able. The moose was shot at ISO 250, 300mm aperture f3.2 center focal point. Tight crop and no sharpening or color enhancement. Again, I think were off to a good start. These are from converted raw files using the Photoshop techniques available with DNG files. So these are my thoughts and results of my tests today. Would like to see other results and thoughts. Thanks everyone, I think I can sleep tonight.
 

Attachments

  • ObanFocusExample.jpg
    ObanFocusExample.jpg
    132.9 KB · Views: 1,850
  • ObanFocusExample2.jpg
    ObanFocusExample2.jpg
    257.7 KB · Views: 1,897
  • MooseFocusExample.jpg
    MooseFocusExample.jpg
    143.4 KB · Views: 1,878
Upvote 0
This whole situation worries me a little. Anyway I was not going to get a 5DmkIII soon because of the price and the fact that my 5DmkII pleases me (for static subjects) and is not used too much at the same time. However being a Canon fan with a lot of lenses I would like to know that everything is OK with 5DmkIII so as to continue relying on Canon. I do not want to switch. I have started using Canon since 1988 with an EOS620 and built upon it. Let's hope that what was mentioned here was an exception.
 
Upvote 0
PhotoMoose said:
One more example of my test. The dog was shot at ISO 800 70-200 f2.8, 1/8000, center focus point and Al Servo. I think it is more than except-able. The moose was shot at ISO 250, 300mm aperture f3.2 center focal point. Tight crop and no sharpening or color enhancement. Again, I think were off to a good start. These are from converted raw files using the Photoshop techniques available with DNG files. So these are my thoughts and results of my tests today. Would like to see other results and thoughts. Thanks everyone, I think I can sleep tonight.

While these are really nice Photomoose, they look a bit soft to my eye at 100% crop. I remember seing similar shots taken with a 1DIV which were much sharper (I think from The Digital Picture with Brian or maybe it was Brian S, from this forum)...
 
Upvote 0
Yes, that is my point. I made the switch from my mrkIV on the promise of improved focus. The 5Dmrk3 is much faster to focus, but not seeing better results. The improvements in high ISO is noticeable, but for my application (wildlife and scenic) shooting in the dark is not really applicable. Sharpness and speed of focus is key for me. Could the 61 point focus system be so technically advanced that many users will have difficulty using it and achieving focus? Could it be that advanced? I have established that I'm a fool for selling my mrkIV, now what is a fool to do? Return the 5Dmrk3 and than purchase what? I'm asking myself if it is time to jump ship? Considering the issues with this focus system not sure that the 1DX will be any different, after all the 1DX has the same focus system. The only other advantage of the 1DX is the weather proof body and 14 fps. I don't shoot in the rain and unless your shooting hummingbirds I don't need 14fps either. I have one more test and calibration test to do today, then its decision time. It still bothers me that canon released a camera with focus issues and DPP software that is not up to the task. Canon has been quoted saying that the software fix will take 'some time'. REALLY? What where they using when they tested this camera?
 
Upvote 0
Is the sharpness problem due to the DPP software problem, or is it due to lenses needing microadjustment? With my 5D2, all my L zoom and prime lenses did not need any microadjustment and it seems that the focus is factory calibrated to zero MA for most lenses. I use this MA method with great results so far http://www.northlight-images.co.uk/article_pages/cameras/1ds3_af_micoadjustment.html I am just so surprised that the 5D3 has so much sharpness issue and I hope that it is more of the DPP software issue rather than a hardware issue. I am holding back my purchase of the 5D3 until the dust settles on this and a couple other issues with the 5D3.
 
Upvote 0
That is part of my disappointment. I have not used the DPP software because of the known issues. I'm using the Adobe DNG converter and photoshop and than maybe importing into lightroom, but going from the converter to photoshop is the best option at this time (actually its the only option). At the risk of repeating myself, I'm uncomfortable with the fact that the camera is released without any software that can process the images. Its a big leap of faith to assume that all is well with the raw images when cannon has not released software that can process them and the raw images where not available to the pre-testers either. Since I'm not using the DPP software it seems to be a hardware issue, but It could be operator error. Maybe the focus system is so advanced that it requires schooling to achieve acceptable results. I just don't know. Learned a big lesson, never buy a pre-released product. My father taught me to never throw away your old shoes until you have a good new pair. I forgot that lesson this time.
 
Upvote 0
PhotoMoose said:
That is part of my disappointment. I have not used the DPP software because of the known issues. I'm using the Adobe DNG converter and photoshop and than maybe importing into lightroom, but going from the converter to photoshop is the best option at this time (actually its the only option). At the risk of repeating myself, I'm uncomfortable with the fact that the camera is released without any software that can process the images. Its a big leap of faith to assume that all is well with the raw images when cannon has not released software that can process them and the raw images where not available to the pre-testers either. Since I'm not using the DPP software it seems to be a hardware issue, but It could be operator error. Maybe the focus system is so advanced that it requires schooling to achieve acceptable results. I just don't know. Learned a big lesson, never buy a pre-released product. My father taught me to never throw away your old shoes until you have a good new pair. I forgot that lesson this time.

Have you tried ACR 6.7 beta?
 
Upvote 0
That happened to me when I was testing for any focussing issues with the 24L.
Opened to 1.4 and tripod mirror up focus using the middle point. The focus point and the actual focus spot was off by a little and a -5 adjustment did it perfectly.

Lent the camera and lens to a friend and he said it's soft at f/8 doing some landscape shots. Calibrated it back to +0 and everything was fine but not when it's at 1.4 >.<
 
Upvote 0
PhotoMoose said:
Yes, that is my point. I made the switch from my mrkIV on the promise of improved focus. The 5Dmrk3 is much faster to focus, but not seeing better results. The improvements in high ISO is noticeable, but for my application (wildlife and scenic) shooting in the dark is not really applicable. Sharpness and speed of focus is key for me. Could the 61 point focus system be so technically advanced that many users will have difficulty using it and achieving focus? Could it be that advanced? I have established that I'm a fool for selling my mrkIV, now what is a fool to do? Return the 5Dmrk3 and than purchase what? I'm asking myself if it is time to jump ship? Considering the issues with this focus system not sure that the 1DX will be any different, after all the 1DX has the same focus system. The only other advantage of the 1DX is the weather proof body and 14 fps. I don't shoot in the rain and unless your shooting hummingbirds I don't need 14fps either. I have one more test and calibration test to do today, then its decision time. It still bothers me that canon released a camera with focus issues and DPP software that is not up to the task. Canon has been quoted saying that the software fix will take 'some time'. REALLY? What where they using when they tested this camera?

I hear you buddy. I ask myself some of the same questions. I really hope the 1DX will prove better. There are too many of us experiencing softness in the image for it to be purely user error. Something is weird here. I am also starting to doubt that DPP issue and the fact we dont have a full release verison of LR4 that support the 5D mkIII will solve this. I picked up my 5D mkII last night again and man I found all my pictures to be sharper!

:-\ :-\ :-\
 
Upvote 0
JR said:
There are too many of us experiencing softness in the image for it to be purely user error. Something is weird here.

Or several somethings.

1) Recall there were many complaints about the 7D's AF system, too, when that camera was released. So, there's a component of AF system complexity. Using an AF mode not well suited to the task (e.g. Spot AF for sports) can result in blurry shots.

2) Many people don't have a strong understanding of how an AF system works, in particular the fact that an actual AF point is larger than its representation in the viewfinder (even with Spot AF), and that the AF system will lock onto the feature that gives the greatest phase difference at the proper orientation regardless of where in that AF point's area the feature falls. You know you're focusing on the eye, but if the hairline falls under the selected AF point, too, the camera may lock onto that as a stronger contrast - it neither knows nor cares that the eye is centered under the AF point.

3) There is a known issue with DPP that results in soft images with RAW conversion in High Quality mode (but not High Speed mode). At this point, ACR to DNG seems to be the nest approach, but it's an RC not final, and it doesn't run natively in LR yet.

So, it may be that if you combine all of these different factors, that accounts for the '5DIII image softness' - different people with different problems, all leading to the same result. Of course, there may be an actual hardware problem at play here, too - time will tell.
 
Upvote 0
I have the same problems with some of my lenses. I'm finding that my 200 f/2 is front focusing by about 50mm at f/2 from 3m. What worries me is that it is not every shot that has this problem?
More than this is also the abnormal noise from the IS system in AI Servo mode that I now avoid with this lense while I wait for a fix form Canon.
I also agree that this new SUPER focus system seems like it needs a SUPER user to get it to perform as I hope it can :-\

Update:
50 f/1.2 front focus
200 f/2 front focus
70-200 f/2.8 ii @ 70 spot on @ 200 front focus
400 f2.8 ii back focus allot!
:-\ :o
not nice so every lens is different and with every doubler the trouble doubles :'(
 
Upvote 0
I'm still in that paranoid group pondering if my lens/body is soft or not, and exchanging or more tests, altering settings etc. I kept trying out shots on our golden, trying to land her eyes, on the rare instances I did get her to stand still (which is no small feat for a 1 yr old pure golden, oh my gosh the energy and add!!!). I'm pretty convinced by like, two or three of every ten or more photos that I take, when I pixel peep just the standard jpegs at 100% that it's not soft because on the ones I do land, it's blowing me away. The soft ones, I can tell the focus was someplace else, be it her fur which in certain areas lands super sharp, or it was some other setting that threw something off.

I'm pretty sure a lot of the issues and focus problems are, in fact, my user error, and focusing on something so small and zooming in to 105mm on the kit lens. Just wish I could figure out how to embed/attach photos to my posts (could someone please shoot me a pm or something at least to say the easiest way they've done this??).

Anyone else have any updates on what steps they took for testing their 5D M3's?
 
Upvote 0
neuroanatomist said:
JR said:
There are too many of us experiencing softness in the image for it to be purely user error. Something is weird here.

Or several somethings.

1) Recall there were many complaints about the 7D's AF system, too, when that camera was released. So, there's a component of AF system complexity. Using an AF mode not well suited to the task (e.g. Spot AF for sports) can result in blurry shots.

2) Many people don't have a strong understanding of how an AF system works, in particular the fact that an actual AF point is larger than its representation in the viewfinder (even with Spot AF), and that the AF system will lock onto the feature that gives the greatest phase difference at the proper orientation regardless of where in that AF point's area the feature falls. You know you're focusing on the eye, but if the hairline falls under the selected AF point, too, the camera may lock onto that as a stronger contrast - it neither knows nor cares that the eye is centered under the AF point.

3) There is a known issue with DPP that results in soft images with RAW conversion in High Quality mode (but not High Speed mode). At this point, ACR to DNG seems to be the nest approach, but it's an RC not final, and it doesn't run natively in LR yet.

So, it may be that if you combine all of these different factors, that accounts for the '5DIII image softness' - different people with different problems, all leading to the same result. Of course, there may be an actual hardware problem at play here, too - time will tell.

Sorry missed your post neuro. These are all valid points and likely apply to a lot of the situation we are seeing in various threads. However i can only speak for myself here, but realizing the complexity of the af system, i did make sure i used only very simple mode for myself at first like single center points and still got some mixed result. I the end i lost patience to figure out why my new $4500 camera (had bought the kit) was not up to par to my current one.

Looking back i did take a lot of great shots, but the softness was still there...i just hope the 1dx produce more crisp image like my mkii does...
 
Upvote 0
I got my 5D3 last Friday 13th April! It is as sharp as a knife with all my Zeiss lenses. MA for MF did help a bit. LV tests were perfect. Not less sharp than the 5D2.
Borrowed a 70-200/4 IS for testing AF. AriHazegis MA procedure over MacPro-based LV. the lens did need +5 at 70 and 0 at 200. Tack sharp on wall testing. In real world shooting the AF tends to miss if misused: like Neuro says, AF is still only a machine, a hair in front of an eye and "off she goes".
Cheers
Wickid - I hope you get your second copy sharp and with light-leak-tight LCD!
 
Upvote 0
I found the same thing, HOWEVER, I have solved it. It doesn't really miss different at different distances, it's just that invisible adjustments up close, makes a MASSIVE difference at longer distances. I own only f1,2 and 1,4 lenses (for the most) where this AFMA can be a total bitch.

I have found that going +-5 steps at the time is pointless, as you can get almost any results. I for example found my 85 L front focus like nothing else at +4 but spot on at mfd, and then +5 it's dead on perfect at any distance. I never thought one step would make such a difference.

SO I have been back and forth, but all my lenses which suffered from missing way different on different distances all work at all distances now.

My 50 L looks to be quite a bit in front using the Spyder LensCal, but in real life it's dead on. So trying, in good and even light with real life subjects are extremely important. Also, indoor light for calibration is almost useless. But once calibrated properly in great light, it works in lower light also.

Just fwiw, I'm not saying no people can have an issue with their camera, but I got the same stuff, but was able to correct using real life subjects and one step at the time adjustment. It seems to get better after a +5 adjustment, but it really should be +3 for example, was hard for me to catch.
 
Upvote 0
Status
Not open for further replies.