Hi guys, I was not planning to post a follow up to the original thread, however, since yesterday I have received many emails requesting a follow up post regarding the performance of 7D AF system when tracking birds in flight, so I decided to post a second thread.
Since the raptor shoot last weekend, I have been able to speak with CPS and learn more about the features of the new AF system, according to them the 7D AF system is supposed to be a step up from the TTL-SIR module currently used in 1DMKIII and 1DSMKIII cameras. The main feature of this new AF module is that it is coupled to the RGB meter and uses color information from the 63-zone RGB CMOS to help track the subjects in AI-servo mode. No other EOS camera has this feature.
This morning, I met with another shooter who also had a 7D, the two bodies were purchased new from different vendors at different times and the serial numbers were completely different in their entire digits.
We went to Palo Alto Baylands nature preserve to do a shoot, a world-famous site for shore birds, I made about 500 photos. The conditions were very different from those in Half Moon Bay, in this case most species are bright white against low contrast shallow water or sky in early morning light. Unlike Raptors, shorebirds like fly in more predictable patterns, usually straight lines or large circles so the behavior is less erratic.
I did not change any of the camera settings since the raptor shoot.
Here is what I learned:
• Both cameras behaved similarly, there is no indication that my camera is a “bad copy” or it needs calibration.
• When using a static subject (a chain with lots of contrast attached to a post) both cameras exhibited unstable behavior in AI-servo mode and 4-point expansion, both would randomly lose focus and lock on BG in this mode, so by now I am confident the performance of AF expansion mode is erratic at best, I did not use this mode for any of the subsequent shots.
• In single center AF point, tracking Willets against plain BG, results are mixed there are sharp and soft shots within a sequence.
• When tracking Willets against plain BG at close range in Zone AF mode, camera had much better success rate, delivered higher numbers of in focus shots, and it was able to capture “peak of landing” moment (see one sample below)
• Static focus stability with zone AF still remains mixed, please see example.
• I also tested static AF stability in spot AF mode, in some cases camera does focus on tiny features (see the spider example) consistently and delivers a sequence of sharp photos and it some cases it does not deliver a sharp photo at all.
• Overall The keeper ratio was much higher compared to raptor results and camera was able to capture the “peak of action” moment, even if AF was just slightly soft. In other words there were few “full blown” AF shots as in the raptor case.
It is now clear to me that Canon’s new AF system is optimized for speed and it performs best when tracking high contrast, large subjects against low contrast plain BG with a different color. It does not perform well in the cases of erratically moving camouflaged raptors against busy BG, in these cases camera AF is apparently fooled by similarity of color between subject and BG, in these conditions a Canon 1DSMKIII TTL-SIR or the Nikon MULTICAM3500FX module is clearly superior. Overall, my impression of the new Canon AF systems remains mixed, it has potential but it also requires fixes , with current firmware I can see it lead to situations like those experienced in summer 2006. I am sure that there will be people who disagree and will attest to its 100% perfectness, but this is my honest opinion. I hope that Canon provides a mode that can disable color-tracking feature and improves overall AF stability, especially in expansion mode and under busy BG conditions.
Regards,
Arash
Willet landing, EOS 7D and 400 f/5.6 L