E-M1 CAF compared to D800 question

whumber wrote:

Great summary, just two points to add. First, like Photo Pete said the E-M1 is quite decent at tracking non-accelerating subjects in good light. Once the light level drops however, the C-AF becomes much less reliable. Second is that the initial focus acquisition time in C-AF mode is quite long. I was shooting a road bike stage race this weekend and brought the E-M1/40-150 along with my 1DX/70-200 and shot them side-by-side in a few situations. The E-M1 did pretty well in a few scenarios that fit the bright light/non-accelerating/plenty of time to lock focus requirements. Often times, however, the E-M1 struggled mostly due to the long focus acquisition time.

At one point I was camped out on a steep downhill section leading into the finishing straight. The racers would pop out from around a corner at ~50mph and quickly blast by me. The 1DX had no trouble locking focus pretty much instantaneously and tracking them even at f/2.8. I tried a few times with the E-M1 but because the initial focus acquisition time is so long, I wasn't able to get even a single in focus shot using C-AF.

Overall though, the E-M1 is still pretty decent at C-AF as long as you have plenty of time to lock focus, good light, and a non-accelerating target.

I agree with the somewhat slow acquisition time with the FTs lenses as well. However, with the 50-200 SWD I find that quick manual pre-focus solves most of this problem (it helps that I first started photographing moving targets with manual focus film cameras). I generally have no problem with targets where I can manually pre-focus with the EVF (the EC14+50-200SWD has mechanically linked focus)

I am not sure how the FTs lenses differ from the mFTs lenses in low light. I personally use Low Sequential (which is only PDAF with the FTs lenses) for the live view and find that the E-M1 tracks fairly well with relatively low light with fast targets which move in unpredictable ways(at least for bats). However, these are typically against a more homogeneous background which makes focus simpler.

See attached bat image below at 1/400 second, f5 and ISO 20,000 (sky lightened in the image - it was very low illumination) from one of my first set of attempts. That is about as low light as I would probably ever want to use a crop sensor for a moving target. My success rate on those conditions was 36%, which may sound low. However, I shoot with Rls Priority C = On (shutter release prior to focus without focus confirmation), so the initial images will not be in focus. In addition, since I used the dot sight, I could only zone pre-focus by setting the focus slightly shorter than infinity, resulting in more initial unfocused images. Also included are those situations where I failed to successfully follow the bat and the camera had to refocus after I corrected the tracking.

This percentage improved to 56% with better light (1/500, f4.9, ISO 3200) for the image I posted in my first response. That improvement may have been partially due to the camera focusing better with more light, but I was also able to follow the bat better with more light, so I don’t know if the improvement was due to the camera or the photographer.

I have never used the Nikon D800, but I would expect it to be somewhat better than the E-M1, since Nikon has been developing their PDAF CAF for far longer than the first generation Olympus mirrorless camera with PDAF.