Again, I set a LEGO figure in the centre of the image frame, and found the time from pressing the shutter release button to the camera taking the image. I turned on the camera just before testing the autofocus, which means that the focus is near infinity when first pressing the shutter release button.
I did the test under two different conditions. The first was in dim light, with a black background: There is artificial lights, and rather dim at around EV5. The distance from the camera to the LEGO figure was about 0.6m.
The second was with daylight coming in through the windows, and white background. The lightning was about EV9.
Here are a couple of examples:
Panasonic GH1, Lumix 20mm, dim lights, black background
Panasonic GH2, Lumix 20mm, dim lights, black background
Panasonic GH1, Lumix 14-140mm @ 140mm, daylight, white background
Panasonic GH2, Lumix 14-140mm @ 140mm, daylight, white background
And these are the timings I found:
|Lens||GH1, dim||GH2, dim||GH1, daylight||GH2, daylight|
|Lumix G 14mm f/2.5 pancake||0.43 seconds||0.43 seconds||0.37 seconds||0.20 seconds|
|Lumix G 20mm f/1.7 pancake||0.57 seconds||0.53 seconds||0.53 seconds||0.40 seconds|
|Leica Lumix DG 45mm f/2.8 macro||1.07 seconds||0.93 seconds||0.73 seconds||0.43 seconds|
|Lumix G 14-42 @ 14mm||0.43 seconds||0.33 seconds||0.30 seconds||0.20 seconds|
|Lumix G 14-42 @ 42mm||0.47 seconds||0.40 seconds||0.37 seconds||0.20 seconds|
|Lumix G HD 14-140 @ 14mm||0.40 seconds||0.33 seconds||0.27 seconds||0.17 seconds|
|Lumix G HD 14-140 @ 50mm||0.53 seconds||0.57 seconds||0.40 seconds||0.23 seconds|
|Lumix G HD 14-140 @ 140mm||0.70 seconds||0.53 seconds||0.50 seconds||0.33 seconds|
What we see here, is that the timings are remarkably similar in the tests done with dim lights. The GH2 has a small advantage to the GH1, especially with the fast focusing zoom lenses. But in this test condition, there is little to gain by using the GH2.
On the other hand, when testing the cameras with more light available, the difference is larger. The GH2 really excels in this test condition.
My previous test gave a larger difference between the two cameras. However, the GH1 firmware has been upgraded several times since that test, and so has the lens firmware. It seems that the GH1, with the up to date firmware, is still very capable.
As for the accuracy of the focus, it is very good with both the GH1 and GH2. Both cameras use CDAF (contrast detection autofocus). This means that the image sensor checks the actual image for focus before the camera takes the picture.
In contrast to SLR cameras, which use PDAF (phase detection autofocus). This means that there are separate AF sensors behind the mirror, which check the focus in some spots in the frame. These sensors must be calibrated to the image sensor, a process which is costly and complicated. Users of SLR cameras often worry that the camera/lens combination is back-focusing or front-focusing, i.e., that the AF sensors are noe correctly calibrated. This is something that users of Micro Four Thirds don't need to worry about.
Modern DSLRs can also use CDAF, which is generally refered to as "live view". However, this focus mode is often quite slow on DSRLs, since few lenses are optimized for CDAF. All native Micro Four Thirds lenses, and some Four Thirds lenses, are optimized for CDAF.