Panasonic Lumix GH3 Predictive AF Tests

Posted May. 16th, 2013 by Daniel J. Cox

Awhile back I wrote about the new Panasonic GH3 and how it performed in Kenya, Africa. I was quite pleased overall but I knew there was one other test that had to be conducted to really see if this camera was up to the standards as my beloved Nikons. Those of you who follow this Blog may recall my Predictive AF test with the Nikon D4, D800 and D600 a few months back. Like those tests, for the GH3, I chose to borrow a good friend and his dog to put the Lumix GH3 through a series of predictive AF tests that are guaranteed to tell the real story when it comes to shooting action subjects. You can read my original Nikon post for more details, but in short, I have a dog sit out front of the camera at 50 yards, her master throws the ball over my head, and the dog charges full speed towards the camera. I’ve not found any better subject than a fast dog coming straight at the camera. The GH3 was set to its fastest – 6FPS with the AF in Predictive Mode, single AF sensor, right in the middle of the EVF frame.

The Bozeman Hawks host a regional track and field meet in Bozeman, Montana.

The Bozeman Hawks host a regional track and field meet in Bozeman, Montana. Click on the image to see a series of track and field images captured with the GH3 in predictive AF.

I’ve uploaded every image I shot so you can see how difficult it is for the photographer to keep the GH3 on the target. Any image where there is no dog in the frame is a situation where the dog bolted out of view, either left or right, and due to not being able to see that movement, the frame is missing the subject. As you’ll see, this happened several times. It also happened with the track athletes but not nearly as often since they were so much slower than the dogs.

Here is the link to the GH3 test on Track Athletes.

Here is the link to the Gh3 test on running dogs.

Tired pooch takes a break

Tired pooch takes a break.

Add Your Voice!
There are 13 comments on this post…
  1. Edward AndersonOn Aug. 14th, 2014

    A thought as I read through your reviews Dan – and while I realize this review is for the GH3 – there is now the GH4 have you had any hands on experience with the GH4? – are these issues still present?

    • Portrait of Daniel J. Cox

      Daniel J. CoxOn Aug. 14th, 2014

      Edward, I currently am shooting 2 GH4’s and they are a huge step up from the GH3. I plan to write about them in more detail in the not too distant future. Thanks for stopping by.

  2. Paul LatoucheOn May. 21st, 2013

    Daniel,
    Can you tell us how many of the dog pictures were 1 star vs 3 stars vs 5 stars (regarding focus), like you did for the Nikons? Or a rough estimate? I guess there won’t be many 5 stars…

  3. AdamOn May. 20th, 2013

    Try this with a black lab…I use a pen epl3 and that camera has a difficult time. I think because the camera thinks the dog is a shadow.

  4. AdamOn May. 20th, 2013

    Hi – I am really impressed with the results and your efforts on doing this! It seems that the CDAF system can keep up with the PDAF systems but the limiting factor is now the EVF vs. OVF. Eventually, the EVF’s will get better to the point where they will be able to replace the OVF for sports photography but in the meantime have you heard about some users adding a red dot sight to the camera? With a $50 RDS and a hotshoe adapter, I have seen some great results from point & shoot users. With a RDS on the GH3, it would make a great combo!

    • Portrait of Daniel J. Cox

      Daniel J. CoxOn May. 20th, 2013

      Adam, you’re right on as far as your conclusion that the limiting factor is the EVF for Mirrorless cameras. I’ve spoken to a Panasonic engineer who says the build up of images within the viewfinder, when shooting at high FPS, is something they have chosen not to change at this point. I told him that until they basically get streaming video to run through that EVF they will never compete in the action category. If anybody building these cameras would borrow a fast dog and do the test I did they would realize there is a huge problem inherent in the mirrorless systems. Then maybe that would send them back to the drawing board to engineer a real fix for the dismal performance in high speed action situations. Thanks for joining the discussion.

  5. Paul LatoucheOn May. 20th, 2013

    As I told you in one of your previous posts, you cannot track movement within the frame accurately with the GH3 set at high speed (6 FPS). You will only get picture previews and won’t see what you are aiming at. If you use medium speed, you will get some “live view”, following a black curtain between pictures, so you will be able to keep your subject within the frame. Never used GH3, but it works this way on the GH2. Of course, medium speed would be a bit slow for your dog test.

  6. RichardOn May. 20th, 2013

    Try the same test with the 35 to 100 lens I bet you get a different outcome

    • Portrait of Daniel J. Cox

      Daniel J. CoxOn May. 20th, 2013

      Richard, I don’t think so. It’ snot the speed of the lens, the lens did jut fine. It’s the buildup of images within the EVF that blocks your view of the fast moving subject. It has nothing to do with the lens. Thanks for your input however.

  7. Andrew SmallmanOn May. 19th, 2013

    Hi Daniel. Thanks for posting. I have spent some time with the GH3 on moving subjects. You can read more about it on my blog at http://cameraergonomics.blogspot.com.au The following findings from my studies on the matter may be of interest.
    1. Set Burst Mode to M, not H. M gives nominally 4 fps, actually 4.6 fps in my tests with a fast card. M gives you AF, AE and Live View on each frame. In M Burst, my GH3 shoots 28 RAW frames in 6 seconds before slowing down. Actually Burst mode L is not bad either, nominally 2 fps but actually a bit faster with a fast lens.
    2. The Lumix 100-300 mm lens is slower than any of the more recent models. I mean it is slower to focus, the aperture diaphragm is slower to operate and it’s fps rate is less than the Oly 75-300 or Lumix 45-150 for instance. I wish Panasonic would upgrade this lens’s operating speed.
    3. I need to do more tests comparing [Focus Release Priority] Release-vs-Focus to see if it makes a difference.
    4. I agree the EVF is the rate limiting factor with follow focus [predictive focus] work.

  8. Steve BarnettOn May. 19th, 2013

    What will work is the combination GH3 (or G5) and Oly 75-300. For whatever reason it’s a lot better behaved in AFC mode than the 100-300 on a G camera or the 75-300 on an EM5. Use it in bright light, stop it down 1/2 stop , set it to 4/second bust (so as to keep live view) and it will work very well.

    • Portrait of Daniel J. Cox

      Daniel J. CoxOn May. 20th, 2013

      Thanks Paul. I don’t think so. It’ snot the speed of the lens, the lens did jut fine. It’s the buildup of images withinthe EVF that blocks your view of the fast moving subject. Richard, I don’t think so. It’ snot the speed of the lens, the lens did jut fine. It’s the buildup of images within the EVF that blocks your view of the fast moving subject. It has nothing to do with the lens. Thanks for your input however.

Add your voice to this conversation

Your email address will not be published. Required fields are marked *

In an effort to combat spam, your comment may be held for a brief moderation period.

This site uses Akismet to reduce spam. Learn how your comment data is processed.