Panasonic GH7 Eye detection \ PDAF logic - How does it work?

[Interceptor121]

The fish in the two images is a Longstriped blenny

Both images are shot with Aninal eye detection and PDAF underwater

The first image is taken at f/10 the second is at f/2.8





You will see that in the first image the camera did not focus on the eye but just behind the second image instead is on the left eye.

I would have expected the first image to be more in focus than the second that has razor thin depth of field however the opposite happened

This makes me think that Panasonic PDAF focusses at the aperture dialled in, like sony mirrorless. It does not focus wide open and the close at shooting as I happens with DSLR.

So the camera nails the second shot but not the first.

I believe this is not just Panasonic but how PDAF works on mirrorless as the AF is done on sensor.

This means that the camera focuses better wide open than it does stopped down which is rather interesting.\

Sony has a setting that limits the aperture stop down to a number like f/11 to achieve focus which panasonic does not have but to be honest there is not a hell of a difference

For the record the camera with the macro lens is quicker to focus with DFD single AF even engaging the focus limited the focus motor move quicker in DFD (!?)

EDIT I am looking inside the lens while focus and it does not seem like it is stopping down unless you enable constant preview therefore I am at loss why it would have different performance

--
If you like my image I would appreciate if you follow me on social media
instagram http://instagram.com/interceptor121
My flickr sets http://www.flickr.com/photos/interceptor121/
Youtube channel http://www.youtube.com/interceptor121
Underwater Photo and Video Blog http://interceptor121.com
If you want to get in touch don't send me a PM rather contact me directly at my website/social media

 

[Interceptor121]

Update I have done more tests and I think I have understood what is going

firstly the camera without constant preview focuses with the lens wide open this is not an issue as long as you dont use a very fast lens (faster than 2.8)

The benefit is that with fast lenses the camera will focus quicker in low light

the flip side of the coin is that in bright scene when the camera is hitting base iso with the lens wide open and you need a fast shutter speed the camera may have challenges if you have a very fast lens

With the array of f2.8 and slower tele lenses this is not an issue but if you were using an f1.4 prime in bright light this could be overwhelmed

I no longer own such lenses so I am not able to stress test but in some situations I have seen the camera hunting

overall the pdaf system seems well designed

one word of caution the camera takes longer to focus in photo mode with constant preview on as it goes back and forth the aperture range when pressing the shutter to focus

 

[Chris R-UK]

Focusing wide open and then stopping down to take the image was a requirement in DSLRs because of the way that the PDAF sensor array worked. I believe that this was a geometric issue rather than a light issue, and there is no such limitation in MILCs, whether using CDAF, DfD or OSPDAF. However, AF certainly gets slower in an MILC in very low light, so there may be advantages in using wide open focusing.

My first serious MILC, a Panasonic GH2 with CDAF , had continuous shooting problems with certain early Panasonic lenses, particularly the 100-300mm Mk1, because the diaphragm mechanism couldn't close down and open up again fast enough between shots. This limited the fps rate to about 2.5fps on my camera when using the 100-300mm. So early MILCs certainly used wide open focusing.

It is my belief now that most MILCs still use wide open AF when using AF-S, but when using AF-C they may switch to stopped down focusing at higher fps rates. It is very difficult to get information on this. Some tests that I have carried out on my OM-1 cause me to think that on that camera there is probably a changeover above 20fps. I believe that Sony cameras may use stopped down focusing much more - this was a subject for discussion on Sony forums a few years ago. Stopped down focusing presumably reduces diaphragm wear.

I imagine that is certain circumstances, certainly in low light but possible underwater as well, all cameras may switch back to wide open focusing.

 

[Interceptor121]

Focusing wide open and then stopping down to take the image was a requirement in DSLRs because of the way that the PDAF sensor array worked. I believe that this was a geometric issue rather than a light issue, and there is no such limitation in MILCs, whether using CDAF, DfD or OSPDAF. However, AF certainly gets slower in an MILC in very low light, so there may be advantages in using wide open focusing.

My first serious MILC, a Panasonic GH2 with CDAF , had continuous shooting problems with certain early Panasonic lenses, particularly the 100-300mm Mk1, because the diaphragm mechanism couldn't close down and open up again fast enough between shots. This limited the fps rate to about 2.5fps on my camera when using the 100-300mm. So early MILCs certainly used wide open focusing.

It is my belief now that most MILCs still use wide open AF when using AF-S, but when using AF-C they may switch to stopped down focusing at higher fps rates. It is very difficult to get information on this. Some tests that I have carried out on my OM-1 cause me to think that on that camera there is probably a changeover above 20fps. I believe that Sony cameras may use stopped down focusing much more - this was a subject for discussion on Sony forums a few years ago. Stopped down focusing presumably reduces diaphragm wear.

I imagine that is certain circumstances, certainly in low light but possible underwater as well, all cameras may switch back to wide open focusing.

 

[Chris R-UK]

Focusing wide open and then stopping down to take the image was a requirement in DSLRs because of the way that the PDAF sensor array worked. I believe that this was a geometric issue rather than a light issue, and there is no such limitation in MILCs, whether using CDAF, DfD or OSPDAF. However, AF certainly gets slower in an MILC in very low light, so there may be advantages in using wide open focusing.

My first serious MILC, a Panasonic GH2 with CDAF , had continuous shooting problems with certain early Panasonic lenses, particularly the 100-300mm Mk1, because the diaphragm mechanism couldn't close down and open up again fast enough between shots. This limited the fps rate to about 2.5fps on my camera when using the 100-300mm. So early MILCs certainly used wide open focusing.

It is my belief now that most MILCs still use wide open AF when using AF-S, but when using AF-C they may switch to stopped down focusing at higher fps rates. It is very difficult to get information on this. Some tests that I have carried out on my OM-1 cause me to think that on that camera there is probably a changeover above 20fps. I believe that Sony cameras may use stopped down focusing much more - this was a subject for discussion on Sony forums a few years ago. Stopped down focusing presumably reduces diaphragm wear.

I imagine that is certain circumstances, certainly in low light but possible underwater as well, all cameras may switch back to wide open focusing.

Sony and nikon focus stopped down in all scenarios

sony to the dialled aperture nikon fixed to f/5.6 this do indeed create issues in low light

i dont know the canon rf system but Panasonic is the only one that focuses wide open I think this is due to mft always having dof but have not checked the s5iI

I am pretty certain that my E-M1.2 focused wide open in S-AF. I came to this conclusion when investigating the focus distances reported in the EXIF data a few years ago. My faster lenses gave very different answers at longer distance to my slower lenses, and an f/1.8 lens gave all distances as "infinity". My assumption was that these differences were caused by different DOFs and that the camera was therefore focusing wide open.

I came to a similar conclusion when investigating my OM-1 fps rates in one of the continuous shooting modes. Reducing the set aperture from wide open reduced the fps rate from 20fps wide open down to as low as 13fps at f/11. This could have been due to the lower light in stopped down focusing, but somebody else on the forum told me that his scientific tests had in fact shown that the diaphragm was opening up between shots.

This was in Silent Sequential mode which is variable up to 20 fps. The next fastest mode, SH2 at 25 fps (which doesn't vary with set aperture) behaves quite differently and is possibly both stopped down and not focusing on every shot.

Thanks for the info on Sony and Nikon. I always thought it possible that Sony used stopped down focusing and I now remember hearing what you told me about Nikon.

 

[Interceptor121]

Focusing wide open and then stopping down to take the image was a requirement in DSLRs because of the way that the PDAF sensor array worked. I believe that this was a geometric issue rather than a light issue, and there is no such limitation in MILCs, whether using CDAF, DfD or OSPDAF. However, AF certainly gets slower in an MILC in very low light, so there may be advantages in using wide open focusing.

My first serious MILC, a Panasonic GH2 with CDAF , had continuous shooting problems with certain early Panasonic lenses, particularly the 100-300mm Mk1, because the diaphragm mechanism couldn't close down and open up again fast enough between shots. This limited the fps rate to about 2.5fps on my camera when using the 100-300mm. So early MILCs certainly used wide open focusing.

It is my belief now that most MILCs still use wide open AF when using AF-S, but when using AF-C they may switch to stopped down focusing at higher fps rates. It is very difficult to get information on this. Some tests that I have carried out on my OM-1 cause me to think that on that camera there is probably a changeover above 20fps. I believe that Sony cameras may use stopped down focusing much more - this was a subject for discussion on Sony forums a few years ago. Stopped down focusing presumably reduces diaphragm wear.

I imagine that is certain circumstances, certainly in low light but possible underwater as well, all cameras may switch back to wide open focusing.

Sony and nikon focus stopped down in all scenarios

sony to the dialled aperture nikon fixed to f/5.6 this do indeed create issues in low light

i dont know the canon rf system but Panasonic is the only one that focuses wide open I think this is due to mft always having dof but have not checked the s5iI

I am pretty certain that my E-M1.2 focused wide open in S-AF. I came to this conclusion when investigating the focus distances reported in the EXIF data a few years ago. My faster lenses gave very different answers at longer distance to my slower lenses, and an f/1.8 lens gave all distances as "infinity". My assumption was that these differences were caused by different DOFs and that the camera was therefore focusing wide open.

I came to a similar conclusion when investigating my OM-1 fps rates in one of the continuous shooting modes. Reducing the set aperture from wide open reduced the fps rate from 20fps wide open down to as low as 13fps at f/11. This could have been due to the lower light in stopped down focusing, but somebody else on the forum told me that his scientific tests had in fact shown that the diaphragm was opening up between shots.

This was in Silent Sequential mode which is variable up to 20 fps. The next fastest mode, SH2 at 25 fps (which doesn't vary with set aperture) behaves quite differently and is possibly both stopped down and not focusing on every shot.

Thanks for the info on Sony and Nikon. I always thought it possible that Sony used stopped down focusing and I now remember hearing what you told me about Nikon.

in SAF the newer GH7 and G9M2 still use DFD nothing has changed there and the performance of autofocus is excellent