NEX jpeg settings affect AF?

[bakubo]

1. Several people have explained how if you shoot raw you can lower the contrast, use natural or portrait picture style, etc. to get more dynamic range in the EVF. This has come up with regards to the NEX cameras and also about the Panasonic and Olympus m4/3 cameras.

2. Something else that has recently come up with regards to the Panasonic and Olympus m4/3 cameras is that if you shoot raw and increase the contrast, sharpness, use vivid, etc. then the CDAF can often focus better and lock-on faster in low contrast situations. Also, that when using continuous focus that it can better keep up with subject movement. There have been several threads about this recently on another forum with multiple people trying it. Still some controversy and various possible explanations. Of course, also the typical fighting, name calling, etc. mixed in so lots of garbage in the threads too.

Unfortunately, #1 and #2 use opposite jpeg settings and they will change the EVF.

I thought I would pass on this so that anyone who has a NEX and wants to experiment can. I have no idea whether it might apply to NEX also or not.

[David Kilpatrick]

There is a way of testing this using focus peaking. If the peaking kicks in sooner, and appears stronger, with higher contrast set then CDAF may indeed work better - but the EVF will look horrible.

Here's the actual truth, because I know how CDAF works. Higher contrast if fed to the CDAF detection process will result in LESS ACCURATE focusing but FASTER lock. Lower contrast results in the opposite, more accurate focusing but it will take longer. Extremely low contrast may result in no ability to focus at all.

You can improve CDAF by using a wide aperture, the optimum which Sony apparently think works is f/3.5, because this is the default for video even if you have a faster lens. One of the big hidden issues with the NEX cameras is that you are always focusing at working aperture (my review of the NEX-7 will try to cover this when finished...). So if you want to shoot at f/16, you end up focusing at f/16 too. That works OK with a 200mm lens but with the 16mm for example it almost removes the AF operation entirely. The lens rarely refocuses at all, because at f/16 it believes the image is sharp no matter what you do.

This is where the A77 scores over the NEX-7 - you focus wide open when using very small apertures, and this enables you to pick the exact plane of focus precisely - even if you use focus peaking. The errors in true focus point which result from focusing at small apertures, even regular settings like f/8, are significant with the NEX-7 especially and those who choose to work with manually opened and stopped down classic legacy lenses are bypassing this inherent flaw in the system.

David

[InTheSky]

There is a way of testing this using focus peaking. If the peaking kicks in sooner, and appears stronger, with higher contrast set then CDAF may indeed work better - but the EVF will look horrible.

Here's the actual truth, because I know how CDAF works. Higher contrast if fed to the CDAF detection process will result in LESS ACCURATE focusing but FASTER lock. Lower contrast results in the opposite, more accurate focusing but it will take longer. Extremely low contrast may result in no ability to focus at all.

You can improve CDAF by using a wide aperture, the optimum which Sony apparently think works is f/3.5, because this is the default for video even if you have a faster lens. One of the big hidden issues with the NEX cameras is that you are always focusing at working aperture (my review of the NEX-7 will try to cover this when finished...). So if you want to shoot at f/16, you end up focusing at f/16 too. That works OK with a 200mm lens but with the 16mm for example it almost removes the AF operation entirely. The lens rarely refocuses at all, because at f/16 it believes the image is sharp no matter what you do.

This is where the A77 scores over the NEX-7 - you focus wide open when using very small apertures, and this enables you to pick the exact plane of focus precisely - even if you use focus peaking. The errors in true focus point which result from focusing at small apertures, even regular settings like f/8, are significant with the NEX-7 especially and those who choose to work with manually opened and stopped down classic legacy lenses are bypassing this inherent flaw in the system.

David

Good observation, this is currently my way of focusing properly with the NEX 7, I'm doing manually the SLR process, focusing first wit with the larger aperture and after that going down.

something that is starting to annoying me is the camera don't want to go down more than 1/60 in A mode when you are filming. That makes me decide to almost always film in M mode.

Other stuff about the aperture, the Sony AF to Nex adapter move the Aperture like always returning from the larger to the correct position, it don't do a smooth transition when changing the aperture as if you are using a E-mount lens. This again, make me start shooting movie with a aperture manual adjustment adapter when I want to shoot film with Minolta AF mount.

David, In your review, tried to review the wide angle adapter for the 16mm 2.8. On the Nex 7, I have been able to get very interesting result (when you do some correction in Lightroom).

regards,

Frank

[David Kilpatrick]

I've no problems with the wide and fisheye adaptors on the NEX-7. Maybe it's just the subjects I shoot, but the issue of colour vignetting is not present, CA is well corrected in ACR7.1, and the sharpness is much better than I would have expected.

You can't use 1/30th with 50/60p video. I think you can use 1/30th if you set 24/25/30p video (on all cameras). 1/60th is the normal standard longest shutter speed for videos, I've had 1/20th out of a Canon at 24p but I think it's approximate.

David

[bakubo]

There is a way of testing this using focus peaking. If the peaking kicks in sooner, and appears stronger, with higher contrast set then CDAF may indeed work better - but the EVF will look horrible.

Here's the actual truth, because I know how CDAF works. Higher contrast if fed to the CDAF detection process will result in LESS ACCURATE focusing but FASTER lock. Lower contrast results in the opposite, more accurate focusing but it will take longer. Extremely low contrast may result in no ability to focus at all.

Thanks for that info. I was suspecting something like that. There is one guy who has shown a few series of some pretty surprising photos of fast running dogs done on different days in different locations using continuous/tracking focus with the Olympus E-M5 set to vivid mode. He has shown several sets and says vivid makes the continuous/tracking AF much better. Of course, he is not saying that it is as good as a PDAF DSLR, but just surprisingly good for CDAF. In this case, the improvement in keeping up with the fast running dogs overwhelms the disadvantage of slightly worse AF accuracy, I think.

With my Panasonic G3 and Olympus E-M5, since I shoot raw, I prefer to turn the contrast down and use other JPEG settings that give better DR in the EVF. Even doing that it seems that the AF for both cameras is quite fast and accurate, but I don't use continuous/tracking AF.

You can improve CDAF by using a wide aperture, the optimum which Sony apparently think works is f/3.5, because this is the default for video even if you have a faster lens. One of the big hidden issues with the NEX cameras is that you are always focusing at working aperture (my review of the NEX-7 will try to cover this when finished...). So if you want to shoot at f/16, you end up focusing at f/16 too. That works OK with a 200mm lens but with the 16mm for example it almost removes the AF operation entirely. The lens rarely refocuses at all, because at f/16 it believes the image is sharp no matter what you do.

This is where the A77 scores over the NEX-7 - you focus wide open when using very small apertures, and this enables you to pick the exact plane of focus precisely - even if you use focus peaking. The errors in true focus point which result from focusing at small apertures, even regular settings like f/8, are significant with the NEX-7 especially and those who choose to work with manually opened and stopped down classic legacy lenses are bypassing this inherent flaw in the system.

It is hard to understand why the NEX 7 would do that. Maybe a firmware update will change it. The E-M5 and G3 focus at the widest aperture no matter what the working aperture is set to.

[David Kilpatrick]

Henry, the NEX-7 opens up the aperture in single-shot mode (at least in some conditions - I'll need to test more). It does not open it up in continuous AF mode except for briefly when you first press the shutter. After that, it focuses at working aperture as the subject moves, until the shot is taken.

I'm also not sure right now whether the NEX-7 always opens the aperture to focus, for example if the light is very bright - the reason I say this is the supposed extra-short shooting lag of 1/50th of a second. I have seen how long it takes to do an open/close aperture for focusing indoors with the AF assist and I'm sure there is not the same lag outdoors.

Anyway, I am definitely wrong to say you always focus at working aperture - but the AF-C setting does put you in that position even indoors in dim light, and causes the lens to hunt around as a result.

I now have to advise against using AF-C indoors... and I will need to check the way the aperture works in sunlight. With AF-C set the effects are pretty awful, the 16mm jumps in and out of focus. I had noticed something similar with the Tamron 18-200mm on AF-S.

David

[bakubo]

My check of the G3 and E-M5 were very simple and not nearly as extensive as what you are doing. I just used S-AF, attached the Panasonic 20mm f1.7 pancake lens, set the aperture to f11, and then positioned the camera so that I could see into the lens. I could see that the aperture was wide open. I then half-pressed the shutter release to AF and saw that the aperture remained wide open. I then fully pressed the shutter release to take a photo and I could see that the aperture closed down for the exposure and then re-opened. I don't know how these 2 cameras behave in the various situations you mention though.

[David Kilpatrick]

The NEX parks its aperture closed fully every time it is switched off, I guess so that there is little risk of sun directly damaging the sensor. It opens up when you switch on, then closes to what you have last preset. S-AF then opens it briefly to focus and closes it, this is the 'wobble' and brief flicker in brightness you see on the screen every time you press the shutter - the lens goes in and out of focus and the image brightens and darkens then stabilises. But in C-AF it only does this once, and then tries to focus at working aperture.

BUT I was almost right - it does not open up fully in bright light. If you set f18 for example the lens seems to open up to around f8 to focus, in daylight - not fully. So focusing is definitely not as accurate as it might be.

David

[bakubo]

David, this post might be interesting to you since Timur is also doing a lot to understand how the CDAF works, except for the E-M5:

http://forums.dpreview.com/forums/read. ... e=41701999

He has posted lots of good stuff over time and I gather he just enjoys discovering all the details of how cameras work. He and the E-M5 dpreviewer got into a long discussion on the forum before the E-M5 review was posted nailing down how some things worked and the results ended up in the review, I think.

[Vidgamer]

Interesting. I don't often have a need for AF-c, but I wonder if it works?

[bakubo]

Interesting. I don't often have a need for AF-c, but I wonder if it works?

Which camera are you using?