Why are the autofocus sensors clustered around the middle of the frame?

[Daniel Browning]

I will try using live-view to achieve accurate manual focus, and exactly where I want it...

I think you'll be amazed by how much more accurate your focus is with 10X live view than any other method (autofocus, OVF, anything). The problem, right now, is the handheld ergonomics. Holding the camera out in front of your face to see the LCD like a neophyte digicam user is not very fun. That's why video shooters like me invest in a big shoulder-mount rig, which does great double duty for stills shooting with live view:



www.redrockmicro.com/redrock_dslr.html
www.chrosziel.com/...ata/chrosziel/media/doc/InfoPage-DSLR_english.pdf
www.cavision.com/pictures/5DMII/5DMII.htm
store.zacuto.com/DSLR-Gunstock-Shooter-Kits/
www.cinevate.com/catalog/index.php?cPath=34
www.thecinecity.com/tcc/home.php?cat=279

Aha, I had been wondering why with larger apertures the DOF in the viewfinder seemed to be much greater than in the actual picture -- is it because only a central part of the lens is used for the OVF via the mirror, thus simulating a smaller maximum aperture?

Actually, it's because of the design of the viewfinder screen. The standard screens don't scatter enough light to see accurate DOF past f/3.5 -- they're more on the side of an aerial image. There are optional screens available (and ground glass screens from third parties like Katz Eye) that scatter much more of the image, giving you far more accurate DOF. The downside is that the viewfinder becomes much darker. Some people find the Eg-S unusable at f/4 because it's too dark (personally, I find it usable. I even use f/5.6 by letting my eyes adjust).

Negative shutter lag?! [img]/emoticons/emotion-6.gif[/img] No thanks. I feel like the 40D is too fast already at 65ms shutter lag. [img]/emoticons/emotion-4.gif[/img]

I would *love* negative shutter lag. It would let me overcome my own delay. Say the camera takes 65ms, but my human reaction is so slow that I take 300 ms. The combined time is 365ms. If I set the shutter lag to negative (-365ms), then it cancels out my slow reaction time. I would have to measure how slow my reaction time is, of course.

[apersson850]

Given the geometry in the diagram, there does not seem to be any reason why the AF mirror and sensor assembly cannot be made wider--that is to say, it could conceivably be made to stretch lengthwise across the frame. This would not be reflected in the cross-sectional diagram. There must be other considerations as to why this is not done.

Agreed. I don't know the reason either.

That's because the secondary image forming optics in the phase detection AF system works with light rays that have passed through the lens at different positions. You get the best accuracy by bending both rays the same amount through the prisms. You can get away with not doing that, but not to any extent, which excludes points too far to the side of the image.


To get high-precision AF points to work, you need these rays to pass furhter apart, so that your base line can become longer. Thus you need lenses with larger max openings. But once you've reached the required opening, making it larger doesn't make anything better. Your AF system is already using the longer base line, and having more space beside it will not help you.

[photosurfer]

Hold on a moment. . .a negative shutter lag? Would that involve having the camera constantly shooting, then when you press the button it "reaches back" into the memory and chooses the image that it took 300 ms before you pressed the button? I believe that's the concept behind one feature on the Casio EX-F, is that the same idea?

[Daniel Browning]

Hold on a moment. . .a negative shutter lag? Would that involve having the camera constantly shooting, then when you press the button it "reaches back" into the memory and chooses the image that it took 300 ms before you pressed the button?

Yes.

I believe that's the concept behind one feature on the Casio EX-F, is that the same idea?
<div style="clear: both;"]</div>

Yep - same idea.

[photosurfer]

very cool stuff. . .thanks for the clarification.

[Daniel Browning]

You get the best accuracy by bending both rays the same amount through the prisms. You can get away with not doing that, but not to any extent, which excludes points too far to the side of the image.

Thanks for the explanation. I'm prepared to take your word for it, but according to my minimal level of understanding, it's not clear why it would be necessary to bend both rays by a different amount through the prisms in order to have points on the far sides of the image. As far as I can tell, it would be possible to retain the same angles by using larger secondary image forming lenses, larger autofocus sensors, and longer distance between the focusing plane and secondary lenses as well as autofocus sensors. My best guess was that the larger equipment necessary is too bulky, expensive, and/or difficult to calibrate to within tolerance. But again, I don't really know.

[Colin500]

I will try using live-view to achieve accurate manual focus, and exactly where I want it...

I think you'll be amazed by how much more accurate your focus is with 10X live view than any other method (autofocus, OVF, anything). The problem, right now, is the handheld ergonomics. Holding the camera out in front of your face to see the LCD like a neophyte digicam user is not very fun. That's why video shooters like me invest in a big shoulder-mount rig, which does great double duty for stills shooting with live view:

Tried it, works great!


(In this case it was with the EF-S 18-55 IS at 55mm without a tripod, AF just couldn't lock on the small tree that I wanted to focus on; however without a tripod, it would have been very hard to focus with 10x live-view without IS, even though there was plenty light to shoot without IS otherwise. I.o.w., IS can be useful for MF even when it's not needed because of the lighting conditions. Why is there no 24-70 IS?)


Regards, Colin