Which technique for manual focusing gives you the best and most reliable results?
I use a 5D with manual focusing lenses and the standard screen. First, I tried focusing through the viewfinder, but the results were far from reliable and many pictures were out of focus. Now I use focusing confirmation adapters for most lenses. Stopped down, this works well, but with fast lenses the camera indicates in-focus for a larger range than is actually in-focus. And the problem varies with the lenses, too. With the Zeiss 85 1.4 for example more images are in focus at 2.0 than with the Zeiss 100 2.0 at 2.0. I use TS-E lenses, too, and with manual focusing with the Canon 90 2.8 I get much better results.
So what gives the best results? Better focussing screens? Cameras with live view? Other cameras (1D?) with better viewfinders and/or more accurate focus confirmation? Focus bracketing?
Yes. A better focus screen will help. More practice will help too.
Live View obviously works great for static slow shots but not good enough for action shots. I have Ec-B on 1 series body and it works great. I have Ee-S screen with Hi-Lux treatment for the 5D and it works great too. The split screen is not good for actions shots but great for macro. 1D viewfinder looks same as 5D.
Thanks for the reply! Do you use focus confirmation with the 5D or are you focusing visually only? I've used an analog manual focus system for many years and the results were much better than now with the 5D. I owned a 1Ds before and my impression was, that more of the pictures I saw in the viewfinder as in-focus have been actually in-focus.
Leon Noel wrote:
Precision matte screen can get dark with slow lenses, and split prism screen tends to have problem with metering.
.
Yes. That's the case for stock Ee-S. However, it is not the case for the Hi-Lux treated version. The Hi-Lux treated looks more like stock Ee-A with better contrast.
Leon Noel wrote:
There's also eyepiece mods for 5D using Nikon DK-17M or DK21-M eyepieces.
.
Yes The DK-17M works great. It has higher magnification than the DK21-M.
The 2X viewfinder magnifier can only do center focus. Not as useful.
I use mechanical adapters so no AF confirmation. I do have AF confirmation adapters but they're less accurate than visual.
Sounds like you need to adjust your eye piece to match your vision. That's what the dial is for. Hope this helps
The 5D has manufacturing tolerance problems with the mirror box and the focusing screen tray. People have had problems with manual focusing, because the focusing screen may not sit in the exactly correct position. You may need to place shims under the focusing screen to correctly position the focusing screen in the image path.
The focus confirmation chip may not be accurate...mine isn't with my 135/2.8 mf nikkor...focus is correct at the point of rotation when the light goes OUT as I focus towards INF...run some tests on yours...mine is off but it is consistently off..weird but true, the LED is correct on my other lenses...oh well...
i shoot the 5D/40D combo and use the Ee-S screens. i have the original Canon FD-EOS adapter with a focus confirmation chip. when focusing the confirmation chip works pretty accurate but not dead on. there is play from when the chip confirms focus and when the focus point deactivates while continually focusing. i use the confirmation chip to give me an indication of when focus is in range and then tweak focusing to get dead on. same is true for an EOS EF lens in manual mode. with that said i find my focusing confirmation chip is most accurate when focusing from the near side (MFD) to infinity ie. i have more in focus keepers when i stop focusing just when the chip activates a focusing point when approaching critical focus from the minimum focusing distance end of the focusing range of a given lens.
when shooting BIF or any moving target i totally disregard the focusing confirmation chip and focusing points. i find that the in focus keeper rate increases when i use the entire view finder as a focusing aide. that is, instead of trying to keep the focusing point on the target and continually focusing to keep the focusing confirmation chip and focusing point active i keep the target in the view finder and keep that target in focus where ever it is located in the view finder and then pull the trigger. i have been shooting FD manual lenses 300mm f4l, 300mm f2.8l, 600mm f4l, and 800mm f5.6l, etc., on my 5D, 20D, & 40D continually for almost 3 years now and have been developing techniques to improve the in focus keeper rate. it can be quite frustrating to be shooting a burst only to discover on the computer that your most valued shots where not critically in focus.
live view is a God send for static manual focusing. can't wait for the 5D MkII with live view.
to test the accuracy of a focusing confirmation chip one needs to set up an optical bench to determine the focusing characteristics of the individual chip specifically to observe the focusing range vs. focus confirmation activation point. once focusing parameters have been ascertained focusing manually becomes some what easier. i will note that in my case i have one adapter and one confirmation chip to deal with. chip characterization can get quite laborious when dealing with multiple chip scenarios.
s23chang, i have 2 brand new Ee-S, 1 for my 5D, and 1 for my 40D waiting to have them Hi-Lux treated but i don't have a link to Hi-Lux. can you post a link. best regards and thank you in advance.
R.H. Johnson II
i ain't easy but it can be done using the afore mentioned techniques as demonstrated by this example image.
Canon EOS 5D iso 200 ~ Canon FD 800mm f5.6l ~ 1/1600 @ f8.0
Edited by R.H. Johnson on Mar 17, 2008 at 01:41 PM GMT
Focus is not an exact point. We may like it to be, but it's actually a range of acceptable sharpness. When using a focus confirmation adapter you'll often find a range where the green dot will light up. That range in our opinion usually starts/ends at front focus, to in focus, to back focus.
When focusing with an AF chip, I first see what the range of focus is (ie - where the light lights up). Then I decide if I want to bias the final focus towards the front or the rear of that range. If you shoot the same lens often enough, you'll begin to get a feel for which lens work best slightly biased towards the front or rear.
The next thing to consider is which focus points you're using and the lens' wide open MTF. As we know the lens' MTF drops off towards the edges/corners. If you're using an outer AF point, the AF system struggles if the lens has a low MTF. I can visually see this in the viewfinder with a Canon 50L F1.0.
Another variable is our physical lag time. When focusing the light goes on, did you rotate focus slightly past that point due to reaction time? You may need to rock the focus back and forth a bit if the DOF is really small.
Lastly, if you shooting with a really fast lens, avoid focus & recompose is possible. That can work if you're shooting with apertures of F5.6 or so (depends on subject distance).
If the image is particularily important, cover your bets and take 3 shots. One where you think focus should be and then 1 slightly front focused and another slightly rear focused. Memory are cards cheap, memories aren't - so don't be afraid to fill a card.
Ee-S
don't focus/recompose
don't trust AF confirm
get the 5D calibrated properly using an f1.4 lens
or swap it for a better one
use glints to minimise circles of confusion
bracket by constantly shooting while trying to perfect focus
don't get anal - slight off-focus isn't always a reason to condemn a good picture
'I've never had problems with a split circle focusing screen (1Ds and 1Ds2)'
been there done that, you will loose metering accuracy with the 5D and split circle focusing screens. the op has a 5D. the clarity of the split prism allows more light to the metering crkts. the metering inaccuracy is not linear it is exponential and changes with f-stop. some have speculated that because the split prism darkens with f-stop when stopping down that is the source of non-linear behavior, i tend to agree. there is no firmware custom function setting to compensate for a split circle focusing screen. i suspect the 40D has the same problem. the 20D however, uses a different metering system that does not pass through the screen and metering works dead on with a Hoada diagonal split circle focusing screen for me. there have been extensively detailed discussions on this forum on the topic of metering vs. focusing screens that contain a wealth of knowledge on this topic.
PSquared63 wrote:
The 5D has manufacturing tolerance problems with the mirror box and the focusing screen tray. People have had problems with manual focusing, because the focusing screen may not sit in the exactly correct position. You may need to place shims under the focusing screen to correctly position the focusing screen in the image path.
Hello,
This might come a bit late, I have some serious focusing issues with the EE-S screen that I installed yesterday. The focus is ever so slightly off when the lens is manually focused, it needs 1 or 2 degree turn (towards infinity) to the image to be tack sharp. I might have to build those shims. Anybody have info on this? =)
r.h. johnson wrote:
when shooting BIF or any moving target i totally disregard the focusing confirmation chip and focusing points. i find that the in focus keeper rate increases when i use the entire view finder as a focusing aide. that is, instead of trying to keep the focusing point on the target and continually focusing to keep the focusing confirmation chip and focusing point active i keep the target in the view finder and keep that target in focus where ever it is located in the view finder and then pull the trigger.
I agree completely, this is why I've chosen to use a camera that's optimized for manual focus.
Does anybody know of a good company (or individual) that will "tweak" a camera body or lens to the degree we are speaking? Like paying someone to optimize your camera for MF with specific alternative lens and adapters (just as an absurd example). Would be great to have confidence in some lab somewhere where you could pay someone to tell you really what part of this complex system is out of kilter and who could then do very fine tune-up.
DocsPics wrote:
Does anybody know of a good company (or individual) that will "tweak" a camera body or lens to the degree we are speaking? Like paying someone to optimize your camera for MF
You mean rip the AF system out & replace the mirror, prism and viewscreen?
I will more systematically try them out, especially exploring the behavior of the AF confirm.