Canon's Real sensitivity in AF

static808 wrote in post #7039957
completely agree with you on the real world shooting part. however, to me the initial concern wasnt about all of us having difficulties shooting blank walls... it was about canons published specs being better than the competition's, but real world use showing the contrary.

i apologize if my initial post seemed rude...no disrespect intended.

Again, not to carry this out, but what Im saying is that verifying their published specs by shooting a poorly lit blank wall doesnt seem very "real world" to me.

Thats all.

static808 wrote in post #7039957
completely agree with you on the real world shooting part. however, to me the initial concern wasnt about all of us having difficulties shooting blank walls... it was about canons published specs being better than the competition's, but real world use showing the contrary.

i apologize if my initial post seemed rude...no disrespect intended.

Where do I find Canon specifically claiming better low-light focusing than Nikon or Sony? Or that it can focus on a "lightly textured wall?" The manual specifically says that it can't focus well on low contrast subjects.

RDKirk wrote in post #7040262
Where do I find Canon specifically claiming better low-light focusing than Nikon or Sony?

Check the manufacturer's specs pages for Low light auto-focusing (usually rated in EV). Canon 5D is rated to EV -0.5.

TheHoff wrote in post #7040300
Check the manufacturer's specs pages for Low light auto-focusing (usually rated in EV). Canon 5D is rated to EV -0.5.

Nope, sorry, not the same thing. I worded my query very specifically.

Given a target with significant contrast, their system does, indeed, focus down to EV -0.5. Canon never said their system focused at lower light any anyone else's, nor did they ever say their system could focus on a "lightly textured wall."

If the Sony can focus on a "lightly textured wall" at any light level, it's doing better than I'd expect any system to do (and I know that Canon's has trouble with "lightly textured walls" under any circumstances.

Of course, I don't really know what "lightly textured wall" means to anyone else. Stucco? Popcorn? Concrete block?

RDKirk wrote in post #7040412
Given a target with significant contrast, their system does, indeed, focus down to EV -0.5. Canon never said their system focused at lower light any anyone else's, nor did they ever say their system could focus on a "lightly textured wall."

If the Sony can focus on a "lightly textured wall" at any light level, it's doing better than I'd expect any system to do (and I know that Canon's has trouble with "lightly textured walls" under any circumstances.

Of course, I don't really know what "lightly textured wall" means to anyone else. Stucco? Popcorn? Concrete block?

Yeah, it's two different issues I think. FWIW, Canon 1 Series and the Nikon D3/D700 are rated to -1 EV...no idea about the Sony. And I think most Canon/Nikon bodies will focus below the published specs...with enough contrast, but don't quote me on that.

But then the Canon system doesn't really work at any light level when there is little/no contrast. I actually had no idea that Nikon or Sony could do this, and I'm sort of surprised that they can. I've always had problems with backlighting and a low contrast subject w/ my 30D and 5D (the camera would want to jump to a higher contrast target), and I wonder how much, if any, the Nikon system is an improvement in this situation.

It would seem that the newer Nikon has an improvement, even over the 1DMKIII's abilities, which for the first time, on paper, match the Nikon specks, but are still not as sensitive in practice.

When I had my 1D as a new camera I could fail to AF on an office block's blank backwall in broad daylight whereas my Nikon D100 could. I've posted this numerous times, but I know few understood. This was not on the blank wall, but using center point focus and the office block just as part of the scene, not a close up of the wall.

I took my 1DMKIII to the same office block and it did three things, it did lock focus instantly and it did hunt and miss and hunt and lock in about equal proportions, so by that reckoning I'd estimate the 1DMKIII's bland sensitivity is about 2/3rds better, my 1D would fail and give up, my 1DMKII was the same ~ never tried my 1DMKIIn, I've said I would but didn't.

It is difficult for people to understand the differences, a lot of the Nikon AF was always better than Canon's ~ except for the AI Servo ~ up to the D300/D3 the Nikon sports shooter came back from any event with millions of perfectly focused audience pictures as Nikon's Continuous AF system ALWAYS locked on the background and would NOT reacquire AF without the photographer releasing the system and starting again ~ Canon's AI-Servo was like six years ahead of Nikon and in Space Age Technology while Nikon were in the stone age.

This applies to the AI-Servo equivalents ONLY [NB], otherwise the Canon system was blind and needed a white cane and the Nikon system had cat's eyes and could see in the dark.

Nikon, from the D300, has updated it's Continuous AF to the point that it is nearly as good as Canon's ~ Canon is still the best sports machine, but system wide it's still lacking against the improvements ~ Nikon's still superior single AF, near matching servo and vastly improved flash [IMO, get the AF right and the flash follows].

bacchanal wrote in post #7040557
, and I wonder how much, if any, the Nikon system is an improvement in this situation.

c

RDKirk wrote in post #7040412
Nope, sorry, not the same thing. I worded my query very specifically.

Oh I hear what you said, I was just too lazy to Google the other specs.

Canon claims the 5D2's "Autofocus working range" goes to: -0.5 EV

Sony claims for the a900: 0 EV

Nikon claims for the D700: -1 EV


While that may not be a Canon advertisement claiming "Our autofocus works in lower light than the competition" you can simply compare the specs as provided by the manufacturer since that is how they represent their cameras. Of course I realize this is a pointless exercise since "working range" is not a definition and it certainly is not a standard followed by the three manufacturers.

So to correct the LL review, Nikon claims their AF works in lower light, but Canon does claim the 5D2 should work in slightly lower light than the a900.

So, according to recent threads, Nikon's build quality, resolution, flash, and now even auto-focus is all better than Canon's. What is the world coming to?

XterraJohn wrote in post #7040941
What is the world coming to?

An end.

Long time reader here, but my firts reply.

The OP has turn off the AF assist lamp of the Sony? If I put the ST-E2 on my 1DMkIII I has no problem with the AF in any surface.

Welcome out of lurk mode.

And yea, the AF assist lamp would make a huge difference, you're right. The original article didn't mention that, or if they were using the off-center 5D2 points, both of which would handicap the Canon. Give them both AF-assist beams and use the center point and maybe the performance would be more equal. (Of course you have to strap $200 to the top of the 5D2 and limit yourself to one point, but hey...)

TheHoff wrote in post #7041096
Welcome out of lurk mode.

And yea, the AF assist lamp would make a huge difference, you're right. The original article didn't mention that, or if they were using the off-center 5D2 points, both of which would handicap the Canon. Give them both AF-assist beams and use the center point and maybe the performance would be more equal. (Of course you have to strap $200 to the top of the 5D2 and limit yourself to one point, but hey...)

Thanks for the welcome!!!

We have to weight in the isue of the assist lamps. The last Canon DSLR that use one was the D60.

I like to use Live View for blank featureless walls

OP, perhaps when you start shooting people, not walls, you'll see that the 5d and 1d are not bad at all even in very low light.

AF assists work, sure. Live View, that's a another one. If my camera can AF unaided on a featureless wall, then I know it will AF on almost anything.

One exceptional camera was the Olympus E10, it used invisible Infra-red and it could AF on any surface, even smooth AND, wait for it, in total darkness ~ no other camera could do this.

But it had two drawbacks, this was effective up to 15feet only and it suffered from warm lighting blindness. That meant that from hotel to hotel the AF performance would be different at distances greater than the 15Ft limit of the IR. But it was fantastic for my socials in night clubs.

My 5D or 1DMKIII with 580EXII and its assist beam on can struggle a wee bit in night clubs ~ total darkness and the distractions of flashing lights means I sometimes have to move people ~ don't like doing that as the whole point is catching a happy snap ~ but both do pretty well though. I like shooting people in the dark and dragging the shutter to pull in the lights which are flashing and not constantly on.

basroil wrote in post #7041188
I like to use Live View for blank featureless walls
.

GerBee wrote in post #7042407
AF assists work, sure. Live View, that's a another one. If my camera can AF unaided on a featureless wall, then I know it will AF on almost anything.

One exceptional camera was the Olympus E10, it used invisible Infra-red and it could AF on any surface, even smooth AND, wait for it, in total darkness ~ no other camera could do this.

But it had two drawbacks, this was effective up to 15feet only and it suffered from warm lighting blindness. That meant that from hotel to hotel the AF performance would be different at distances greater than the 15Ft limit of the IR. But it was fantastic for my socials in night clubs.

My 5D or 1DMKIII with 580EXII and its assist beam on can struggle a wee bit in night clubs ~ total darkness and the distractions of flashing lights means I sometimes have to move people ~ don't like doing that as the whole point is catching a happy snap ~ but both do pretty well though. I like shooting people in the dark and dragging the shutter to pull in the lights which are flashing and not constantly on.

Shot night life stuff with an XT and mkiii, 430ex both, and both had very good keeper rate (xt>50%, mkiii>80% and in tougher environments), what i suggest is putting the camera on servo and flip the switch to slave mode on the camera. Believe it or not, it'll still fire normally (ettl and all), but it will also blink the af illluminator even with ai servo. Moving people are less of a problem that way.