Canon's Real sensitivity in AF

[QUOTE]One surprising test result was uncovered while doing some low light noise tests. I found that at very low light levels (ISO 1600 / f/2.8 / 1/20 sec) the 5DMKII lost its ability to focus on a lightly textured but blank white wall, while the A900 had no difficulty at all, locking on instantly. The Canon would simply hunt. This is contrary to the specs, which show the Canon as having better low light sensitivity. Once again, when it comes to real world vs published specs, I always go with the real world results. [/qiote]

from: http://www.luminous-landscape.com …cameras/a900-5dmkii.shtml

I've complained for years about this phenomenon. Way back I posted about my new 1D being unable to focus on an office block in daylight while my D100 could, similarly my D100 could focus on a bland wall ~ even my 1DMKIII still has difficulty, even if it is much, much better.

So, is the actual ISO selected a factor in getting sharply focussed pics?

I don't know if I'd go with focusing on a blank white wall as a real world focusing example at all. I've not got enough experience with the variety of Canon bodies or Sony bodies to make a fair judgement about AF, but I've never had problems with my 40D's AF. It's managed to lock focus with an F/4 lens in extremely dark places through glass and water without ever having a problem. As for ISO, it will have no impact on focusing, as the camera's ISO is it's sensitivity to light, which is achieved by amplifying the output from the CMOS sensor. The focusing is done by an entirely separate mechanism that doesn't use the CMOS sensor at all - phase ttl focusing uses the lens and the AF sensor. See more here - http://en.wikipedia.or​g …Autofocus#Phase​_detection

However, contrast detect AF may be affected by ISO, as the camera uses the main CMOS sensor to autofocus with contrast detect. If the ISO is too low, the sensor wont have sufficient contrast to be able to determine focus or not. I would however imagine that using Contrast Detect the camera employs some method of automatically upping its ISO accordingly, but I could be wrong.

[QUOTE=GerBee;7022577]

One surprising test result was uncovered while doing some low light noise tests. I found that at very low light levels (ISO 1600 / f/2.8 / 1/20 sec) the 5DMKII lost its ability to focus on a lightly textured but blank white wall, while the A900 had no difficulty at all, locking on instantly. The Canon would simply hunt. This is contrary to the specs, which show the Canon as having better low light sensitivity. Once again, when it comes to real world vs published specs, I always go with the real world results. [/qiote]

from: http://www.luminous-landscape.com …cameras/a900-5dmkii.shtml

I've complained for years about this phenomenon. Way back I posted about my new 1D being unable to focus on an office block in daylight while my D100 could, similarly my D100 could focus on a bland wall ~ even my 1DMKIII still has difficulty, even if it is much, much better.

So, is the actual ISO selected a factor in getting sharply focussed pics?

D100 has a built in AF assist bulb last i knew... Try the mkiii with an external EX flash and you'll see it focuses faster than the D100 again

I've long turned off the AF assist on the D100 ~ it's not a question of AF speed ~ but Nikon, since the F90 [film] has had a reputation for being able to focus on plain bland surfaces.

This feat has yet to be matched by anything Canon has ~~ the reference to the 5D2 suffering from this problem suddenly as the ISO went up is intriguing. It's something, I had never taken cognisance of ~ was my ISO selection effecting the AF ability of my Canon cameras?

I don't know and had not thought it could be a possibility, but we know that ISO selection with the Canon does effect noise with a rise in noise between each segment and not a linear noise increase, thus ISO640 can be worse than ISO800 ~ one would not have expected that either.

[QUOTE=basroil;7023374​]

GerBee wrote in post #7022577
D100 has a built in AF assist bulb last i knew... Try the mkiii with an external EX flash and you'll see it focuses faster than the D100 again

Im actually in pain from laughing so hard!

I guess if you want to take pictures of featureless walls, Nikon is obviously the way you should go!

AdamLewis wrote in post #7039492
Im actually in pain from laughing so hard!

I guess if you want to take pictures of featureless walls, Nikon is obviously the way you should go!

So you're saying you've never had trouble locking on to a low contrast subject?

GerBee wrote in post #7038727
This feat has yet to be matched by anything Canon has ~~ the reference to the 5D2 suffering from this problem suddenly as the ISO went up is intriguing. It's something, I had never taken cognisance of ~ was my ISO selection effecting the AF ability of my Canon cameras?

I think you're misunderstanding the cause and effect. The result is the autofocus hunting -- the cause is not the ISO setting, the cause is the low light environment. The ISO setting is simply reflective of the low-light conditions... so he was saying "as your ISO goes up (and the scene gets darker), the AF will hunt more" but it has nothing to do with what you set the ISO at.

bacchanal wrote in post #7039531
So you're saying you've never had trouble locking on to a low contrast subject?

Is that what I said?

OK, I'll have to re-read the article ~ I got the impression he just changed the ISO for the test, and not that it was darker ~ but maybe it was darker.

Yes, I think now I took the wrong interpretation from the piece.

TheHoff wrote in post #7039533
I think you're misunderstanding the cause and effect. The result is the autofocus hunting -- the cause is not the ISO setting, the cause is the low light environment. The ISO setting is simply reflective of the low-light conditions... so he was saying "as your ISO goes up (and the scene gets darker), the AF will hunt more" but it has nothing to do with what you set the ISO at.

BTW, I agree with him. Canons do seem to hunt in low-light with relatively featureless targets. I haven't used in AF Nikon in 10 or 12 years, though, so I can't compare.

One point to consider, though, is that the reviewer did not mention which focus point of the 5D2 he was using. As we all know, the center is markedly better than the rest.

OK, a story. Long, long time ago, in a land that time forgot, there was a poor photographer trying to earn a crust using OM gear.

Here he was in front of the Bishop at Confirmation time and trying to focus and frame as fast as possible so he can get more clients in.

One day a big bad wolf enters the Cathedral, armed with the newest Nikon 8008 he was able to P&S and flash in a millisecond, he soon had queues and the poor photographer just got poorer.

Next season, after praying to St Anthony his bank manager funded a revolutionary new machine, the Nikon F90s sporting a fly on a lump of coal in a coal cellar AF system ... actually I think it was the black cat ~ my memory fades ...

But woe betide the bad wolf and his obsolete 8008s, the F90s could AF on the child's dress, the Bishop's scarlet robes or the Priest's whites ~ all the poor photographer had to was compose and fire, compose and fire, compose and fire, soon the entire Cathedral was queueing behind him and the money was stacking up so much that he was no longer poor.

But what had changed, asks a puzzled child ~ no it was not the touch of the master's hand ~ it was an AF ssytem that could accuratly AF on most bland and featurless items, it did have a problem with one shcool uniform, Petrol Blue ~ just Petrol Blue ~ it would hunt, so we avoided that shool ...

AdamLewis wrote in post #7039492
I guess if you want to take pictures of featureless walls, Nikon is obviously the way you should go!

GerBee wrote in post #7039786
OK, a story. Long, long time ago, in a land that time forgot, there was a poor photographer trying to earn a crust using OM gear.

Here he was in front of the Bishop at Confirmation time and trying to focus and frame as fast as possible so he can get more clients in.

One day a big bad wolf enters the Cathedral, armed with the newest Nikon 8008 he was able to P&S and flash in a millisecond, he soon had queues and the poor photographer just got poorer.

Next season, after praying to St Anthony his bank manager funded a revolutionary new machine, the Nikon F90s sporting a fly on a lump of coal in a coal cellar AF system ... actually I think it was the black cat ~ my memory fades ...

But woe betide the bad wolf and his obsolete 8008s, the F90s could AF on the child's dress, the Bishop's scarlet robes or the Priest's whites ~ all the poor photographer had to was compose and fire, compose and fire, compose and fire, soon the entire Cathedral was queueing behind him and the money was stacking up so much that he was no longer poor.

But what had changed, asks a puzzled child ~ no it was not the touch of the master's hand ~ it was an AF ssytem that could accuratly AF on most bland and featurless items, it did have a problem with one shcool uniform, Petrol Blue ~ just Petrol Blue ~ it would hunt, so we avoided that shool ...

Ok. Now I hurt worse. Thats a wonderful story!

AdamLewis wrote in post #7039564
Is that what I said?

thats what you implied. if someone stated an issue with something and i started "laughing so hard that i was in pain", it pretty much means i have never experienced the same issue. if i share the concern, and i started laughing, then that would have been just plain ignorant and inconsiderate.

so which category do you fall in??

static808 wrote in post #7039850
thats what you implied. if someone stated an issue with something and i started "laughing so hard that i was in pain", it pretty much means i have never experienced the same issue. if i share the concern, and i started laughing, then that would have been just plain ignorant and inconsiderate.

so which category do you fall in??

Neither.

Im laughing at people testing AF on mostly blank walls in poor lighting. To me, its as pointless as people spending hours toiling over a ruler checking focus.

Its great that you want to use "real world" examples, but I find testing focusing ability against a poorly lit featureless wall. If you want "real world" tests, why not go use it in the "real world"!

Just my 1 cent.

AdamLewis wrote in post #7039911
Neither.

Im laughing at people testing AF on mostly blank walls in poor lighting. To me, its as pointless as people spending hours toiling over a ruler checking focus.

Its great that you want to use "real world" examples, but I find testing focusing ability against a poorly lit featureless wall. If you want "real world" tests, why not go use it in the "real world"!

Just my 1 cent.

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.