[runjmb]

I am a lifelong Pentax user. I don't think I have ever held a Canon or Nikon DSLR in my hands, until Sunday. I was at Target and picked up the intro level DSLR from Canon. While I didn't think much of the camera, the autofocus was great. I would describe it as Zippy! It was quick and silent. It made my K100D look like a tortoise.

Has the K20D improved its autofocus? I have been finding myself in some situations where having better AF would certainly benefit me. It almost made me think about bailing out on the SS Pentax. Then I remembered that I own over $1500 of Pentax gear.

[kthung]

I've never tried an SDM lens for pentax, but I was wondering if for everyone else, when autofocusing, does the camera go past the focus point slightly, backwards past the focus point again slightly and the lock in, essentially taking three turns to lock focus? This always happens, even if I manually prefocus on a point it will go forward slightly, back slightly then lock. Kind of annoying, but I've learned to adjust my shooting to it.

Are SDM lenses any different? I only have a K100DS so I'm also interested if the K20D is better as well. I am equally invested in Pentax as well so I don't think I'll be jumping ship. Besides I dislike the ergonomics of Canon bodies, and their midrange cameras only take CF cards. And Nikon glass is way too expensive. No constant aperture zooms under $1000.

[rparmar]

I was at Target and picked up the intro level DSLR from Canon. While I didn't think much of the camera, the autofocus was great. I would describe it as Zippy! It was quick and silent. It made my K100D look like a tortoise.

But was it accurate? Since you mention the tortoise, "slow and steady wins the race" is perhaps an apropos phrase.

[cpopham]

Has the K20D improved its autofocus? I have been finding myself in some situations where having better AF would certainly benefit me.

As an owner of the both a K20, and a K100, I can say that, yes, the K20 is definitely faster, and so it should be with its high voltage lithium battery.

Which brings me to the other thing - one day, on a whim I tried Lithium AAs (non-rechargable) in my K100, and was shocked to find the AF was noticably faster compared to the NiMH AAs I was previously using. This makes a lot of sense, actually, when you consider that the Lithium AAs are typically providing a voltage which is 0.3V higher, per battery...

I've never tried an SDM lens for pentax, but I was wondering if for everyone else, when autofocusing, does the camera go past the focus point slightly, backwards past the focus point again slightly and the lock in, essentially taking three turns to lock focus?

Are SDM lenses any different? I only have a K100DS so I'm also interested if the K20D is better as well.

I can only speak for the 16-50, but it can hunt in exactly the way you describe in very poor light. I've noticed it seems to happen much less, than I used to notice the kit lens doing. I presume this is mostly related to the larger maximum aperture, since I haven't noticed that a DA 40 hunts much on either camera.

But was it accurate? Since you mention the tortoise, "slow and steady wins the race" is perhaps an apropos phrase.

This isn't exactly a fair criticism, but I've been quietly thinking this about a friend's Canon (400D). It seems to focus very quickly, but some of the time when she shows me full sized pictures I see something isn't quite right with the focus. It's possible that it's user error, or just a really poor kit lens, I guess.... but I do wonder...

[smcclelland]

You won't find too much of a difference in the 350/400D/450D models and if there is it will be minor. The sweet spot with Canon AF is using USM Canon lenses or HSM Sigma lenses which greatly improves the focus accuracy/consistency comapred to non-USM motorized lenses. If you're a center point shooter it will all be the same really but if you're like me where you use the outter AF regions then you'll really notice a big jump going to the 30D/40D/50D models as they are 9-point AF cross-type sensors with the ultra-sensitive center point using an f2.8 lens or faster.

I've not used a Pentax K20D so I cannot vouch there and will leave that for the more experienced Pentax shooters but in the Canon camp you have to "pay to play" for AF performance. My 40D tends to hunt in extremely low light even with my 50 1.2 but my 1DMk3 locks on in an instant and holds focus nicely.

[Scottie137]

I have not had a chance to use the K20 outside the store, but its absolutely slower than my 40D (which I am selling to buy a K20...). I have posted in other threads, but I did find there to be accuracy issues (not major ones, but present) with my Canon. I mainly used either the 28-135 IS USM kit lens or the Canon 50 1.8, so I was not using high end lenses. This absolutely could have had something to do with it, but with my old K100 and the general purpose lenses I used (kit lens, 18-250, FA 50mm 1.4), I rarely was out of focus due to the AF system, though the k100 is dramatically slower than the 40D...

[cabstar]

I think the main reason why the Pentax seems slow is due to the AF confirmation. I use the 50-135mm DA* in photo pits at arenas & compared to my Tamron 16-50mm it locks on almost instantly, it very rarely hunts & I shoot under some really bad conditions without flash. The Tammy does hunt now & again but more so than the DA* lens.

When I have the 50-135mm on I would say 95% of my pics are in focus on my Tamron this goes down to approx 75%.

I prefer having focus confirmation & once used to it, it really does help improve your photos.

Links

DA*
Tamron

[UnknownVT]

I am a lifelong Pentax user. I don't think I have ever held a Canon or Nikon DSLR in my hands, until Sunday. I was at Target and picked up the intro level DSLR from Canon. While I didn't think much of the camera, the autofocus was great. I would describe it as Zippy! It was quick and silent. It made my K100D look like a tortoise.
Has the K20D improved its autofocus? I have been finding myself in some situations where having better AF would certainly benefit me. It almost made me think about bailing out on the SS Pentax. Then I remembered that I own over $1500 of Pentax gear.

Unfortunately it doesn't look as though the Pentax AF has been speeded up much since the K100D -

Popular Photography is one of the very few places where they actually test the AF speed -




Note: the horizontal (light level) axis on the K20D graph is misaligned -
it should be the same as the other graphs.

However having said that - almost any AF (unless it can't focus, hunting) is WAY faster than we can focus manually - so how did anyone managed to take fast action shots with manual focus lenses?

Pre-Focus on the spot where the action is likely to occur and trip the shutter when one sees the shot - then the lag is <0.1sec which is faster than most human reaction time.

[Ratmagiclady]

However having said that - almost any AF (unless it can't focus, hunting) is WAY faster than we can focus manually - so how did anyone managed to take fast action shots with manual focus lenses?


Doesn't matter how good the AF is, it can't anticipate like a human mind can. Very neat and a great tool, but AF can only react, or at best extrapolate, no matter *how* quickly it does it. With familiar MF lenses I can be as quick to drop into fine focus as you could be on an AF button: in fact, my hands can often drop me into decent focus before the camera's fully to my eye. (Age changes the equation, there, of course.) But 'how it was done,' wasn't even about trying to tell a computer what to do: trained hands and eyes, millions of years of instinct, you don't even think about it. It's like how the ball players catch or hit a ball, you don't sit and calculate, you just do.


That said, the K20d is pretty happy for me, faster than *my* eye than I could expect to focus an f5.6 lens in the dim, certainly. We'll see what I think when my 1.4 arrives.

[lol101]

But was it accurate? Since you mention the tortoise, "slow and steady wins the race" is perhaps an apropos phrase.

Unfortunately (and as far as my limited experience goes...) it's not the case.

When I did side by side testing between K10, K20, 30D and 5D, (mounted with their respective 50-135SDM or 70-200f4 L) the Canon weren't only evidently faster, they were also more consistantly precise.

To me it's THE perfectible point in today's Pentax DSLRs.

But don't get me wrong, it is not to the point where Pentax 's AF is "blown out of the water", "unusable" or other trendy internet stupidity I have seen, I would rate it as very good while Canon's one is extremely good (but still imperfect).

I couldn't keep both and Canon offered me some undeniable technical advantages plus a way to "affordable" FF so I kept the Canon and financed it with my Pentax collection but read this

The Online Photographer: Sony vs. Nikon vs. Canon

and you'll see that real world differences have shrink to paper thin between recent DSLRs FF or not, fast AF or not (look for the part when the K20 enters stage).

[Scottie137]

Canon weren't only evidently faster, they were also more consistantly precise.

See this is why forums are useful. I had a subjectively (never tested the camera directly against one another) different experience with my 40d. I found it to be less precise (although MUCH faster) when auto focusing. I am an admitted newbie (relatively) though and its possible I just wasn't using the Canon to its potential related to focus point selection, etc...

[lol101]

See this is why forums are useful. I had a subjectively (never tested the camera directly against one another) different experience with my 40d. I found it to be less precise (although MUCH faster) when auto focusing. I am an admitted newbie (relatively) though and its possible I just wasn't using the Canon to its potential related to focus point selection, etc...

But are they really?

It is quite possible that I had a miscalibrated K10 (K20 was better in almost every aspect but still not quite as reliable as the Canon) or exceptional Canons, or a subpar 50-135, or... , ..., too many possibilities and not enough samples one way or the other to draw general conclusions I'm affraid...

To me, the turning point was when I saw I was often getting sharper images out of the 30D+70-200L combo than from the K20+50-135... but it could be 90% due to the lenses that were used.

One way or the other, it seemed reasonable to pick the combo offering the best resulting IQ from what I could see (I finally added the FF variable to the equation since I do a lot of low light portraits).

However, the fact that I am still posting here should tell you that I will probably try the future Pentax offering and wouldn't hesitate to jump back once I find one that gets the AF and IQ to the level of what I see from my 5D+70-200L... plus all the much nicer "Pentax features" that we all come to love because they just make taking pictures a much more enjoyable moment...

[Mallee Boy]

Unfortunately it doesn't look as though the Pentax AF has been speeded up much since the K100D -

Popular Photography is one of the very few places where they actually test the AF speed -




Note: the horizontal (light level) axis on the K20D graph is misaligned -
it should be the same as the other graphs.

However having said that - almost any AF (unless it can't focus, hunting) is WAY faster than we can focus manually - so how did anyone managed to take fast action shots with manual focus lenses?

Pre-Focus on the spot where the action is likely to occur and trip the shutter when one sees the shot - then the lag is <0.1sec which is faster than most human reaction time.

I am not sure what these charts prove, other than the difference between the cameras is down to decimal points of a second in speed.....how many of us can pick that sort of difference in real life.?

I find with sport for example that I still prefer to set up, research the shot and shoot manually or pan with auto focus. It seems to work for me.

Cheers.

[pentaxmz]

I am not sure what these charts prove, other than the difference between the cameras is down to decimal points of a second in speed.....how many of us can pick that sort of difference in real life.?

I find with sport for example that I still prefer to set up, research the shot and shoot manually or pan with auto focus. It seems to work for me.

Cheers.

With all due respect, are you serious?

With a little more than a second of difference; that's a huge amount of difference. And these tests were done in a controlled studio setting. In real-world conditions, the difference could be even more.

But even one second could be the difference between getting the shot and not!

These graphs prove to me what I always suspected. I have always known that Canon and Nikon seemed superior in AF and my Pentax cameras.

Last night I discovered that it is quite possible that my eight year old MZ-S has better AF than my K20D, at least in a twilight setting. Where my K20D would hunt and give up, my MZ-S had no trouble finding a focus lock (using same lens). Now why is that?

[rparmar]

According to the charts a K20D is actually a lot better than the K100. Since I use the K100D, this is of interest to me. At a light level of zero it finds focus in about 1.1 instead of 1.8 seconds. And the best it does is 0.35 instead of 0.51s. For those who think 150ms makes no difference, think again. A target traveling at 40 km/h can move almost 2 metres in that time.

Accuracy means more than speed, though. And there's no telling what lenses the cameras were tested with, which would certainly change at least the low light scores.