[Urkeldaedalus]

Their comment could mean that the AF speed isn't where it should be, or it could be a last straw they are clinging to in order to give the nod to the Nikon D7100 over the K-3.

I've given up on DPReview as a reliable source of information after they failed to address the many obvious questions regarding their K-5 II vs D7000 scores. They scored the Nikon D7000 with better build quality than the Pentax K-5 II and much higher performance (despite its miniature buffer size compared to the K-5 II).

You may want to look at the inexplicable scores for the Nikon D7000 and the Pentax K-5 II yourself.

There was quite a discussion back then with DPReview staff involved, with the staff picking up on spelling mistakes and other small errors, correcting them in the review, but not properly responding to the fundamental questions about the inexplicable scoring. Based on my suggestion, DPReview at least added "built-in sensor shift shake reduction feature" as a "pro" in their summary list of the K-5 II review (imaging that not being there in the first place!) but ignored many other important questions and never explained their odd -- to say the least -- scoring.

If Imaging-Resource mentioned a lack of AF speed, I'd be listening. With DPReview it could be anything from dilettantism to unashamed brand favouritism.

I saw another funny example of this recently when the Olympus OMD EM-1 came out. They said something to the effect of "users of 4/3 glass on the E-5 will be disappointed with the AF performance of these lenses on the EM-1." They were soon cowed into admitting that this was not based on side by side testing, it was actually on the reviewer's recollection of shooting with an E-5.

There might be some institutional bias toward the big two at Dpreview, but I think more often than not with them their review mistakes are just laziness, or as you said dilettantism, which of course would just let that bias slip in even easier where it exists. Those random passive aggressive remarks do seem to be a trademark of some of their work: "Well it sure would have been a nice camera if..."

With all this said, I wouldn't be that surprised if the AF on the K-3 isn't much faster than its K-5 brethren. Probably a bit snappier and more responsive, but not a ton faster. There are physical limitations of the SDM motors for example, and I imagine if they were using a review body Dpreview was talking about the DC motor in the 18-135.

I have photographed sports (in terrible lighting conditions) with all of the Pentax bodies from the K20D through the K-5 (including the K-30, not including the K-50). Things have gotten much better since the K20D days, but the focus tracking based on color with the additional focus points will likely be the biggest boon for sports shooters, along with the big buffer for continuous spray and pray shooting. The camera getting confused and choosing the guy in the wrong jersey in a crowd has resulted in a lot of the missed shots I have had previously. It looks like they already grouped the AF points with this in mind, but I hope there are some options for grouping AF points like in the K-30.

I'm also excited to see if the multi-pattern AWB can help with the atrocious color balance issues you can get with high school sports lighting. I swear it sometimes looks like the players are playing in several different dimly lit gyms after I get home and look at the photos

[bwDraco]

Static AF responsiveness is probably about the same as the K-5 II under various lighting conditions. Dynamic AF performance has probably improved significantly, though. The K-3 is almost certainly better at reliably locking onto and tracking a fast-moving subject than the K-5 series. I'd like to see it compared with the likes of the EOS-1D X alongside the K-5 II to see just how much tracking performance has improved, and to see how well it fares against pro-grade C&N equipment.

Ricoh designed the new AF system with sports photographers in mind. It would be very surprising if there was no significant improvement here.

--DragonLord

[deadwolfbones]

One thing you need to remember about camera review sites: if they don't have hard test data (numbers) for certain metrics (like AF speed), they go on memory. They don't keep every camera ever just sitting around the office. They're loaners, and they go back when the review is done. They also don't spend nearly as much time with each camera as owners do, so owners pick up on things reviewers don't, given enough time.

Take it from a professional camera reviewer!

[monochrome]

One thing you need to remember about camera review sites: if they don't have hard test data (numbers) for certain metrics (like AF speed), they go on memory. They don't keep every camera ever just sitting around the office. They're loaners, and they go back when the review is done. They also don't spend nearly as much time with each camera as owners do, so owners pick up on things reviewers don't, given enough time.

Take it from a professional camera reviewer!

Fair enough - but it sure would be nice if camera reviewers (and I make no implications about your fine work) would so state in thier reviews.

[clackers]

Fair enough - but it sure would be nice if camera reviewers (and I make no implications about your fine work) would so state in thier reviews.

Professional reviewers of most consumer items - including cars and computers - have editorial deadlines. They may have a weekend with the object, a press release and Google to work with.

[monochrome]

Professional reviewers of most consumer items - including cars and computers - have editorial deadlines. They may have a weekend with the object, a press release and Google to work with.

Yeah, I get that. I object to a reviewer making a comparison from "memory" and not stating it was from memory. The implication in an otherwise-factual review is that the comparison is also an actual, factual comparison.

[solar1]

has barely used a pre-production camera with 0? firmware. Saying what he said at this point imo is assinine.
He knows quite well not to go there - no statements should be made until at least FW 1.0
Why he jumped on this is pretty curious.

FalconEye wrote: Anybody noticed this sentence in the DPR preview?

"Camera performance is generally very snappy, though we were hoping that the new AF system would be faster."

[deadwolfbones]

Yeah, I get that. I object to a reviewer making a comparison from "memory" and not stating it was from memory. The implication in an otherwise-factual review is that the comparison is also an actual, factual comparison.

Part of what you should expect from reviews is subjective impressions (alongside empirical data). The problem is that sometimes it bleeds over so you get subjective impressions of things that should be empirically provable, etc. That's either because the reviewer doesn't have the facilities to accurately test the thing in question, because of time limits, or because of laziness.

[Rondec]

I know this isn't about DPR, but in particular since they were bought by Amazon, it feels like their goal is to sell cameras and they definitely soft pedal any issues with Nikon and Canon cameras, while pounding the smaller players for tiny deficiencies (the K5 II "build issues" or, the lack of a dedicated video button). There is no perfect camera. All a review site can do is present cameras as they are, weaknesses and all, but when you start straining at gnats to create problems with certain brands, that is a problem.

[audiobomber]

I would love to see the Laboratoire FNAC and Chasseurs D'Image results. As far as I know, DPR has no way to test for auto-focus. They just comment based on their gut feel.

[falconeye]

I would love to see the Laboratoire FNAC

FNAC and field reports are the only sources I consider.

And for FNAC, the nearest distance with good focus is my only performance measure, inverted and manually corrected for outliers. It is the inverted nearest distance which measures magnification change speed where the AF starts to fail. A very sensitive number if only it were measured consistently.

[audiobomber]

FNAC and field reports are the only sources I consider.

And for FNAC, the nearest distance with good focus is my only performance measure, inverted and manually corrected for outliers. It is the inverted nearest distance which measures magnification change speed where the AF starts to fail. A very sensitive number if only it were measured consistently.

Sorry, I misremembered. I meant to say FNAC and Color Foto, not Chasseurs D'Image. You started a thread about some comparative AF tests they conducted:
https://www.pentaxforums.com/forums/pentax-news-rumors/154373-pentax-k-5-exce...ocus-test.html

[D4rknezz]

I am tired of lab testing autofocus , and have less need of a fast AF then an accurate AF system. One that is accurate in game. I know its accurate in lab - but the focus points are too big : it never is accurate for me in real life shooting. The red blob of focus confirmation confirms and is always off, because it confirmed focus lock on a whole face. Especially their wide angle lenses.
OMG. Trying to get a couple portrait with the couple off to the side and a really nice landscape view is ...well, impossible.

I hope this fixes it.

[kenyee]

I'm also excited to see if the multi-pattern AWB can help with the atrocious color balance issues you can get with high school sports lighting. I swear it sometimes looks like the players are playing in several different dimly lit gyms after I get home and look at the photos

That's because of their weird lighting (sodium vapor?). You have to set your shutter speed so it's a multiple of their frequency or something wacky like that...

[kenyee]

The tech sounds great but the example in the video is not a good one...

+1
Do Nikon's football player running through tackles test w/ long 2.8 telezooms and I'll be impressed