@smeggypants

I have a K20D and K-5. Just tried focus comparison between the two of them. Both were right on to about .5 EV and then both FF at 0 and below a bit. [so I did get the FF but with both of them and only at .5 EV and below]. I understand why you are sending in your K-5. Nowhere near what you are experiencing. And as you have reported, my K20D did a good job too, keeping up with the K-5.

I'm not sure what to make of this post.....

Firstly it makes perfect sense when installing a firmware update that makes changes to the AF system to make sure the calibration of the lenses is still correct. As it happened they were not.

"opportunity to bias my tests? Are you suggesting I'm deliberately posting tests to show a problem when there isn't one?


And what do yo mean by 'a little knowledge' ? Seems an uncalled for assumption on the level of my knowledge there don't yo think?

Hmmm??

I think you might be reading too much into what he's saying. The thing is, it's actually hard to directly compare 1.02 to 1.03 because of this recalibration step in between. One thing to do would be to not recalibrate and see what happens. Of course recalibrating kindof makes sense, but it's another place for error to creep in. Was the calibration perfect before? Is it perfect now? I think these different possibilities are what he means by "opportunities to bias." Biases are not all intentional.

I think the "little knowledge" comment may just be another reference to the whole issue of unintentional biases. Lots of scientific experiments go awry even when the people performing them are quite smart.

Nevertheless, it kinda seems to me like your k5 is not working properly.

OK ..... the background .... I know the Sigma 30/1.4 was accurate focussing on my K-5 on 1.02. For two reasons. I've been using it for shoots extensively and at 1.4 it needs to be accurate so I made sure I calibrated it properly.

New firmware appeared and had changes to the autofocussing system. NOT recalibrating or checking calibration would be the recipe to introduce error. In order to get the test as comparably accurate I had to start off with both cameras and both lenses focussing properly in normal daylight levels. This I did.


Yup that's certainly a consensus I agree with.

thanks for the tests smeggy, I'm going to run my own set of tests and see if its a K-5 issue or if its your camera's issue.

Thanks Werticus.

I look forward to the results.

I'm not sure what can be made out of these comparison photos, as the perspective/setup/distances seem to change with each iteration. You've got to keep variables like those consistent to prevent confounding.

Center focus point or auto?

It will be 6 hrs or so till its dark here

+1 that's exactly what i want to post too, the tests charts doesn't looks very accurate...
Why K5 test charts are not at the same angle than the K20 ?
Why the charts of the K5 are not centered the same way than the K20 ?

Isn't (at least) dubious that only Smeggy has done a serious test? Look at the topic about fw 1.0.3 and you'll see that they just chatter about the firmware but no one didn't attached a photo test.
I tend to believe Smeggy and Christine Tham and, of course, Falconeye, when he'll publish his study on fw 1.0.3. What I believe too, is the fact that Pentax just improved a little bit the focus, but didn't eliminate the ff.

If you want some phototests, watch this thread : Whose FF issue have been fixed with the K-5 1.03 Firmware ? [Page 1]: Pentax SLR Talk Forum: Digital Photography Review

Edit: oh, and a good test should be to compare the same K5 with a 1.02 FW and a 1.03 FW, not a K20 against a K5...

I saw already those "tests" on dpreview. Better Smeggy's rapid tests using AF charts. And, by the way, I'm interested about K5 tests not KxxD, so I can separate them.

I don't know about that. First of all, there is no doubt but that Smeggy pants has a defective camera, but I find it is better to take photos of newspaper on a flat surface, rather than a forty five degree angle photo of a focus chart (particularly with a wide angle lens). Falconeye himself suggests that this is more useful. Secondly, Peter (Asahiflex) has posted several real life before and after photos with the firmware that show marked improvement. These type of photos are much more like what I would shoot in real life and therefore more useful to me.

Pentax Photography & Camera Forum - PentaxForums.com

Hmmm, I doubt it. Not just Smeggy's camera, but JayBee's, Christine Tham's and who knows how many other. If Pentax launches so many defective cameras on the market, it is a big problem, too.

Another big problem is that the good impressions (about the camera) are of ones who had good impression about the anyway and were not bothered by ff.