[mrmentera]

Dont know if this is exactly the right forum for this subject but there is a lot of people on it and a lot of testing and AF-adjust so I chosed it. I´ve tried to do some focusadjust earlier but I always come with different results from time to time, trying different charts and patterns on the screen. I´ve never got this confident feeling as others on the forum seems to get when they test and adjust. Now when the new firmware came I thought the time was right to make an effort. I want also to put som focus on relability and other faktors of importence when doing an AF-adjust.
The mekanic stuff in the lenses and the screwdrive mekanism for example. I hear that my lenses behave diffrently when focusing. The FA50 has av very rough sound and gives a feeling of much bigger "cogs on the wheels" than my FA77 and Tamron 17-50. The easiest lens to get a result that can be reproduced is the FA77, then the Tamron and the hardest to get it with is the FA50.

I worked with a testchart shown below and som batteries. First I had black batteries and got one result, then I took silvery batteries and got another result and both differed from the charts result. With my Tamron I got some FF when I between every shot focused on something behind the batteries and I got some BF when I between every shot focused on on something closer. It seem to have bad brakes. Piuuh!

I have not written down all the numbers because I wanted to find severel testdesign that had some interreliability first, but I dont come to that point. I just take a lot of shoots, first on the middle battery, the same kind of battery on the left and then on the right, adjust, go to the black batteries, adjust, go to the chart, adjust. I dont get a result that is consistent betwen theese tests.

How can everyone else on the forum get so consistent and reprodusable data?
Is my K-5 (with 1.3) totally ****ed up? Are my lenses ****ed up? Or is it me???
Or is there a lot of testing on the forum that actually isn´t reliable in all aspects and may cause some confuse?
It would be nice to have some feedback on the aspects of the lenses mechanical construction that affect how AF behaves. Thanks.

Attached Images

   

---

Last edited by mrmentera; 03-11-2011 at 08:53 AM.

[Snydly]

Ok, been doing a lot of tests ... I will say there is some improvement, but the AF performance on my K-5 is still way behind the K20D. I'll post pictures tomorrow sometimes as it's 3:30an here, but here's the test I just did ....

K20D - Sigma 30/1.4 - exposure f1.4 1/60s ISO160 [apox EV6.5] tungsten light.

AF test chart focus spot on!

K-5 - Sigma 30/1.4 - exposure f1.4 1/60s ISO160 [apox EV6.5] tungsten light.

AF test chart focus spot on!

proves the above the two cameras perform equally well with the same lens and same settings at EV6.5


now taking the light down to EV2 ...

K20D - Sigma 30/1.4 - exposure f1.4 1/60s ISO3200 [approx EV2]

object in room - AF spot on

K-5 - Sigma 30/1.4 - exposure f1.4 1/60s ISO3200 [approx EV2]

object in room - phase AF way front focusses. contrast AF nails it.


---------------------------------------


Current conclusion: My K20D focus works perfectly at EV2 in all situations i've tried. the K-5 AF has worked in some situations at EV2 but there's many situations including the test above where it has failed miserably.

A lot more testing required.

Sorry to go off topic slightly here, but I've asked this question before and although many people tried to explain, I still didn't get it.

How do you determine and quantify EV? Do you have a light meter with an EV readout? I see people give an EV number and I would like to understand EV and how it's calculated or measured.

Thanks,

Bill

[Smeggypants]

Sorry to go off topic slightly here, but I've asked this question before and although many people tried to explain, I still didn't get it.

How do you determine and quantify EV? Do you have a light meter with an EV readout? I see people give an EV number and I would like to understand EV and how it's calculated or measured.

Thanks,

Bill

I use this ...


Ultimate Exposure Computer


There is a formula but I don't have excel installed and I've yet to find an online calc

[dcmsox2004]

some info about ev.... i'm learning about something i'd never quantified/qualified before... never had a problem with my k-x....hope this helps... dave m

Exposure value - Wikipedia, the free encyclopedia

reading further will talk about light meters...
Determining Exposure

[ursamajor]

I use this...

Use this.

[bxf]

...
How can everyone else on the forum get so consistent and reprodusable data?
Is my K-5 (with 1.3) totally ****ed up? Are my lenses ****ed up? Or is it me???
Or is there a lot of testing on the forum that actually isn´t reliable in all aspects and may cause some confuse?
It would be nice to have some feedback on the aspects of the lenses mechanical construction that affect how AF behaves. Thanks.

I've been asking myself the same questions, because I also obtain inconsistent results - consistently The one thing that I have observed up to now (preliminary observation only), is that my Pentax kit lenses seem to behave better than my Tamron 28-75mm and Sigma 50mm f.4 (which won't even work in Liveview). This statement is not to be taken at face value, since I haven't performed any meaningful tests yet, and perhaps a good AF adjustment session will cure some of my problems.

[Smeggypants]

Use this.

unfortunately it won't allow teh dialling in of values, only preset full steps. for example 1/5s ISO320

If anyone knows of an online calc where you can actually type in any value I would be most appreciative

[MrPetkus]

Smeggy - I've read your posts here and elsewhere within the forum. Why don't you exchange your K5? I've seen a dramatic increase in AF reliability with 1.3 and suggest an exchange as your best course of action. Unless you've passed the return window...

[ursamajor]

Hmmm, Smeggy, you must use your own brain, my friend, just a little.

[ASa]

Smeggy, I suggest you put it bold in your profile: I'm about to get my K-5 replaced by a new unit. So people will realize you're actually going to do that..

[Smeggypants]

Hmmm, Smeggy, you must use your own brain, my friend, just a little.

Interesting

Well the old withered cranium is quite capable of extrapolating between fixed values on those calcs and is a computation I already do on this

The other operation my old withered cranium is capable of is writing my own online EV calculator in PHP and hosting it on a page at my own forum.

In fact I may just do this in the near future, but currently don't have time.

[dadipentak]

Wait a minute! I'm a big-time K-5 fanboy but I don't think there's any value served by taking gratuitous pot-shots at Smeggy.

[MrPetkus]

I, for one, wasn't taking any pot-shots. If his exchange window hasn't been breached, I would suggest a replacement as a remedy to comparison headaches.

Wait a minute! I'm a big-time K-5 fanboy but I don't think there's any value served by taking gratuitous pot-shots at Smeggy.

[dlacouture]

How can everyone else on the forum get so consistent and reprodusable data?

Well, here are my thoughts on your problems:
- batteries are not good targets (specially when grouped like you did), as 1) they are rounded, and 2) you don't know what battery the AF module (whose size is far greater than the small red square) will pick as a target...
- keeping the background "open" can lead to another set of problem, as your camera can pick another different light source (say, a window in a tungsten-lit room, or a fluo lamp) and apply the wrong AF adjustment (seeing how your background is overexposed, this is a possibility).

Now you can take the following things into considerations:
- using focus test charts will get you the "perfect" AF fine-tune, but could still produce a slight FF if you always shoot faces in low-light situations.
- focus tests are "prefect" targets, where the contrast is high. Real-life targets will restrict you to higher Ev, as they often have less contrast.

[Christine Tham]

Wait a minute! I'm a big-time K-5 fanboy but I don't think there's any value served by taking gratuitous pot-shots at Smeggy.

I agree.

I haven't read any of SmeggyPant's posts, but if he is saying he still encounters FF with the new firmware, I believe him.

I am also finding with my camera that the FF problem (which on my camera occurs around EV 0-2) is still there, but reduced in magnitude.