[areidjr]

Exifs was stripped by flickr in smaller size; original size has exif : AF-k5-test - a set on Flickr
No "maybe is" - all is ok (not only in chart tests).

As I said, if you're happy, all is good. Your incandescent test appears to have been shot around 6-7 EV, which is above the level where the FF issues show up.

I'm not saying that your K-5 isn't perfect, just that the examples you posted don't rule out the low tungsten FF issue.

Be happy and shoot away.

[Class A]

No "maybe is" - all is ok (not only in chart tests).

Unless I missed one, none of your images indicates a <4EV scenario, which is necessary to provoke the error.

You can calculate the EV value for yourself:

EV = log2 (f-ratio^2/shutter-speed) - log2 (ISO-setting/100) - EV-compensation

I'm not sure about the requirement about the image brightness value required to make the above formula an accurate indication of the scene brightness. I'm assuming that camera settings are only a good indicator for the scene illumination if the image brightness at the relevant spot is 18% gray (or whatever the real camera exposure target is). I haven't checked this, just plugged in some of your EXIF data into the above formula.

The good news: For your low-light scenarios your K-5 works.
The bad news: Your tests may not have any relevance for the <4EV FF problem.

[jolepp]

With my K-5 EV 4 would seem to be close to the threshold: EV 4 in tungsten light results in ok phase AF, @EV 3 FF is evident with a 1:2.8 lens. The max. aperture should have a direct bearing to the absolute light intensity at which this becomes a problem as the amount of light available to the AF sensor depends on this directly; I would expect e.g. a 1:2 lens to work at EV 3 and fail at EV 2 with my K-5. Then again AF FF/BF behavior in general is specific to a particular copy of a lens, so actual results might differ, but this would seem like the most reasonable prediction as such.

It is possible that some copies of the K-5 do not suffer from tungsten/low light FF at all, but I cannot recall posted test photos supporting this (?), I suppose it is fair to say that the assertion that there are K-5 bodies that do front focus significantly in low tungsten light has been so proven (see: https://www.pentaxforums.com/forums/pentax-k-5-forum/129738-k-5-user-auto-foc...t-results.html).

[Smeggypants]

Well, I try K-5 with my lens too. In my opinion all is OK:

DAYLIGHT:
1. K-5 FA77 daylight


2. K-5 FA43 daylight


3. K-5 DA*50-135 daylight (at 135)


4. K-5 DA*200


5. K-5 DA*16-50 daylight (at 50)


INCANDESCENT LIGHT (some photos, all OK)
1. K-5 FA77 night


2. K-5 DA50-135 night (at 135)


CANDLE LIGHT

1. K-5 FA77 Candle


2. DA*50-135 Candle (at 135)

Your pics are taken in too much light EV4 and above.

try at EV2

you can calculate exposure for EV here

Ultimate Exposure Computer

[jackseh]

Yep, I'm asking too much for a camera to AF correctly and come without globs of glue on the sensor. But what the hell, what's $1,500 amongst Pentax and its loyal customers.

To me, that is unacceptable for a $1500 product. However, all indications show that Pentax has worked diligently with its customers who reported faulty K-5s to ensure that their problems are resolved. The level of customer service speaks for itself. I've read several of your threads that were started and closed. Are you intentionally being acrimonious ? Whether you are or not, that's certainly not the way Pentax treat its customers.

[simico]

It is possible that some copies of the K-5 do not suffer from tungsten/low light FF at all, but I cannot recall posted test photos supporting this (?)

https://www.pentaxforums.com/forums/pentax-k-5-forum/128562-relevant-news-re-...ml#post1356037

I've done hundreds of "real" (not test) photos near and below EV4 since then. But I guess some people just don't want to believe something if it doesn't fit their theory...

[Tommot1965]

With my K-5 EV 4 would seem to be close to the threshold: EV 4 in tungsten light results in ok phase AF, @EV 3 FF is evident with a 1:2.8 lens. The max. aperture should have a direct bearing to the absolute light intensity at which this becomes a problem as the amount of light available to the AF sensor depends on this directly; I would expect e.g. a 1:2 lens to work at EV 3 and fail at EV 2 with my K-5. Then again AF FF/BF behavior in general is specific to a particular copy of a lens, so actual results might differ, but this would seem like the most reasonable prediction as such.

It is possible that some copies of the K-5 do not suffer from tungsten/low light FF at all, but I cannot recall posted test photos supporting this (?), I suppose it is fair to say that the assertion that there are K-5 bodies that do front focus significantly in low tungsten light has been so proven (see: https://www.pentaxforums.com/forums/pentax-k-5-forum/129738-k-5-user-auto-foc...t-results.html).

I doubt that very much....but Im sure some are better than others..and certain lenses and focal lengths show it more that others to,

so get a reasonable functioning K5 in low light, coupled with a lens that also good, then the problems are not as bad as other users......but all K5's, im sure suffer FF in low light to some degree.......of that Im convinced... whether that's a huge problem for the customer, only the individual can answer that .

what Ticks me off..is post after post of people complaining that a $1500 cameras doesn't do this or that...well it don't and if that's a massive problem..that get shot of it.or dont buy it till its sorted ..Crikey mate


Im interested to see Falk's white-paper....he promised it this week ??

none of this rant is directed at you Jolepp, just quoting your thought is all

[jolepp]

https://www.pentaxforums.com/forums/pentax-k-5-forum/128562-relevant-news-re-...ml#post1356037

I've done hundreds of "real" (not test) photos near and below EV4 since then. But I guess some people just don't want to believe something if it doesn't fit their theory...

Thanks for pointing this out. However the Sigma 1:1.4 is 2 stops faster than 1:2.8 and so EV 2.3 with that would mean an amount of light for the AF sensor to work corresponding to EV 4.3 for a 1:2.8 lens where my K-5 seems to work fine. Trying the Sigma around EV 1 in tungsten light would seem like a way of reproducing the problem - or not .

[Class A]

Thanks for pointing this out. However the Sigma 1:1.4 is 2 stops faster than 1:2.8 and so EV 2.3 with that would mean an amount of light for the AF sensor to work with corresponding to EV 4.3 for a 1:2.8 lens where my K-5 seems to work fine.

One may need to bear in mind that the AF module may not be able to capitalise on the difference between f/1.4 and f/2.8. It probably has its own lower limit of a widest aperture, in the same vein as the focusing screen makes all differences between f/2.8 irrelevant.

[filorp]

Guys is anyone of you able to check whether k5 will front focus at something like 10EV but with strong red/orange light (limited light spectrum) or the problem depends just on light intensity?

[jolepp]

Guys is anyone of you able to check whether k5 will front focus at something like 10EV but with strong red/orange light (limited light spectrum) or the problem depends just on light intensity?

I have not seen that kind of thing posted, however, in my case there is a difference at low (< EV 4) light level and tungsten / CFL / LED (see here) it would seem that it is about light intensity and source (IR?).

[dlacouture]

Strange that nobody remarked on my post about Neon lighting and back-focus...
Using two neon strips (for guitar lighting!), at close range (thus meaning quite high light levels), I ended with a consistent 15cm back-focus on my perfectly daylight-tuned K5 (with a FA50/1.4 at around 2m)...

For the record, my K5 suffers sometimes from a slight FF from EV 4 downwards.

[jolepp]

One may need to bear in mind that the AF module may not be able to capitalise on the difference between f/1.4 and f/2.8. It probably has its own lower limit of a widest aperture, in the same vein as the focusing screen makes all differences between f/2.8 irrelevant.

Good point. I was just going with the simple idea that bigger aperture = more incoming light = more of that (higher intensity) on the AF sensor(s) = problem (if any) shows up at a lower absolute light intensity (at the front element).

In general it is hard to keep track of the posts. With this tungsten-low-light-AF-FF issue having (at least a (strictly) 'relevant news' thread (like the current one was meant to be), a 'post your tests' one (which this was meant to be) and a general discussion not falling under the former (which just about any thread related to the issue tends to become might help. This opinion has the benefit of hindsight, of course .

[SteveB]

There's quite a few 'tungsten' AF problem posts on the D7000 building up on DPR's D7000 forum.............. Does ANY D7000 focus properly in tungsten?: Nikon D90 - D40 / D7000 - D3000 Forum: Digital Photography Review

[philbaum]

There's quite a few 'tungsten' AF problem posts on the D7000 building up on DPR's D7000 forum.............. Does ANY D7000 focus properly in tungsten?: Nikon D90 - D40 / D7000 - D3000 Forum: Digital Photography Review

SteveB,
Thanks for bringing this subject up. I've mentioned it once or twice, but folks don't seem to think its related. Coupla days ago, i shot live indoor theatre dress rehearsals with 3 others, one had K20 like me, one had D7000 and the other D90.
A. Everytime i looked at the Nikon shooters, the focus assist lights were on for each shot, and its a WHITE light, not the GREEN light chosen for frequency purposes by the K5 engineers. (by the way, the K20 cameras we had did not use any assist light and we didn't seem to have any problem focusing - the lights weren't as dim as some previous plays)

I think this announcement usually doesn't get more attention because the D7000 owners most often talk about backfocusing, but not always. While the K5 owners talk about front focusing entirely. Perhaps thats due to the effect of the white focus assist light in low artificial light environments, i don't know.

What's similar is that the front focusing or back focusing is beyond the capability of the fine AF adjust tools, like -19 or more (on Nikon). Sure, cameras have long been subject to Tungsten focusing effects, whats different is the magnitude of the errors - i think - perhaps someone has a different perspective.

But the focus outrage among new owners seems to be an order of magnitude above the ordinary level of complaints on previous dslrs.

I can understand a bit why neither Nikon or Pentax want to talk on the record about this issue, auto focus issues are chronic, and many times due to inexperience of the operator. If i were in one of those companies (and i'm not), i'd be concerned that every customer would send in their cameras, instead of learning how to operate them in the first place.

But i don't doubt there is an issue here due to the several experienced shooters who have mentioned it on these forums. What we don't know is whether Nikon or Pentax can come up with firmware changes to mitigate the problems.

I'm still holding off buying a K5 because i don't know how much this issue would affect my shooting, which often is indoor and low light. But i don't often shoot at f2.8 and wouldn't need it to be particularly accurate below 6'. I wouldn't like it but could live with those limitations. the only way i'm going to figure out if its going to bother my shooting is just to go buy the K5. Both of us K20 owners were really impressed with the D7000 that our friend had. He was shooting routinely at 6400 and getting what appeared to be good shots while we were shooting at 2000 to 2500 iso. The D7000 owner did seem to be getting some highlight clipping which i've read some complaints about.