[falconeye]

However for the LV - Light Value that determines the AF limits -
it is reflected light that matters -
because that is what the camera "sees" -

I think I didn't express myself clear enough.

What you say is correct, but that wasn't my point.

If a vendor says AF in -3EV (like K-5II) or -2EV (like D800), then it usually means incident (not reflected) light (scene illuminance) and a bright black&white target, i.e., optimum conditions. At least, this is my suspicion without any further detail, always being sceptical when it comes to marketing dep. specs. AFAIK, there is no ISO or CIPA standard AF target.

However, under identical conditions, auto-metering will give a reading like -1EV (or 0EV resp.) exposing the white as 18% gray (if auto-exposure is still working at all). In other words, you won't see AF working at such low EV combinations in the real world, esp. if the focus target has normal contrast, not an ideal 100% black&white one. Moreover, you must add at least one additional EV to guess AF performance in tungsten light because of a lack of illuminance in the green channel.

In other words, -3EV (Pentax) may mean 0EV (in practice). And both numbers would still be correct. But don't get me wrong, that's still pretty dark

---

Last edited by falconeye; 11-08-2012 at 08:34 AM.

[UnknownVT]

I think I didn't express myself clear enough.

What you say is correct, but that wasn't my point.

If a vendor says AF in -3EV (like K-5II) or -2EV (like D800), then it usually means incident (not reflected) light (scene illuminance) and a bright black&white target, i.e., optimum conditions. At least, this is my suspicion without any further detail, always being sceptical when it comes to marketing dep. specs. AFAIK, there is no ISO or CIPA standard AF target.

Got you, thank you.

So your suspicions are
manufacturers may use a bright white target (with a focusing mark) to measure their AF lowest light level -
that would indeed inflate their lowest limit by several stops -
since zone 5 or V (18% gray) to zone 10 or X (white)
is a 5 stops difference......

FWIW -
I do a lot of photography in what I consider ridiculously low light conditions on an almost weekly basis -
with a K-x - the light levels in some areas are below the metering and AF spec limits.

ISO5000, f/3.5, 1/5sec, -0.7 comp; 18mm
metering segment read outs:


Perhaps seeing the original unprocessed (other than re-sizing) might help to visualize the scene a bit better?
EXIF attached

according to the metering segment it is about -1.3LV at the face
this is about -1.3 stops below the spec'd metering,
and -0.3 stops below the spec'd AF limit of the K-x -
metering is supposed to be 0LV and AF -1LV, but using a f/1.4 lens -
I am using f/3.5 which is -2 2/3 stop slower/dimmer..

[JinDesu]

I think I didn't express myself clear enough.

What you say is correct, but that wasn't my point.

If a vendor says AF in -3EV (like K-5II) or -2EV (like D800), then it usually means incident (not reflected) light (scene illuminance) and a bright black&white target, i.e., optimum conditions. At least, this is my suspicion without any further detail, always being sceptical when it comes to marketing dep. specs. AFAIK, there is no ISO or CIPA standard AF target.

However, under identical conditions, auto-metering will give a reading like -1EV (or 0EV resp.) exposing the white as 18% gray (if auto-exposure is still working at all). In other words, you won't see AF working at such low EV combinations in the real world, esp. if the focus target has normal contrast, not an ideal 100% black&white one. Moreover, you must add at least one additional EV to guess AF performance in tungsten light because of a lack of illuminance in the green channel.

In other words, -3EV (Pentax) may mean 0EV (in practice). And both numbers would still be correct. But don't get me wrong, that's still pretty dark

In consideration - the k-5II has been shown to perform (and perform well) in conditions where the k-5 has failed.

I would hope and believe that Pentax's method of measuring AF EV sensitivity was kept the same between their camera models - which gives us an understanding as to the actual improvement when compared to Pentax cameras we currently own (the -3EV of the k-5II vs the -1EV of the k-5/k-r/k-x).

Would be great to have a standardized method of measurement for AF EV sensitivity, but we could do real life tests between cameras of different brands to set relative benchmarks - and with benchmark information, we could review the manufacturer numbers and normalize their scales.