K20D metering

Hi all,
Just an interesting thing that came to me in a bit of inspiration... having read about the Leica M9 and the issues that beset the M8.

Had to test it first, but the K20D meter is sensitive to IR

Try this:
* Darkish room
* Spot metering, Av mode
* Take a remote control and point at centre of meter area
* Press remote and the exposure changes!

Can anyone try this with the K7? I wonder if they've introduced an IR cut filter over the AF and/or metering. It may well explain some of the issues with flash metering and focus under tungsten light

Also, has anyone got an IR-cut filter they could try in front of the lens??

Matt

Just to add a little more info to my previous post (in case it was confusing!)...

The Leica M8 has big issues with Infra-Red (IR) bleeding which seriously affected the image quality. M8 users had to put an IR-cut filter on the front of the lens to resolve the issue. Thankfully for those rich enough to afford Leica gear, the new M9 has a much better IR-cut filter in front of the sensor.

So, that led me to thinking about dSLR design. The sensors have IR-cut filters (in fact the K20D & K10D have quite a severe one compared to the original *ist-D which wasn't too bad for IR photography). But of course the optical path for the AF sensor (which sits beneath the mirror) and the exposure metering sensor (which sits on top of the pentaprism) is different.

Then I wondered about what's different between the K20D and the K7, and why the flash exposure seems to be more "what you expect" on the K7. Of course, the exposure meter is completely different - now a 77 segment meter.

So, my little test using a standard Infra Red remote control (actually a "6-in-1" remote with a heavy duty IR LED) shows that the metering is sensitive to IR. Consequently, if you have a situation where there's a fair bit of IR light - but no visible light - then the metering may say "Wow! That's bright, better stop down a bit", but then the sensor will not record the IR component at all. It's possible that you'd get what appears to be an underexposed shot.

And of course, for flash, the P-TTL system uses the same 16-segment exposure meter to determine the flash power. At the moment I can't prove my assertions because I don't have a light spectrometer handy. However, it may be that the flash guns output a fairly broad spectrum of light which may include some IR (and possibly UV) components. If that's the case, then it might be that the metering is seeing these IR components and making a wrong choice.

Finally, there were reported issues of focus errors in low light levels with tungsten lamps. I've seen it myself - snapshots of chicks under a warming lamp etc. And of course, there'll be a goodly amount of IR in that! It may well be that the AF mechanism is picking up on the contrast in the IR region and getting a 'dud' focus sometimes. (Anyone who's used manual SLRs of the old days will remember the focus adjustment markers for when you are shooting IR )

Anyone care to comment, or help me out with some comparative tests on the K7??



Matt

Very interesting Matt!

Not able to add anything about the K7 - still have the humble K10D.
But sometime ago I had wondered whetehr the under exposure of the K10D metering (and *istD before it) was in part down to UV. I never put a UV filter on the lens to do a 'before and after' scientific test. It was more of a feel. The problem occurs often, but I did notice a bit more with bright sky landscapes, even having tried adding exposure compensation!

I guess the comment about IR affecting everything would equally apply and probably makes more sense!

Andy

Well, anyone has an IR cutter filter to try some flash pics?
We know that reflections are a big problem with PTTL, so you just have to do a Before-and-after test with a flash, a reflective surface, and an IR-cut filter.

If the pics are the same, then IR has no effect on flash exposure...