[Lowell Goudge]

Very useful post, Lowen. Thanks.
As for light scatter this was my point - lenses in focusing screen could cause such behaviour (as K10D/D20D screen is similar to rear projection TV screen). But I do not belive it as my primary suspect is software implementation.
Anyway, my proposed solution is to make special configuation text file. And you will be able to scroll throught lenses defined in this file, and each will have SR distance, exposure compensation for each diaphragm value, etc.

The more I look at DSLR metering the more I notice how dated this thing is.
My panny TZ5 meters much more accurately then GX20 in matrix mode :-)

Having gone through this issue, all I can say is that the focusing screen on the K10/K20 is a real issue in terms of metering, and you will find, if you substitute an *istD screen in it's place that M lenses will then meter much better.

Just as a little bit of information, note that the thread I attached the link to was about using teleconverters.

If you take a sigma TC, which is a pass through only TC for the aperture contacts, and use it, the lens will have a metering error. for example, my sigma 70-200 F2.8 needs to under expose by .7 stops with a 1.4x TC and 1./3 stops with a 2X TC. on the K10D, on my *istD and K7 (both of which use an XX-60 screen) the TC does not impact the metering,
ALSO, if you use pentax's SMC-F 1.7x AF TC on a lens, the metering is correct. The difference, the Pentax TC modifies the lens data and aperture so that the camera knows the real aperture of the lens and TC combo.

The only way that this can be happening is if the non linearity as a function of aperture is compensated for in the camera using the Fstops reported by the camera through the TC.

That is why the same behavior exists on A lenses when taken out of A mode as exists on K lenses, once the camera does not knot the actual aperture it is totally lost.

Edit note

As an afterthought, this might explain why there have been complaints about DA lenses and metering issues. The problem of light angle is not specifically limited to the focusing screen but also to the sensors. the biggest design difference between DA lenses and the rest is not the coating on the rear element to stop reflections back from the sensor causing loss of contrast and flair, but the optical formula is changing to have the light hitting the sensor at a higher angle, to avoid vignetting due to the performance fall off of the sensor with light at low angles. This change in design will undoubtably cause a different scatter than older lenses that were designed to use with film.

---

Last edited by Lowell Goudge; 07-30-2009 at 12:25 PM.

[tr13]

I prefer to use Occam razor in this field. All this light angles must not cause errors more then 0.2-0.3 stops.
Anyway, full configuration file could solve all issues by applying automatic EV correction.
I don't belive that we'll be able to fully understand all exposure metering in Pentax firmware. But finding EV correction value address is much simpler.
And yes, for A lenses camera knows starting aperture (but for zoom lenses it don't work - aperture becomes not correct during zoom) and diaphragm have different lever rate. Both could be fixed.
Pentax programmers could do it in a week. It is not hard job having firmware source.

[Lowell Goudge]

I prefer to use Occam razor in this field. All this light angles must not cause errors more then 0.2-0.3 stops.
Anyway, full configuration file could solve all issues by applying automatic EV correction.
I don't belive that we'll be able to fully understand all exposure metering in Pentax firmware. But finding EV correction value address is much simpler.
And yes, for A lenses camera knows starting aperture (but for zoom lenses it don't work - aperture becomes not correct during zoom)

note that later zoom lenses (including one from sigma that I took apart) modified aperture information when zooming, and they still do today, to correctly report aperture when it changes, although I only have 2 such lenses as my others (zooms) are constant aperture hence aperture is always the same regardless of focal length

and diaphragm have different lever rate. Both could be fixed.
Pentax programmers could do it in a week. It is not hard job having firmware source.

note lever rate does not change on variable aperture lenses, the lever spans an allowable range the entire range shifts therefore lever rate remains the same.

As for older lenses, the way to resolve this is to enter the maximum aperture and then use a dual step metering, where you press the green button and it meters, stops down, meters again, and knowing the change inlight, as well as the maximum paerture it could control a lens, or enter maximum aperture and use the control lever to control the lens, but since the pentax body is designed for a linear motion of F stop (i.e. area) with displacement, but the K &m lenses have a diameter vs displacement motion, the exposure accuracy would become worse and worse as the lens is stopped down

[tr13]

As for older lenses, the way to resolve this is to enter the maximum aperture and then use a dual step metering, where you press the green button and it meters, stops down, meters again, and knowing the change inlight, as well as the maximum paerture it could control a lens, or enter maximum aperture and use the control lever to control the lens, but since the pentax body is designed for a linear motion of F stop (i.e. area) with displacement, but the K &m lenses have a diameter vs displacement motion, the exposure accuracy would become worse and worse as the lens is stopped down

Lever is displaced by firmware to predefined positions. We can change this during firmware hack :-) Or we just can make EV adjustments contained is special text file.
So, you approach problem from user point of view, and I try to look at this as programmers problem. and it is clear that PEntsx is able to make very accurate M metering, even better then that A lenses have.