Firmware 1.10

[Godfrey DiGiorgi]

On May 1, 2007, at 10:05 AM, Cory Papenfuss wrote:

>> My Katz Eye screen is a plain matte fresnel surface without metering
>> aids, a custom design for scribed lines.
>>
>> Yes, I talked with Rachel Katz about the problem I was seeing at
>> length. Her take on it is that Pentax reshaped the scatter of the
>> screen for the K10D model to provide more brightness at the ocular,
>> which has the effect of reducing the amount of light at the metering
>> sensors. So they use a different calibration curve for the sensors,
>> based on the notion of less light going their way compared to the DS,
>> to give accurate metering.
>>
>> I conjecture that there's an error in the calibration curve used when
>> the camera is in the green-button stop-down metering mode. The
>> progression of error as the aperture deviates from wide open is
>> regular and it seems to me that the problem is an incorrect
>> multiplier or lookup table.
>>
>> Godfrey
>>
>  	... or just a variation in the lens?  Doesn't the angle of light
> coming in depend on not only the aperture, but also the design of the
> lens?  As such, it's probably only possible to do an "average"  
> correction.

If that were true, then TTL metering in the general case wouldn't  
work consistently for different lenses in any exposure mode at  
different lens openings.

Consider the SLR focusing screen as a rear projection screen, with  
the illumination source being the lens. Its built-in fresnel lens  
collimates what would otherwise be relatively simple hemispherical  
scatter, to first order approximation, and directs it towards the  
ocular.

The metering sensors are designed to read light intensity from that  
scattering rear projection surface, they don't intercept the light  
directly from the lens. The scatter induced by the screen's matte  
focusing surface ameliorates the variations from that would otherwise  
be seen with a clear screen.

(When special purpose near-clear and clear screens were available in  
the past for cameras like the Nikon F series, charts regarding TTL  
metering errors, both for total ambient curve and for focal length  
adjustments, were included. Most of the time, they recommended not  
using the TTL metering due to variabilities like this..)

Godfrey