Experienced Nikon and Minolta photogs DON'T switch out of matrix
metering which is extremely accurate on those monikers. Even the
itty-bitty point and shoots.
I'm going to design a rock bottom matrix metering solution for you on
the fly. I'm going to divide the sensor area into a 3x3 grid and assign
the light intensity in each grid section to a binary number from 0 to 3
and store the result in two bits. I'll use the JPEG luminosity values
to represent light intensity. binary values from 0-63 will be assigned
0, 64-127 will be assigned 1, 128-191 will be 2 and 192-255 will be 3.
I now have a grand total of 9 x 2 bits = 18 bits. In total they can
represent 256K unique values.
Now I go out and take zillions of well exposed photographs taken by
professional photgraphers. I bring them back to the lab and analyze
them all with a photometer. I divide them into the same grid of 9 cells
and assign each grid an exposure value of 0-3 just like the camera is
going to do. Eventually, I discover that the composite value 165,293
represented by those 18 bits was properly exposed by the pro using EV
10. I build a 256K array of EV numbers and use the index values
computed from the array as an index to the EV table.
Does this work? I haven't a clue. But that's my simplistic view of a
matrix metering system. The Japanes version, of course, can't be
complete unless it incorporates lots of AI and fuzzy logic as well. :-)
Chuck Norcutt
Ali Shah wrote:
> I am a newbie to all of this. I have read and perhaps
> mis-understood, however it seems that Matrix Metering
> is not very dependable. Experienced photogs switch out
> of Matrix Metering. I just heard about of Matrix
> Metering recently...and no clue how it is different
> from SPOT, Center Weighted, etc.
==============================================
List usage info: http://www.zuikoholic.com
List nannies: olympusadmin@xxxxxxxxxx
==============================================
|