new 50/1.2 $1600 (continued...)

gml said:

And I never argued against this one.

Click to expand...

Then you're arguing against the reason given (larger sensor)?

--
Lee Jay
(see profile for equipment)

ljfinger said:

You can quantify the portion that is caused by the increase in
sensor size and that's what we've done here (log2 1.6^2 = 1.357
stops).

Click to expand...

According to Phil's review, there certainly isn't a linear correlation:
http://www.dpreview.com/reviews/canoneos5d/page21.asp
How can you prove the validity of your formula.

ljfinger said:

gml said:

And I never argued against this one.

Click to expand...

Then you're arguing against the reason given (larger sensor)?

Click to expand...

I'm arguing against you simplistic theory and the 'evidence' that supports it.

gml said:

According to Phil's review, there certainly isn't a linear
correlation:
http://www.dpreview.com/reviews/canoneos5d/page21.asp

Click to expand...

Since they don't have the same pixel dimensions, and aren't processed in such as way as to match detail, that test is not a test of my claim. It's also a test of JPEGs which mean it's as much a test of the JPEG engines as it is of the sensor performance. I'm only talking about the sensors themselves so I do all the testing in RAW and convert with the same converter.

gml said:

How can you prove the validity of your formula.

Click to expand...

There are several ways. The argument that physical aperture at constant FOV is what really matters makes intuitive sense to many people, but apparently not to you. I do have a couple of drawings that demonstrate the concept conclusively (the very same photons end up in the very same sensor elements in both cases) but I'm not ready to publish the article they go into. I've been too busy with shoots lately and have four more scheduled for September. But I'm working on it.

--
Lee Jay
(see profile for equipment)

gml said:

I'm arguing against you simplistic theory and the 'evidence' that
supports it.

Click to expand...

Well, it is simplistic in that it isolates a single variable - format size. I explicitly ignore (hold constant) other variables which were coined "dirty details" by poster xandxor. But I stand by my theory given my assumptions (all theories are constrained by the assumptions that go into their development so this is not a weakness, it's science).

Fortunately, it seems that, when it comes to Canon cameras of similar generation but different formats, those assumptions largely hold true.

--
Lee Jay
(see profile for equipment)

ljfinger said:

gml said:

I'm arguing against you simplistic theory and the 'evidence' that
supports it.

Click to expand...

Well, it is simplistic in that it isolates a single variable -
format size. I explicitly ignore (hold constant) other variables
which were coined "dirty details" by poster xandxor. But I stand
by my theory given my assumptions (all theories are constrained by
the assumptions that go into their development so this is not a
weakness, it's science).

Click to expand...

It would also be interesting to hear how your theory would explain the difference in noise performance between the new 400D and the 30D. They both have the same size. One has more MPs but according to Canon almost the same noise characteristics.

gml said:

It would also be interesting to hear how your theory would explain
the difference in noise performance between the new 400D and the
30D. They both have the same size. One has more MPs but according
to Canon almost the same noise characteristics.

Click to expand...

Since my theory explicitly ignores these things (technological improvements), it wouldn't. It's about changing format with constant technology. Period. All the testing that gets done here by Phil and Simon, by other reviewers and by users is about teasing out those details because they are important as well. Heck, that's one reason I'm with Canon (better sensor performance per unit of sensor area compared to Sony). But those are other discussions not related to my threory (developed in concert with others) or its claims.

--
Lee Jay
(see profile for equipment)