Originally posted by blende8 I don't buy this test.
It is suggesting that the FA is resolving almost three times as many lines at f4 than at f1.4.
This is not in agreement with my tests.
I have done many 50mm tests and the resolution is almost the same at f1.4 and at f4.
Only the contrast is lower at f1.4.
Here is another comparison, completely unenhanced: 200%
Wieland,
that's an awesome demonstration of the necessary distinction of resolution and contrast.
While in the analog era, full MTF curves were measured on an optical bench (leaving the sensor, AA filter and Bayer demosaicing and (over-)sharpening out of the game), it is now replaced by digital MTF50 figures which are very sensitive to all of the above.
If I'm not mistaken, resolution is typically defined as MTF5 (5% contrast) frequency for MTF curves measured on an optical bench. Because digital MTF curves are so artifact-prone (incl. demosaicing Moiré), one cannot simply use the digital 5% value though.
Anyway, Zeiss lenses resolve about 300 - 400 lp/mm around f/4 and no digital test so far revealed that. I came close to seeing this resolution in reversed lens macrophotography using a Zeiss lens as a front lens and a tele lens as a back lens. I've seen 1.5µm resolution translating into 330 lp/mm.
Originally posted by 0144 77zone metering system gives 77 dots for histogram, if they could increase zones to several hundred and make up some fake points by interpolating zones nearby, they could make a decent histogram.
i don't think we'll see it in this generation though
Originally posted by glanglois I think we'd need more than a few hundred points to get a histogram that's accurate within a few percentage points at each point on the horizontal axis and do so 97% of the time.
But then I don't remember enough math to model this. Anyone?
Well, you get the
exact histogram of the image scaled down to 11x7 pixels (0.2% size). It's not a question of statistical variation but of the presence of high frequency high contrast features within areas of a few hundred pixels. On the other hand, small size specular highlights may be exactly the kind of features you want to ignore when exposing to the right...
I for one would like a 77 field expose to the right metering mode, maybe with 1 stop in reserve as the default.
Originally posted by Christopher M.W.T I dont know much about the new VF other than the improvements relate to how the focus points are projected, it will apparently be similar to what Nikon and Sony does.
How reliable is this info?
A change of focus point display in the VF would point to a more than cosmetic evolution of the AF system, possibly with more and/or different AF sensors.
Last edited by falconeye; 08-30-2010 at 01:15 AM.