HaruhikoT Offline Upload & Sell: Off
|
p.7 #2 · p.7 #2 · Voigtlander Ultron 28mm f/2 Lens tests with PCX Front Filter | |
Hi Fred and Guy,
Thank you for sharing your test results as always.
I'm also shocked about the PCX2500 results. We PCX team struggles this time and I don't have good answer about why this happens.
Fred Miranda wrote:
I could be totally wrong about this but my theory is that the CV 28/2 Ultron optical design is not sensitive to a sensor stack thickness just like the CV 15/4.5 III M-mount.
Have you noticed the CV 15 M-mount works great on Leica and Sony? I think the same applies here.
This is unlikely IMO. This lens' schematic shows symmetric design and also from the picture of the lens itself's backside, we can see its exit pupil size similar to front-side.
That means it has steep ray angle towards sensor so it must be sensitive to the cover glass thickness.
Perhaps this CV28 is, at least Fred's copy is optimized for thicker sensor stack by Cosina?
My previous simulation results for ZM35/1.4, CV35/1.7 and this CV28/2 already took such 'optimization for digital' into consideration, but I set optimized sensor stack to 0.8mm (Leica M9).
If cosina considers compromise between Leica's stack (0.8mm~1mm) and Sony's stack (around 2.5mm), it may be optimized for around 1.5mm. In that case, PCX2500 should be too strong and deteriorates IQ.
I still can't believe Cosina did such aggressive optimization for their M-mount lens, but recently majority consumer of CV lens may be E-mount user so perhaps this is possible story.
Anyway, more test results may give us some hints. I'll wait for Guy's test with PCX2000
I have interest how IQ looks at wider aperture or at closer distance.
How about the field curvature? If PCX2000/2500 is just too strong, FC with these PCX may be n-shape (Corner resolves closer subjects than center).
Haruhiko
|