Miljenko wrote:RobertOToole wrote:...but with the slower old-fashioned manual method I can easily compare files.
Trust me, Robert, it is even easier to compare numerical figures than stare to blocks and traces.
I actually prefer to look at images, some of the wafers have interesting patterns, rather than just stare at numbers. I can see an engineer or accountant might prefer the opposite.
In my experience some sites, not your tests, but some sites that use MTF are useful to compare lens to lens within that website but when you look at the real world results from that lens there is not much correlation between images and the ratings. For example one lens might place 2nd of all time based on MTF but the lens is almost unusable in the real world due to poor coverage and high CAs.
RobertOToole wrote:It would be a different story of course if I had 100 MTF curves on my hard drive already.
I believe that so far you have tested more than 100 lenses. So you actualy could have 100+ MTF curves on your HD. :-)
You have put a huge effort (not to mention investing in all those lenses) to get to very valuable results. However, if you were using MTF measuring methods, results would be universal and comparable over a wider database.
I disagree with you here Miljenko. I firmly believe you cannot compare MTF data from one source to another, or even within the same site or database sometimes due to test parameters changing.
Some sites sharpen Jpeg images for MTF, some use RAW converted files, Some use razor blades, some sample 10 lenses and spin the lens for 5 readings and average. There are too many variables.
Some MTF test sites us an optical bench without even using a camera body.
After all, lens manufacturers are using MTF method with a good reason.
These are just computer simulations based on an ideal lens design. Only one manufacturer shares tested MTF and that is Zeiss so that is not exactly a endorsement. I would argue the opposite. MTF from manufacturers is one way the marketing department can mislead customers with overly optimistic performance promises so these should never be trusted.
Also as far as I know, no MTF lens test site or manufacturer shares MTF data at close up distances or magnification. 100% as far as I know are at an estimated infinity distance. We all know this is not going to be an accurate way to judge a 1x macro lens.
Testing a production lens at magnification the only real way to know how a lens performs.
Although visual comparison can be fun, it can sometimes be missleading. Even USAF 1950 comparison method is not 100% proof as it depends on visual assesment. Tried it all since 2004. and ended with MTF method as the most reliable.
I agree with you here.
Best,
Robert