Menu
What's new
The GetDPI Photography Forum

Great to see you here. Join our insightful photographic forum today and start tapping into a huge wealth of photographic knowledge. Completing our simple registration process will allow you to gain access to exclusive content, add your own topics and posts, share your work and connect with other members through your own private inbox! And don’t forget to say hi!

vignetting corrections and M-coding

S

scott kirkpatrick

Well-known member
Having the M-coder makes it relatively easy to explore (or "reverse-engineer") what the M8 firmware is actually doing so that one can make intelligent choices of how to code various lenses of interest for use with the M8.

For example, I tried my CV21/4 in the M-mount recently with both the Elmarit 21/2.8 pre-asph and the 21/2.8 asph codes and to within the accuracy of my testing, the two corrections were identical. This is based on just comparing the center and one corner in a series of white wall shots in indirect daylight at various apertures with and without a Leica filter installed. To be really sure takes more work analyzing the resulting images, and may take more careful setup, artificial light, etc. The firmware corrections can't be perfect since different light temperatures and different apertures require different corrections, but the M8 does appear to be trying to compensate for the increased luminance vignetting that occurs with the lens wide open.

My day job is keeping me pretty busy at the moment, but what questions do you want addressed next in this area? Which lens coding choices seem most important, what light environments, etc...? I can't promise instant answers, but will try to consider some of these.

scott
 
carstenw

carstenw

Active member
It would be helpful to arrive at some kind of quantitative description of a correction with respect to aperture, so that choosing the best lens type to fake for an unsupported lens would be made easier... That is a larger project, I realise, but that would be tremendously useful.
 
S

scott kirkpatrick

Well-known member
carstenw said:
It would be helpful to arrive at some kind of quantitative description of a correction with respect to aperture,
Click to expand...
The two limiting behaviors, wide open and f/8 or more, are fairly easy to determine. To convince the M8 that your lens, whatever it is supposed to be, is wide open, you cover the blue dot with a finger and the estimated aperture is the widest that the pretend lens can offer. And the estimating procedure with the blue dot usually seems to get f/8 quite accurately. In between, extracting the actual curve of the variation in the correction is hard, because of the noise in each measurement that you need. But it is worth trying, I agree.

scott
 
carstenw

carstenw

Active member
Not only the curve of the correction between apertures, but also from centre to edge... I wonder if simply measuring several points and then interpolating with splines would be good enough? A simple GUI tool should be able to help automate the process. Hmm...
 
S

scott kirkpatrick

Well-known member
carstenw said:
Not only the curve of the correction between apertures, but also from centre to edge... I wonder if simply measuring several points and then interpolating with splines would be good enough? A simple GUI tool should be able to help automate the process. Hmm...
Click to expand...
I turn them into jpgs and use a Matlab jpeg reader to get full curves. The good part of that is that this averages several pixels into a single rgb data point, reducing the noise, which is usually considerable. Sandy McG has a dng reader which extracts the individual (Bayer arranged) r, g, and b data, in leica's compressed format. Both of these have been transformed from their original linear form, which is proportional to the amount of light reaching the imager, but at least we know what Leica has done to the data while making it 8 bit. In either case, the curves that you get have noise in them, so comparing two curves to see if they represent the same or different processes is tricky...

scott
 
You must log in or register to reply here.
Top