Kota Hayashi is making the analysis now. There are some initial results. For a fiber (~50mm away from the center), defocus images (both -4mm and +4mm) do not look significantly different in blue and red.

Now the obscuration fraction is being calculated (at first geometrically) but the values highly depends on how to measure (~15-35%). In any case, however, the analysis does not show any significant difference between blue and red.

Some more details:

These images are from datasets taken on July 27, 2023 (PFI) and May 11, 2023 (DCB). The fiber used here is fiberId=12, which is the one closest to the center on the PFI that I have. The positions of these spots on the detectors: xc~3950, yc(b1)~3240, yc(r1)~2510.

Trying to estimate the flux obscuration geometrically, i.e. just considering area, I found that the result is subject to the way of defining a circle for each spot. Circles defined like the red ones in the figure below, the estimated obscuration was (mask area) / (circle area) ~ 0.65, while it was ~ 0.85 with circles like the blue ones.

Next, we need to consider the actual flux as well, not just geometrically. Possible ways to do so that I can think of include: interpolation of a certain uniform value into a spot image (seems unreal), or 2d interpolation by e.g. scipy.interpolate.interp2d (large obscuration seems to make it difficult).

This seems to be duplicate of this ticket: https://pfspipe.ipmu.jp/jira/browse/PIPE2D-980. You can estimate obscuration directly from the images, but you do have full obscuration results derived from the images. Also please note that you have to also take care that you take into account that obscuration of the flux depends also on the pupil illumination profile - it is more complex than (mask area)/(circle area).

Thank you! And, sorry, I totally forgot that ticket. I actually suggested to use the model image so that the illumination does not matter, but is that your estimation in the previous ticket? Anyway, we should recall the results, which should be include in the ETC anyway.

Yes, the estimation is the previous ticket is the average decrease of the illumination due to obscurations in the image, on SM1, with PFI.

This last comment https://pfspipe.ipmu.jp/jira/browse/PIPE2D-980?focusedCommentId=30597&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-30597 seems to been have prepared for you.