Phase retrieval p9, always converging on the same solution
Hi Vincent,
Probably in relation with the issues raised by David the other day. I noticed that when I run the phasing with devel.p9 on jupyter slurm I always get the same solution, even if my support is not fixed. On the other hand, when I run the reconstruction on scisoft13 with devel.debian9 with the exact same parameters, I get a large variety of solutions, as expected. I guess this has something to do with the PSF... This is illustrated in the following directory: /data/id01/inhouse/maxime/Beamtimes/LTP_HC4050/results/S323/. The parameters I used for the phasing are in pynx-cdi-input_tryo.txt (defaut PSF). The reconstructions performed with the p9 are in run_1_p9 while the reconstructions with scisoft13 are in run_1 and run_2. The notebook is in the same directory. Have you ever noticed that before?
Other example (with more variability): /data/id01/inhouse/maxime/Beamtimes/SIXS_2021_06/results/S_4098/pynxraw/
Merci!
Maxime