Gain ref different size than tifs

[cgisriel]

Gain ref different size than tifs

[cgisriel]

It turns out that the data were accidently not collected in super resolution mode, which explains the problem.

[biochem-fan]

You have to downsample the super-resolution gain reference into the physical size.

You must take harmonic mean, not simple binning or Fourier cropping. This can be done with e2proc2d from EMAN2 but I forgot the command line. If I remember correctly, this was discussed on the CCPEM mailing list before.

Alternatively, you can use relion_estimate_gain to estimate gain from movie frames. Probably this is easier.

[cgisriel]

I was able to make relion_estimate_gain work. Thanks so much for your help!