search

LBEM-CH / focus

High Throughput Electron Microscopy Image Processing Software
http://www.focus-em.org
GNU General Public License v2.0
36 stars 14 forks source link

LMAX setting location/meaning #239

Open Ziloan opened 5 years ago

Ziloan commented 5 years ago

Hi,

When trying to run CTF Correction (Part 1) script on a 2dx dataset recorded in super-resolution mode, it gives the following error: "ERROR, image too large. Increase LMAX. Aborting. ABORTING"

Could you tell me where this LMAX setting is? Could not find anything in the settings.

Please find more information in the attached logfile. verbo3_ctfcorr1.txt

Thank you, Gergely

sthennin commented 5 years ago

Hi,

LMAX is a dimension parameter that defines the one-dimensional size of several buffers in the historic MRC programs. Those F77 programs usually contain lines such as PARAMETER (LMAX=20100) DIMENSION ALINE(LMAX)

Originally, the MRC programs were written for very small images. Over the years, the sizes were increased by us and others in various programs. Current programs within the FOCUS package allow working with 8k x 8k images, I believe.

How large is the image that you want to work with? Which program is acting up?

I guess that you are working with a K3 image in super-resolution mode, which could be expanded to a square image of 10k x 10k size? If so, it would be best if you could send me one such example image, and I try to process it from A to Z and adjust all LMAX and other dimensions of the various F77 programs to become able to cope with those dimensions.

All the best,

Henning.

Ziloan commented 5 years ago

Dear Henning,

It is indeed a K3 image in super-resolution mode with the dimensions of 11520*11520. I am sending an example image to you! Motion correction with Motioncorr 2.1 and CTF estimation with CTFFIND 4.1 worked, (launched from FOCUS) and all the scripts in the 2dx processing pipeline worked up until the point of Ctf Correction (Part 1) where the error occurred. (I am using the 20190507 version of the software) As a quick-fix is there an option to downscale/bin the existing images, so I could progress in the pipeline? Or maybe import the images as if they were acquired not in super-resolution mode, and repeat the pipeline?

Thank you for the quick response, Gergely