Leginon

h1. TIA raw data conversion

h2. TIA raw (*.bin)

The raw format in TIA is a binary file with a header.  Similar to mrc format image files, this can be parsed and converted to numpy array and then converted to any other image format.  A python script that has a few basic functions such as read and readHeaderFromFile can be found in pyami/tiaraw.py

For microED purpose, we choose to save int32 data type since the image we received are gain/dark corrected.

h2. TIA raw to mrc

Within leginon/diffrtransfer.py script this image is converted to mrc format, also at int32.  If you want to do this yourself, you can use pyami/tiaraw2mrc.py to make the conversion.  See the script in the repository for instruction.

h2. mrc to SMV

SMV is a simple crystallography format that is typically in unint16.

To retain the statistical characterization, Leginon/diffrtransfer.py converts each of the diffraction series with a constant offset that is recorded in SMV header as LEGINON_OFFSET

LEGINON_OFFSET is (-1) * minimal of all image min values in the diffraction series.  This is not expected to be very large if the gain/dark reference are correct.  Larger values are usually caused by cosmic ray spike and should be ignored.  Therefore, we set a cut off at 2000.

DIALS does not read LEGINON_OFFSET but wants PEDESTRAL value.  We make an estimate of reasonable PEDESTRAL value as the maximum of all image min values in the diffraction series after LEGINON_OFFSET is applied.  This is defined as such since proper gain/dark corrected image should find its background value averaged at 0 in raw or mrc images.

For example, for a diffraction series that has three images.  Their raw/mrc  image minumal values are (-10000,-145,-200).  LEGINON_OFFSET will be 2000.  Any values in the first mrc image that is less than -2000 are truncated in the corresponding SMV file.  PEDESTRAL value = -(-2000) + (-145)

For more details, please examine leginon/diffrtransfer.py