Jerome Kieffer Mail

From canSAS
Revision as of 13:01, 28 July 2012 by Tobias Richter (talk | contribs) (Created page with "<pre> >Dear Adrian, > >As I explained to Claudio, I do not have time to >attend to your meeting in Upsala, this week-end and I apologies. >Nevertheless here are a couple of i...")
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
>Dear Adrian,
>
>As I explained to Claudio, I do not have time to 
>attend to your meeting in Upsala, this week-end and I apologies.
>Nevertheless here are a couple of information 
>about what I am doing in SAXS, and especially 
>for BioSaxs (Beamline BM29 fromerly ID14-3 at ESRF).
>
>* 2D storage of data: we use FabIO which has 
>been re-written for the "ESRF Data Format" to 
>accommodate multiple frames per file to keep 
>together intensity and pixel variance 
>(uncertainties). Please have a look at the draft 
>of "fabio.pdf". Curated 2D data could probably 
>benefit from HDF5 storage using the NeXus convention.
>
>* Data reduction can be made by various ways ... 
>in pyFAI (see attached submitted publication) we 
>aim at best performances for all ... SAXS being 
>very simple in this aspect (no tilt, no problem) 
>this library can nevertheless be used for SAXS 
>and will be put in production on the BioSaxs 
>beamline this fall. A 40x improvement in the data-reduction speed is expected.
>
>* 1D storage: I have very little ideas about 
>that as some program (ATSAS) enforces 3 column 
>ascii (q/I/sigma) with 1 line header. Unless my 
>scientists accept to change _OR_ Dmitri changes 
>his interface, I see little chance to see 
>another format emerging; especially not the 
>over-vebose XML cansas tries to promote. Once 
>again HDF5 could be an option but we are not using it here (because of ATSAS).
>
>I attached as well some slides presenting what 
>is ongoing on the BioSaxs beamline on the data analysis side.
>
>With my best regards,
>
>--
>Jérôme Kieffer