Charge:

The official charge is given at DM-11819 - Getting issue details... STATUS  (which has been assigned LDM-582).

A synopsis is:

  1. to identify the amount of loss that is considered allowable with respect to the LSST science requirements if a lossy compression were used, 
  2. to provide a recommendation on a lossy compression strategy for use within LSST pipelines along with a quantitative estimate of its impact on the sizing models LDM-144.


Meeting Information:

  • Currently scheduled to be weekly on Mondays at 10am PT (Noon CDT, 1 EDT)
  • Using https://bluejeans.com/343370910
  • A SLACK channel (#lossy_compression_wg) now exists for discussions by this group

Meeting Minutes:

Other Resources:

Members:

  • Robert Gruendl (NCSA, Chair)
  • Paul Price (Princeton)
  • Bob Armstrong (Princeton, WL)
  • Krzysztof Findesien (UW, Alert Production)
  • Sophie Reed (Princeton)
  • Ben Emmons (Tucson, observer)
  • No labels

2 Comments

  1. Gregory Dubois-Felsmann

    The WG should be aware of the mention of compression in the system-level requirements document LSE-30, in OSS-REQ-0111, "Science Image Archiving Reliability", which basically says that images must be archived exactly, except:

    Corruption of images shall mean changes of any image pixel values, with the exception of application of a compression algorithm.  Any such compression algorithm must be shown to have no effect on any SRD requirement.

    So this "no effect on any SRD requirement" condition is something this working group needs to address - perhaps first by defining more precisely what this constraint means, and then by evaluating this condition against specific compression levels.

    1. Gregory Dubois-Felsmann

      Strictly speaking, the way this requirement is written, it applies only to raw images.