[data-issues] UW.LON bad compression

[Chad]

Just my 2 cents.  Something is definitely wrong with that particular data.
The miniSEED records that I looked at did not have any indication of their
encoding format as there are no 1000 blockettes.  I tried to decompress the
data as 16/32-bit ints, 32/64-bit floats, Steim-1 and Steim-2.  None of those
produced anything that looked like real data.  So the data is either badly
packed/compressed or in a different encoding (24-bit ints?)

Many of the DMC tools default to Steim-1 when no encoding is indicated.  Maybe
something Philip is using is defaulting to 32-bit floats?

Chad

Philip Crotwell wrote:
> Hi
> 
> For UW.LON in the POND for this event, I get an error that looks like
> the data is either badly compressed, or has the wrong compression type.
> The mseed record thinks it is 32 bit floats, but there are not enough
> bytes for the number of points in the record. Other events for this
> station might have this same problem, I haven't looked.
> 
> When I try to look at this in wilber, I get an error of "Unable to
> generate plots at this time. Sorry." for this station, so your wilber
> plot generator is probably seeing the same issue.
> 
> This is via DHI, but I suspect the problem is in the underlying data,
> not the server.
> 
> thanks,
> Philip
> 
> Event: Komandorskiye Ostrova Region | 10/20/1992 04:40:02 GMT | Mag: 5.5
> | Depth 33.00 km | (55.525, 166.429)
> 
> UW.19630301T00:00:00.000Z.LON.  .BHZ.19920601T00:00:00.000Z Data
> retrieval in progress
> java.lang.IllegalArgumentException: Not enough bytes for 2014 32 bit
> data points, only 4036 bytes.
>         at edu.iris.dmc.seedcodec.Codec.decompress(Codec.java:73)
>         at
> edu.iris.Fissures.seismogramDC.LocalSeismogramImpl.decodeBits(LocalSeismogramImpl.java:1398)
> 
>         at
> edu.iris.Fissures.seismogramDC.LocalSeismogramImpl.get_as_longs(LocalSeismogramImpl.java:1053)
> 
>         at edu.sc.seis.fissuresUtil.bag.Cut.cut(Cut.java:59)
> 
> _______________________________________________
> data-issues mailing list
> data-issues at iris.washington.edu
> http://www.iris.washington.edu/mailman/listinfo/data-issues