Hi,
I am just wondering how I should handle instrument response deconvolution when the constant is set to zero in pole zeroes files? Is there a standard practice in this case?
Thanks !
Januka.
_______________
Serenity isn't freedom from the storm, but peace within the storm
Januka Attanayake
Postdoctoral Research Associate
Laboratório de Sismologia
Departamento de Física
Instituto Superior Técnico
Av. Rovisco Pais, 1
1049-001 Lisboa, Portugal
University email: januka.attanayake<at>ist.utl.pt
Home page: http://sites.google.com/site/janukaattanayake/
I am just wondering how I should handle instrument response deconvolution when the constant is set to zero in pole zeroes files? Is there a standard practice in this case?
Thanks !
Januka.
_______________
Serenity isn't freedom from the storm, but peace within the storm
Januka Attanayake
Postdoctoral Research Associate
Laboratório de Sismologia
Departamento de Física
Instituto Superior Técnico
Av. Rovisco Pais, 1
1049-001 Lisboa, Portugal
University email: januka.attanayake<at>ist.utl.pt
Home page: http://sites.google.com/site/janukaattanayake/
-
Dear Januka,
If the poles are not all zero then you know the
instrument frequency response but not its absolute
scale. You need the calibration factor of your
system (seismometer + digitiser), which should be
on (or derivable from) the manufacturer's data sheet.
Unfortunately you also have to work out the normalisation
factor of the poles-and-zeros. It's the ratio
polynomial of the complex poles / polynomial
of the complex zeros, at the calibration frequency (i.e. the
signal frequency at which the calibration factor
was measured - should also be in the manufacturer's
data sheet). Fortran or Matlab could be used.
Scherbaum's book, "of poles and zeros", describes this,
and there is a brief mention of it in the appendix of
the SEED manual.
Dividing this normalisation factor by the
calibration factor (in nm/count or nm/s/count depending
on whether your Ps and Zs are in displacement or velocity domain)
gives you the constant to put
with your poles-and-zeros before using the SAC
"transfer" command.
If you don't know the calibration factor or you can't work out
the normalisation factor then setting the constant to
one should allow "transfer" to run and give you a
seismogram corrected for the frequency response of the
instrument, but the units on the output seismogram will be
arbitrary and no use for magnitude calculations or making
into hodograms in combination with other seismograms.
Regards,
Sheila Peacock.
-
Thanks Sheila, it sounds like a doable calculation.
Januka.
_______________
Serenity isn't freedom from the storm, but peace within the storm
Januka Attanayake
Postdoctoral Research Associate
Laboratório de Sismologia
Departamento de Física
Instituto Superior Técnico
Av. Rovisco Pais, 1
1049-001 Lisboa, Portugal
University email: januka.attanayake<at>ist.utl.pt
Home page: http://sites.google.com/site/janukaattanayake/
On Monday, October 21, 2013 2:33 PM, Sheila Peacock <sheila<at>blacknest.gov.uk> wrote:
Dear Januka,
If the poles are not all zero then you know the
instrument frequency response but not its absolute
scale. You need the calibration factor of your
system (seismometer + digitiser), which should be
on (or derivable from) the manufacturer's data sheet.
Unfortunately you also have to work out the normalisation
factor of the poles-and-zeros. It's the ratio
polynomial of the complex poles / polynomial
of the complex zeros, at the calibration frequency (i.e. the
signal frequency at which the calibration factor
was measured - should also be in the manufacturer's
data sheet). Fortran or Matlab could be used.
Scherbaum's book, "of poles and zeros", describes this,
and there is a brief mention of it in the appendix of
the SEED manual.
Dividing this normalisation factor by the
calibration factor (in nm/count or nm/s/count depending
on whether your Ps and Zs are in displacement or velocity domain)
gives you the constant to put
with your poles-and-zeros before using the SAC
"transfer" command.
If you don't know the calibration factor or you can't work out
the normalisation factor then setting the constant to
one should allow "transfer" to run and give you a
seismogram corrected for the frequency response of the
instrument, but the units on the output seismogram will be
arbitrary and no use for magnitude calculations or making
into hodograms in combination with other seismograms.
Regards,
Sheila Peacock.
-