edit · history · print

Software.SigprocIssues History

Show minor edits - Show changes to markup

March 22, 2007, at 08:10 AM by mkeith -
Changed lines 13-14 from:

Workaround: Patch with old re-sampling code.

to:

Workaround: Patch with old re-sampling code, or remember to account for negative values

March 22, 2007, at 08:02 AM by mkeith -
Changed lines 11-14 from:

Versions tested: 4.1,4.2 & 3.8 Versions affected: >=4.1 Workaround: Patch with old re-sampling code.

to:

Versions tested: 4.1,4.2 & 3.8
Versions affected: >=4.1
Workaround: Patch with old re-sampling code.

Changed lines 22-25 from:

Versions tested: 4.1,4.2 & 3.8 Versions affected: all Workaround: A patch: patch. Note that this will mean that the DM is not set to 0 when completely decimating in frequency.

to:

Versions tested: 4.1,4.2 & 3.8
Versions affected: all
Workaround: A patch: patch. Note that this will mean that the DM is not set to 0 when completely decimating in frequency.

March 22, 2007, at 08:00 AM by mkeith -
Changed lines 20-21 from:

When resampling a timeseries using decimate, the DM value in the header is set to 0.

to:

When resampling a timeseries using decimate, the DM value in the header is set to 0, even if no frequency decimation occurs.

March 22, 2007, at 07:59 AM by mkeith - Added two new bugs
Deleted lines 6-20:

Dedisperse

Issue - Baseline causes flux degridation

When using dedisperse with small numbers of channels (<12) and a strong source, the baselining causes the pulse power, and region round the pulse to be lowered. This may happen to a lesser degree with low power pulsars or with many channels, but it does not appear to be significant. This is probably caused when the pulse power is strong enough that single pulses can be detected in a single channel.

Versions tested: 3.8 only
Versions affected: 3.8 (probably affects all versions)
Workaround: Use -nobaseline

Submitted by: M Keith


Fixed Issues (as of 4.2)

Added lines 8-44:

Issue - Acceleration values negative

New acceleration time stretching code causes incorrect acceleration value to be searched. This means that any detections in acceleration are negative of the real value.

Versions tested: 4.1,4.2 & 3.8 Versions affected: >=4.1 Workaround: Patch with old re-sampling code.

Submitted by: M Keith


Decimate

Issue - DM set to 0

When resampling a timeseries using decimate, the DM value in the header is set to 0.

Versions tested: 4.1,4.2 & 3.8 Versions affected: all Workaround: A patch: patch. Note that this will mean that the DM is not set to 0 when completely decimating in frequency.

Submitted by: M Keith


Dedisperse

Issue - Baseline causes flux degridation

When using dedisperse with small numbers of channels (<12) and a strong source, the baselining causes the pulse power, and region round the pulse to be lowered. This may happen to a lesser degree with low power pulsars or with many channels, but it does not appear to be significant. This is probably caused when the pulse power is strong enough that single pulses can be detected in a single channel.

Versions tested: 3.8 only
Versions affected: 3.8 (probably affects all versions)
Workaround: Use -nobaseline

Submitted by: M Keith


Fixed Issues (as of 4.2)

Seek

Changed line 51 from:

Versions affected: 4.1\\

to:

Versions affected: <=4.1\\

March 22, 2007, at 07:49 AM by mkeith -
Added lines 5-6:

Current Issues (as of 4.2)

Added lines 20-21:

Fixed Issues (as of 4.2)

Added lines 26-27:

Fixed - Various patches to normalization routines.

February 15, 2007, at 10:18 AM by mkeith -
Changed lines 20-21 from:
to:

After upgrading to 4.1, seek produces consistently lower SNR values than previous versions.

February 15, 2007, at 10:15 AM by 130.88.9.75 -
Added lines 16-25:

Seek

Issue - Seek in Sigproc-4.1 generates snrs (reconstructed and spectral) a few percent lower than its older counterpart Sigproc-3.8.

Versions tested: 4.1 and 3.8 only
Versions affected: 4.1
Workaround:

Submitted by: R Eatough

February 15, 2007, at 09:12 AM by mkeith -
Changed line 6 from:

Issue

to:

Issue - Baseline causes flux degridation

Changed lines 9-10 from:
to:

This is probably caused when the pulse power is strong enough that single pulses can be detected in a single channel.

February 15, 2007, at 09:11 AM by mkeith -
Changed line 6 from:
to:

Issue

Changed lines 10-11 from:

Fix: Use -nobaseline

to:

Versions tested: 3.8 only Versions affected: 3.8 (probably affects all versions) Workaround: Use -nobaseline

February 15, 2007, at 09:08 AM by mkeith -
Changed lines 3-4 from:
Please sign your entries here.
to:
If you add any issues please sign your name!
February 15, 2007, at 09:08 AM by mkeith -
Changed lines 1-2 from:

Sigproc Issues

to:

Sigproc Issues

February 15, 2007, at 09:07 AM by mkeith -
Changed lines 7-8 from:

When using dedisperse with small numbers of channels and a strong source, the baselining causes the pulse power, and region round the pulse to be lowered.

to:

When using dedisperse with small numbers of channels (<12) and a strong source, the baselining causes the pulse power, and region round the pulse to be lowered. This may happen to a lesser degree with low power pulsars or with many channels, but it does not appear to be significant.

Fix: Use -nobaseline

February 15, 2007, at 09:06 AM by mkeith - dedisperse problem
Added lines 1-10:

Sigproc Issues

Please sign your entries here.

Dedisperse

When using dedisperse with small numbers of channels and a strong source, the baselining causes the pulse power, and region round the pulse to be lowered.

M Keith


edit · history · print
Page last modified on March 22, 2007, at 08:10 AM