-
George found a related issue in our configuration GUI and I can no longer repeat the bug. I am considering it fixed and closing.
2009-01-02 18:48:20 UTC in PAMGUARD
-
The problem appears to be related to the fact that there are no PamGuard events generated when a run is started and stopped and the underlying framework has no knowledge that the sample number (i.e., time) is going to reset itself when start is pressed again.
The PamDataBlocks manage purging their buffers via periodically having removeOldUnitsS() called on them. They then calculate the...
2008-12-30 16:44:34 UTC in PAMGUARD
-
When the spectra plot is set to "apply spectrogram scale", the spectra scale doesn't update when the spectrogram scale is changed.
The workaround is to open and close the settings dialog for the spectra plot itself after making a change. This is simple, so I will set initial priority to low, though it would frustate a beginner.
2008-12-24 12:39:00 UTC in PAMGUARD
-
I get a similar error when running with multiple files. I suspect that there is an odd interaction between the core and the LD. I am going to task our team to spend a couple of hours on it next week and see if they can learn more.
2008-12-23 16:01:08 UTC in PAMGUARD
-
I am not sure if this helps to track down the bug, or makes it harder to repeat. When I switched to a new mode (using a folder with multiple files), that had a 4kHz sample rate, as soon as I went into the LD module the sample rate (and available bandwidth) was correctly understood. In this case I had loaded a prior psf file. When I first observed the bug I had started with an empty psf file.
2008-12-23 15:19:27 UTC in PAMGUARD
-
The following is a dump of the JAVA console for a run using the Likelihood detector running from a wave file (Fs = ~80kHz). Everything works great when run from start (psf file attached), but if I stop and try to run the same file again I start to get a bunch of "Unknown Source" errors and then an out of memory error (see below). The errors seem to be related to the display widgets. I am not...
2008-12-23 15:08:52 UTC in PAMGUARD
-
When configuring the Likelihood Detector (LD), I set up a file based sound acquisition first. I changed the bandwidth to 40kHz in the dialogue to match the sound file. I then went into the LD configuration to set up a Spermwhale target file, but the module was only aware of a 1 kHz bandwidth until I started and stopped processing. Once this had been completed the correct bandwidth was available...
2008-12-23 13:25:04 UTC in PAMGUARD
-
It _might_ be worthwhile to add plots of the Likelihood Detection function. This could be a compressed (MAX across many bands) or individual by band, which may end up creating clutter. There would be one or two likelihood functions (detection functions) per band, depending on whether or not guard bands are defined.
1773110 suggests another option using a level meter display.
This is an...
2008-12-23 01:07:17 UTC in PAMGUARD
-
I would add that this would work for detection functions as well. Some kind of normalized compressed detection time series (e.g. across all detectors of type X) could be displayed to avoid plotting the various functions below a spectrogram, which may take up needed space. The normalization would be by the current threshold ( 1 = at threshold).
The display could also have min/max bars with...
2008-12-20 12:03:16 UTC in PAMGUARD
