A rather quick update after 1.0.1. There were some major bugs that were worked on since, so a rather quick release before working on other things felt important.
So obviously, this is mainly bugfix…
Until now, channel 3A and 3B were not being separated on AVHRR/3 data. In view of other upcoming features/changes, they are now being split according to the configuration downlinked by the satellite. Composites were modified to account for this.
Initial UHD (USRP) supported has been added and tested with X300 and B200 hardware. While things do appear to be working just fine, it’s not impossible some bugs are left so perhaps don’t consider it fully stable just yet.
Thanks to F4GKR for lending the USRPs for support :-)
There was more work on figuring things on LRIT/HRIT. Images are now produced correctly… But work on those downlinks is not over yet!
It was until now required to run a second command manually to finish processing down to imagery in CLI mode. Now, adding the
--finish_processing flag will process the decoded data fully when exiting. If the flag is not present, things will act as did before.
The console log level as well as if you wish to output a satdump.logs file can now be set in settings.
Sometime, if you happeend to record an “empty” pass (eg, JPSS-2 comissioning!) you would end up with products containing no actual data… And the viewer would crash.
OpenJPEG is utilized for GK-2A HRIT, FengYun-4 xRIT and GOES-R GRB decompression. Though unfortunately, there was a bug leading it to call a non-aligned “free” on Windows… Which caused things to crash.
EMWIN files on Windows were being corrupted due to a missing parameter when writing the file.
Quite a few old, forgotten-about DSP issues were fixed… Some that could lead to NaNs, etc locking things up. Oops. This happened very rarely, hence making it so complicated to figure out.
HimawariCast was pretty likely to abort on any bad filename / data.
Processing coud hang on a bad admin message (eg, corrupted). This was extremely unlikely to ever happen as Bzip2 decompression should normally fail first, but it did happen recently!
FY-4x live processing will now be done down to images directly.
The ability to stream frames from one “demodulator” node to others decoding the data has been added on other xRIT downlinks, as it is a rather common configuration on those setups.