On 01-03-11 10:54, Holger Hans Peter Freyther wrote:
I still don't have the right architectual image of GSMTAP i think. Do i need to direct wireshark to sniff on the same interface that tetra-rx opens its udp socket on, with a port filter, or is there a pseudo interface that listens for raw packets on the same UDP port?
There is no pseudo interface. The tetra decoding program sends UDP packets to the IANA assigned port for GSMTAP on the specified ip address. E.g we like to use multicast addresses as this avoid seeing ICMP messages as the destination is not reachable.
The easiest would be to use wireshark's pseudo interface to scan every interface and then filter for udp in the trace and select gsmtap in the filterbar of wireshark.
Ok. I've straced tetra-rx and it opens on 127.0.0.1. So i wiresharked it on lo with a port filter. That should work. But tetra-rx doesn't do a write on the udp-socket. That could of course mean there is no tetra data in the capture.
I've rebuild the usrp1-tetra_demod.py in the latest grc from git trunk, and the 25Khz lowpass gives me a nasty spike on -22Khz. A 20Khz gives me a much sharper idle constellation. Do you see any problem with 20Khz lowpass?
Gr. Sim