VladK wrote: >On Wed, 2006-12-06 at 23:50 +0800, Steve Underwood wrote: > > >>Some of those logs don't contain the debug information, so they are not >>very helpful. The ones that do contain debug show what we have seen >>before - the mediatrix machines are weird. we need to make our software >>compatible with them, but we haven't had enough access to a mediatrix >>box to really debug things properly. Is this mediatrix box accessible on >>the internet, so we would try interacting with it to debug things? >> >>Regards, >>Steve >> >> >> > >My company took mediatrix box from reseller just for testing. It's not >on the Internet, but I could perform necessary tests, just let me know >what kind of tests should be done. > BTW, do you have any comments on Grandstream HT 488 with T.38 ? >Thanks Steve. > > The first set of tests you did all failed in the same way. There seems to be a bug in the Mediatrix box. It sent the following:
Rx 73: data type T38_DATA_V17_14400/T38_FIELD_NON_ECM_DATA + 36 byte(s) Rx 74: data type T38_DATA_V17_14400/T38_FIELD_NON_ECM_DATA + 36 byte(s) Rx 75: data type T38_DATA_V17_14400/T38_FIELD_NON_ECM_DATA + 36 byte(s) Rx 76: data type T38_DATA_V17_14400/T38_FIELD_HDLC_SIG_END + 0 byte(s) It ended a burst of non-ECM data with an HDLC end of signal message. That's wrong, and and my T.38 software didn't like it. I have changed spandsp so it should tolerate this rubbish. There are probably more issues affects the other series of tests you did. However, if you still have the mediatrix box can you try the snapshot of spandsp I just created: http://www.soft-switch.org/downloads/snapshots/spandsp/spandsp-20061207.tar.gz If you still have the equipment, can you try your first series of tests again with this software, please? Regards, Steve _______________________________________________ Openpbx-users mailing list [email protected] http://lists.openpbx.org/mailman/listinfo/openpbx-users
