Any progress report of v 1.4?

Ken, at the possible cost of oversharing gory details here's the scoop.  We have over 40 threads that run in the main radio processor to handle everything the radio has to do.  We made some alterations to the Linux kernel and the software to ensure that events are more carefully scheduled avoiding conflicts and ensuring the most important tasks are always done first.  This included an implementation of priority-based preemptive multitasking, a multi-level locking structure using Pthread rwlocks and mutexes and movement of some algorithms from the processor to the FPGA where it made sense.  All of these things make the radio run cleaner and more efficiently and will be noticeable during normal operation I would say.  All of these changes were implemented and tested across all of the FLEX-6000 line.

On top of this, the SmartSDR client was worked over from a performance standpoint and a number of major enhancements in performance were made.  A lot of these changes had to do with how we handle memory on the PC.  Microsoft tends to hide a lot of this from programmers, which is generally good.  With a high performance program that is expected to draw and communicate a lot of data, we really have to try a number of things and see where we can get the underlying language runtime (CLR) and the OS to do what works best.  The results have been fantastic.  I've seen a number of "marginal" computers go from having difficulty to performing very well in the transition to the new code.  This will help everyone that runs SmartSDR.

As you mention, we did implement a standard audio codec to transport audio between the PC and radio that has excellent performance, excellent audio quality and good self-healing for situations where there is some packet loss.  This codec is expensive in terms of CPU usage, but the results are really amazing.  I literally cannot tell that a codec or any compression is involved.