While I was setting up to record audio from several radio time sources for the leap second last night (my leap second web page is at http://www.febo.com/time-freq/leapsecond-2005), I realized that it would be easy to use the same configuration to measure the delay of the audio output of the two receivers (SDR-1000 and Yaesu FT-817) when both were tuned to the same signal on the same antenna. I ran that experiment today, using WWV on 15MHz. The short answer is that with my setup, the delay is 187.6 milliseconds.
More detailed results, with screenshots, are at http://www.febo.com/geekworks/sdr1k/sdr1k_latency. I was using 1.4.5 preview 4 and the DSP buffer size was set to 2048; the sound card buffer size was 1024. I think these were the defaults when I installed the software, but if they're not appropriate, someone let me know and I can rerun the experiment with different values. More details about the test setup are at the web page. Happy New Year! 73, John