Audio KS Position test failing - different behavior on HLK vs local RRS feed

  • Question

  • Hi everyone,

    I'm currently testing a virtual microphone driver (no hardware) based on WaveCyclic Miniport driver.  

    I'm failing a number of tests in the KS Position test and seeming some interesting behavior on the WDMAudio::KsPosTest::TAEF_Capture_DurationLength#CaptureDefaultMonoFormat#0 in particular.  

    When run from HLK studio, it reports that it waits "20000 msec" (which I assume means 2000ms - 2 seconds) but then reports that "time before waiting period" is 517.775 and then "time at end of wait period" to be 2747.708 (duration of 2229.933).  

    It also reports that the driver streamed 0 bytes of 192,000.  Format of the pin is set to 48k, 16 bit, 1 channel. 


    Measured: 0
    Expected: 192000
    Error: -192000
    Time reading before starting of wait preiod: 517.775 ms
    Time reading at the end of wait preiod: 2747.708 ms
    Measured: 2229.933 ms
    Expected: 2000.000 ms
    Error: 229.933 ms (11.497 %)

    Now the inconsistent part is that when I run the same test locally via command line and TE.exe I get bytes streamed.  

    An example run looks like:

    Measured: 213312
    Error: 21312
    Time reading before start of wait period: 0.00ms
    Time reading at end of wait period: 2228.612
    Measured: 2228.612
    Expected: 2000
    Error 228.612

    Note that the amount of bytes streamed is correct for the measured time.  

    The failure reported is that the duration error exceeds the acceptable limit of 1.000%.

    Why would the amount of bytes streamed change so dramatically when running the test locally vs the HLK studio?  This is on the same machine with the same build of the driver.

    Thanks in advance!!

    Client OS: Windows 10 build 16299

    HLK server: Windows Server 2012 R2 Standard. 

    • Edited by AdamPH Thursday, April 19, 2018 1:40 AM
    Thursday, April 19, 2018 1:27 AM