Announcement

Collapse
No announcement yet.

Delay after attention acknowledge phrase

Collapse
This topic is closed.
X
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

    Delay after attention acknowledge phrase

    George,

    I just started noticing this recently. I am running HS 1.7.7.

    6/6/2004 9:54:17 AM~!~SpeakEasy Debug~!~Received Dev=100 W=True -> yes bill
    6/6/2004 9:54:17 AM~!~SpeakEasy Debug~!~Issuing SpeakProxy: Device =100 Wait=True Text=yes bill
    6/6/2004 9:54:17 AM~!~Speak~!~yes bill
    6/6/2004 9:54:17 AM~!~SpeakEasy Debug~!~Audio has started
    6/6/2004 9:54:18 AM~!~SpeakEasy Debug~!~Audio has ended
    6/6/2004 9:54:18 AM~!~SpeakEasy Debug~!~Audio has started
    6/6/2004 9:54:30 AM~!~SpeakEasy Debug~!~Audio has ended
    6/6/2004 9:54:42 AM~!~Event Trigger~!~Voice Trigger (SpeakTime)
    6/6/2004 9:54:42 AM~!~SpeakEasy Debug~!~Received Dev=100 W=True -> It is now 9:54.
    6/6/2004 9:54:42 AM~!~SpeakEasy Debug~!~Issuing SpeakProxy: Device =100 Wait=True Text=It is now 9:54.
    6/6/2004 9:54:43 AM~!~Speak~!~It is now 9:54.
    6/6/2004 9:54:43 AM~!~SpeakEasy Debug~!~Audio has started
    6/6/2004 9:54:43 AM~!~SpeakEasy Debug~!~Audio has ended
    6/6/2004 9:54:43 AM~!~SpeakEasy Debug~!~Audio has started
    6/6/2004 9:54:46 AM~!~SpeakEasy Debug~!~Audio has ended

    As you can see in the log there are two sets of "Audio has started/ended" for each phrase spoken. The phrase was spoken immediately. I thought it might be my exit (which currently doesn't do anything) so I disabled it and it still happens. This is kind of annoying when the delay is on the attention acknowledge phrase. In this example it took 12 seconds to speak nothing, during which time listening was disabled. If I disable Speakeasy, the problem goes away. What is strange is that the delay on the second audio start/end is sometimes only one second, so it isn't noticeable.

    Thanks,
    Bill

    #2
    Bill,

    This has me at a little loss. I didn't respond before because I've been looking at the code in that area and honestly there is hardly any code in the Audio started/ended area. That routine is only in there to get the Audio event calls from HS so that the plugin can support the IsSpeaking functionallity. There's only about 7 lines of code (including the debug lines) from entry to exit. And the only thing the routines do is set some flags.

    Has anything changed lately on your end?

    Did this do this before you upgraded to 1.7.7?

    Could there be another process running that's sucking up CPU?

    If you notice in the log, when you see the entry of 'Issuing SpeakProxy' that when I pass control of the phrase back to HomeSeer and then HS then calls the TTS engine to speak it. From the time the plugin received the the phrase from HS to the time that it processed everything and passed it back to HS to be spoken it looks like it was all in the same second.

    I'm just curious if anything changed in HS that affects the SpeakProxy support (being you said that disabling the plugin makes it go away) or anything else might have been running on your system. Also, has it happened since you last posted. If it happens more often maybe you can enable a level 3 debug trace and we can see what else may be happening.

    Tx,
    George

    Comment


      #3
      George,

      I disabled the plugin for now. I'll set aside some time this weekend to gather more info.

      Bill

      Comment


        #4
        George,

        I turned on the level 3 trace (thought I had it on before) and the culprit is the slow AB8SS speaker switching. I disabled that and haven't noticed the delay since.

        Bill

        Comment


          #5
          Great!

          I'm working, as I type, on the fast AB8SS switching code that uses the new AB8SS set speakers by bitmap call. That should improve your performance significantly. That's the last thing I'm adding to the 1.2 release of the plugin.

          George

          Comment

          Working...
          X