Announcement

Collapse
No announcement yet.

Speeding up TTS

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

    #16
    Originally posted by ssk View Post
    Thanks Dirk, my comments are below...

    >>The event for "Garage door closed" is wrong. You should change the "Speak the content of the file" to a "speak" action.
    Yes I fixed it yesterday. I was trying many different things to see if it would fix the problem

    >>Secondly, make sure you "debounce" door open/close events so you filter out multiple (false) events. The open has a 30 second time, the close has none.
    Can you provide more details on this please. I have 30 second timer for Open because the event is checking for open status of the door that is at least 30 sec and I want the event to be triggering every 30 sec. For Close, the even will trigger only ones and is the reason why I do not have 30 sec timer for re-triggering.

    >>The log shows you have an issue playing an announcement but the log doesn't show anything else. You should change your settings and REMOVE the "Log Errors only flag" because it suppresses everything else that is being logged. Also turn off the superdebug flag which generates a lot of logging that is not needed to troubleshoot this.
    I changed the log settings as per your suggestion and converted the HS log to CVS. Please link for log file that is converted in to CVS file. The log file is too big to upload.

    >>Are the status of the Sonos devices still being updated properly? What firewall are you using? Is the HSPI_Sonos.exe allowed through the firewall? It is not enough that HS3 is allowed by the firewall, all PIs run as standalone programs and they all need their own allowance.
    I am using windows firewall and here is the screenshot of the plugin settings in windows firewall. every thing is allowed.
    [ATTACH]59702[/ATTACH]

    Please let me know if you need additional information. thank you for your help. looking forward to have this fixed and use my TTS reliably.
    De-bounce it by change it (for example) to If status becomes closed for 2 seconds (or something like that).

    I can't open the .csv file, excel complains it has too many columns or rows. One way is to page through the HS log, this log should not be more that 2 perhaps 3 pages (remember NO superdebug) and you cut/paste each page to notepad, save and upload. Alternatively, you can try the log-to-disk-option (I think it works reasonably well in that version of the PI). You would go to the config page, click it on, which will open the log file, do the announcement, go back to the config page and turn it off. Turning it off will close the log file. Turning back on will overwrite so be careful. Log is at <hs root>\html\sonos\logs\Sonosdebug.txt

    Are your devices still updating in real time (meaning if you change a player from pause to play, it shows up properly in the PI)?

    Are your announcements still working or not at all at this point? The error you posted before typically is when the announcement doesn't work at all. In that case, the announcement file might be corrupted or there is an issue with your setting for text-to-speech.

    Dirk

    Comment


      #17
      Originally posted by dcorsus View Post
      De-bounce it by change it (for example) to If status becomes closed for 2 seconds (or something like that).

      I can't open the .csv file, excel complains it has too many columns or rows. One way is to page through the HS log, this log should not be more that 2 perhaps 3 pages (remember NO superdebug) and you cut/paste each page to notepad, save and upload. Alternatively, you can try the log-to-disk-option (I think it works reasonably well in that version of the PI). You would go to the config page, click it on, which will open the log file, do the announcement, go back to the config page and turn it off. Turning it off will close the log file. Turning back on will overwrite so be careful. Log is at <hs root>\html\sonos\logs\Sonosdebug.txt

      Are your devices still updating in real time (meaning if you change a player from pause to play, it shows up properly in the PI)?

      Are your announcements still working or not at all at this point? The error you posted before typically is when the announcement doesn't work at all. In that case, the announcement file might be corrupted or there is an issue with your setting for text-to-speech.

      Dirk
      Managed to get the small log file. link

      Yes it updates the real time but late like 5 sec.

      The TTS work but very slow. some time it is like more than a min.

      Thanks again for your help.

      Comment


        #18
        Originally posted by ssk View Post
        Managed to get the small log file. link

        Yes it updates the real time but late like 5 sec.

        The TTS work but very slow. some time it is like more than a min.

        Thanks again for your help.
        Looking through your log, I think this test has a big of an issue in that you run the announcement right after a boot of the PI. Update of status of players is 100% dynamic. In this case the PI hasn't received proper player status of all the players. In most cases when players are discovered and events are subscribed to, the player automatically sends status update. I suspect because all the players are already grouped that the grouped players do NOT send updates (that's how Sonos implemented it). So this PI starts, adds all the players, all players are grouped and none sends a status update, so the PI at this point doesn't know the status of the players. Now the announcement starts and the PI is waiting for the source player to go to status stopped, which never comes and nothing happens.
        Simple way is play something and pause it, that will update the status of all the grouped players. In this log, all players are linked in 2 seconds, I suspect it actually played the announcement but because it never received a play event either is waiting for 2 minutes after which it cleans it all up.

        So first advice, don't generate an announcement as the first thing you do after a restart of the PI. Play something on involved players to make sure the PI is 100% up to date on their status.

        Secondly and a bit of a side effect, you have S1 players (first gen) and S12 players (second gen S1). Sonos calls them all S1 players but they announce themselves different in SW, I suspect because they have different (future) capabilities. The version you are using doesn't understand what an S12 player is. In itself not a big deal except that S1 and S12 players can be used to create stereo pairs which is a whole other bowl of wax on how to treat those. You should step up the version 3.1.x.x which can be found in the beta section of the updater. The step up will keep all your events and HS devices in check BUT you will have to recreate your musicDB (read help file if you don't know how) and your radiostationDB (play the radiostations again).

        Last but not least, your source player is the family room which in itself is a stereo paired configuration. You could try to pick another player as source that is not a stereo pair. It is again one of those Sonos idiosyncrasies where in the stereo pair, one player does all and the other is dead as a doornail. I can't tell whether you picked the right one because Sonos gives them the same name (something else that is fixed in version 3.1.x.x) so you might have chose the player that is the "dead" player. The PI should be able to deal with all of this but you created quite a perfect storm in this particular test, with announcement right after restart, all players are linked and your source player might be the dead slave in a stereo pair.

        Dirk

        Comment


          #19
          Dirk, The new plug-in helped a lot. Also I separated my stereo setup to make sure all Sonos speakers are reporting to the PI. Much better now. I still have small lag like 2-9 sec and I think I have to live with that. Now all my TTS are working and none of them are lost. Thanks a lot for your help. Let me know if I can do anything else to completely eliminate the lag.

          Thanks again.

          Comment


            #20
            Originally posted by ssk View Post
            Dirk, The new plug-in helped a lot. Also I separated my stereo setup to make sure all Sonos speakers are reporting to the PI. Much better now. I still have small lag like 2-9 sec and I think I have to live with that. Now all my TTS are working and none of them are lost. Thanks a lot for your help. Let me know if I can do anything else to completely eliminate the lag.

            Thanks again.
            no chance to eliminate the lag, just the way it all works unfortunately

            Comment


              #21
              I understand. Thanks a lot for all your help, Dirk. Really appreciate all the support.
              Last edited by ssk; February 24, 2017, 02:19 AM.

              Comment

              Working...
              X