Announcement

Collapse
No announcement yet.

Voice recognition error in log

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

    Voice recognition error in log

    Whenever I issue a voice command I get the following error in the log. The command works, but apparently something is missing causing the error report. HS3 is at latest level.

    May-26 9:11:33 AM Z-Wave Device: Node 3 Z-Wave Couch Lights Set to 99
    May-26 9:11:33 AM Warning In recognize: System.NullReferenceException: Object reference not set to an instance of an object. at A.c2e3d32053331d1c2520ebd17f9dfa64e.c036addc28d8512f6bb7b434 10cb9c307(String c0c4bb840068d7b8cbc2ae8c658cd4c5a, Int32 c641f32ff78b38533aa8c5cb87c7f4144, String c5848527c62da6fff46dcd41d7bc189ab, String c0cf6c15fc505ce256a93afdc5f2268a9) at Scheduler.clsVR.Recognize_Microphone(RecognitionInfo& vData)
    May-26 9:11:33 AM ASR Device voice command: turn on the Couch Lights

    #2
    How are you issuing your command? Speaker client, Echo, etc. ?
    💁‍♂️ Support & Customer Service 🙋‍♂️ Sales Questions 🛒 Shop HomeSeer Products

    Comment


      #3
      I am using a microphone directly to Homeseer system. ie Speaker Client

      Comment


        #4
        What version of HS3 are you running? Have you restarted your server lately?
        💁‍♂️ Support & Customer Service 🙋‍♂️ Sales Questions 🛒 Shop HomeSeer Products

        Comment


          #5
          This issue started when I upgraded to 3.0.0.258 back in April. I had hoped when I did the latest upgrade it would clear up, but no such luck. Below is the system data I should have added in the first place. I upgraded to Windows 10 from Windows 7 a few days ago (after the latest HS3 upgrade) (or should I say Microsoft upgraded it cause I hadn't yet read the latest warnings about their trickery)

          Current Date/Time: 5/26/2016 12:25:28 PM
          HomeSeer Version: HS3 Standard Edition 3.0.0.280
          Operating System: Microsoft Windows 10 Home - Work Station
          System Uptime: 0 Days 4 Hours 13 Minutes 50 Seconds
          IP Address: 192.168.1.30
          Number of Devices: 213
          Number of Events: 23
          Available Threads: 400

          Enabled Plug-Ins
          3.0.0.14: Applied Digital Ocelot
          3.0.0.17: iAutomate RFID
          1.0.0.11: Random
          3.0.5579.15955: Ultra1Wire3
          3.0.0.36: X10
          3.0.1.25: Z-Wave

          Comment


            #6
            bump

            Comment


              #7
              I had hoped someone might be able to give me a suggestion as to where the problem might be. I did a repair install with no change. I really didn't want to do a complete re-install if there is a possibility of fixing the issue.

              Comment


                #8
                You are not alone

                I am having the same issue running HS .297 Linux on a Rpi3 and using remote Windows machines running speaker.exe. I've experienced other "Object reference not set to an instance of an object" type errors in the HS log over the past 5 or 6 weeks of getting the Linux version running on the Rpi3. When I found out speaker.exe only works on Windows, I started setting up various versions of Windows (7, 8.1, 10) just to play around with it (using the separate download of speaker.exe from HS).

                When I issue a voice command remotely on a brand new install of Windows 10 Pro running on an Intel NUC or Windows 8.1 running as a virtual machine, controlling HS on the Rpi3, I get the same error(s) you got. The command will work, the lights turn on or off, but the log shows:
                12:57:58:2784:[ASR]->Device voice command: turn the Closet Light off
                12:57:58:2864:[Warning]->In recognize: System.NullReferenceException: Object reference not set to an instance of an object
                at A.c2e3d32053331d1c2520ebd17f9dfa64e.c036addc28d8512f6bb7b434 10cb9c307 (System.String c0c4bb840068d7b8cbc2ae8c658cd4c5a, Int32 c641f32ff78b38533aa8c5cb87c7f4144, System.String c5848527c62da6fff46dcd41d7bc189ab, System.String c0cf6c15fc505ce256a93afdc5f2268a9) <0x7014c000 + 0x080a8> in <filename unknown>:0
                at Scheduler.clsVR.Recognize_Microphone (Scheduler.RecognitionInfo& vData) <0x70162008 + 0x03c57> in <filename unknown>:0
                I assume you are running HS on your Windows 10 Home and not just running the speaker.exe app on it?

                I have another, more powerful NUC coming in the mail in a few days and thought I would try to re-install and re-setup everything from scratch to live on that machine, running Window 10 Pro. There was something awesomely cool knowing I was running HS on a $39 Raspberry Pi 3, but if the speed & reliability of HS can be monumentally increased and I can get HS voice control working by running it all on a Windows environment, then that's what I'll do.

                Comment


                  #9
                  Put in an email to support@homeseer.com as I'm not sure a speaker client on a Windows PC issuing VR commands to a Linux installation works.
                  Last edited by Rupp; September 11, 2016, 04:54 PM.
                  💁‍♂️ Support & Customer Service 🙋‍♂️ Sales Questions 🛒 Shop HomeSeer Products

                  Comment


                    #10
                    Also make sure that the dll's installed with the speaker client are the same version as those in .297.

                    Cheers
                    Al
                    HS 4.2.8.0: 2134 Devices 1252 Events
                    Z-Wave 3.0.10.0: 133 Nodes on one Z-Net

                    Comment


                      #11
                      Many months ago saw <filename unknown>:0 kind of errors running between HS3 Linux and Windows machines. Thinking it was just a nomenclature thing relating to paths defined in Linux versus Windows ...backslash versus forward slash stuff.

                      I recall opening a bugzilla ticket on it and over time it was fixed for Linux. Didn't pay much attention to it after a bit. It was relating to the play back of the wave file for TTS versus the use of the SAPI client.

                      VR hasn't been addressed in a long time because of the options of using cloud based VR / TTS.

                      Here did test using Windows 10 SAPI and VR and it worked with no errors in the speaker dot exe and never saw any errors in the Homeseer log.

                      Thinking too the use of a TTS proxy messes a bit with this stuff sometimes. I have here always stuck to using HS stuff without 3rd party proxies.

                      @Randy I am guessing you tested the MS SAPI TTS / VR without Speaker dot exe running to make sure everything was OK eh?
                      - Pete

                      Auto mator
                      Homeseer 3 Pro - 3.0.0.548 (Linux) - Ubuntu 18.04/W7e 64 bit Intel Haswell CPU 16Gb
                      Homeseer Zee2 (Lite) - 3.0.0.548 (Linux) - Ubuntu 18.04/W7e - CherryTrail x5-Z8350 BeeLink 4Gb BT3 Pro
                      HS4 Lite - Ubuntu 22.04 / Lenovo Tiny M900 / 32Gb Ram

                      HS4 Pro - V4.1.18.1 - Ubuntu 22.04 / Lenova Tiny M900 / 32Gb Ram
                      HSTouch on Intel tabletop tablets (Jogglers) - Asus AIO - Windows 11

                      X10, UPB, Zigbee, ZWave and Wifi MQTT automation-Tasmota-Espurna. OmniPro 2, Russound zoned audio, Alexa, Cheaper RFID, W800 and Home Assistant

                      Comment


                        #12
                        Thanks for the followups, guys!

                        Originally posted by Rupp View Post
                        Put in an email to support@homeseer.com as I'm not sure a speaker client on a Windows PC issuing VR commands to a Linux installation works or was expected to work.
                        First paragraph from "chapter 6: Voice Recognition", of the HS3 End User Documentation:
                        Voice recognition is handled by the speaker client application. While voice recognition is not support on Linux, it is supported
                        by running the speaker client on a Windows system and then connecting to your Linux system.
                        That is what I have done.
                        Originally posted by sparkman View Post
                        Also make sure that the dll's installed with the speaker client are the same version as those in .297.

                        Cheers
                        Al
                        Not sure about the dll's, but the stand alone version of speaker.exe on my Windows 10 Pro install was 3.0.0.7. I have since downloaded and installed .297 on the Win10 machine and speaker.exe is still 3.0.0.7. I had to reconfigure it after the full install, so I assume the app was overwritten during the full install.

                        Since the original poster seems to have been running everything on one Windows 10 Home computer, this error doesn't seem limited to remote speaker.exe controlling a Linux install.

                        I'll go ahead and finish the complete install on Windows 10 Pro anyway and see if the issue is still there. Can't setup any X10 devices via CM15A yet because IE is complaining about a missing ActiveX Control for it. As long as the voice commands go through and control my devices, I guess there's no harm done. But as a developer myself, it would drive me nuts to know that something was not behaving as intended.

                        Comment


                          #13
                          Originally posted by RandyInLA View Post
                          First paragraph from "chapter 6: Voice Recognition", of the HS3 End User Documentation
                          The documentation may not be correct. I would ask the HomeSeer technicians directly if you want the definitive answer. support@homeseer.com
                          That is unless you want to continue debugging this on your own.
                          💁‍♂️ Support & Customer Service 🙋‍♂️ Sales Questions 🛒 Shop HomeSeer Products

                          Comment


                            #14
                            Originally posted by Pete View Post
                            VR hasn't been addressed in a long time because of the options of using cloud based VR / TTS.
                            ...
                            Thinking too the use of a TTS proxy messes a bit with this stuff sometimes. I have here always stuck to using HS stuff without 3rd party proxies.
                            Having both the remote ivona.com TTS playing out of the Rpi3's headphone jack while TTS was also playing out of the Windows box showed me how bad the latency for using ivona.com's server was. I commented out the call to ivona in my speak.sh file and downloaded the UK English "Brian" voice directly to my Win10 box. It is SO much faster using TTS this way! Feels like an actual conversation instead of "give command... wait... listen to response..."

                            Originally posted by Pete View Post
                            @Randy I am guessing you tested the MS SAPI TTS / VR without Speaker dot exe running to make sure everything was OK eh?
                            I did test MS SAPI TTS/VR on the Win 7 & 8 virtual machines prior to installing speaker.exe, and it was fine, but went straight to speaker.exe on the Win10 machine.

                            Using Win10 VR/TTS via remote speaker.exe to the Rpi3, the recognition & responses are instant. The communication over my network to the Pi is fast as well. Lights turn on/off right away. I can only imagine how much more instantaneous it will be once everything is running on the same Win10 machine.

                            Comment


                              #15
                              Voice recognition is handled by the speaker client application. While voice recognition is not support on Linux, it is supported by running the speaker client on a Windows system and then connecting to your Linux system.

                              This statement is correct only relative to the use of Microsoft SAPI TTS and VR.

                              That said TTS/VR using the Amazon Echo is totally different. (or cellular phone TTS/VR).

                              TTS only to HSTouch Windows/Android or iOS clients uses a wave file playback. IE: a wave file is generated by HS and copied over to the client. This is not related to VR at all.

                              When you utilize the speaker dot exe program you are using Microsoft SAPI which is totally different. It is fast here running the speaker dot exe's on the the HS3 box with a VB.

                              When you use a TTS speech proxy you are doing that wave file thing.

                              When you are doing TTS/VR in the cloud it is TTS/VR is totally different and unrelated to SAPI.
                              - Pete

                              Auto mator
                              Homeseer 3 Pro - 3.0.0.548 (Linux) - Ubuntu 18.04/W7e 64 bit Intel Haswell CPU 16Gb
                              Homeseer Zee2 (Lite) - 3.0.0.548 (Linux) - Ubuntu 18.04/W7e - CherryTrail x5-Z8350 BeeLink 4Gb BT3 Pro
                              HS4 Lite - Ubuntu 22.04 / Lenovo Tiny M900 / 32Gb Ram

                              HS4 Pro - V4.1.18.1 - Ubuntu 22.04 / Lenova Tiny M900 / 32Gb Ram
                              HSTouch on Intel tabletop tablets (Jogglers) - Asus AIO - Windows 11

                              X10, UPB, Zigbee, ZWave and Wifi MQTT automation-Tasmota-Espurna. OmniPro 2, Russound zoned audio, Alexa, Cheaper RFID, W800 and Home Assistant

                              Comment

                              Working...
                              X