Loading ...
Sorry, an error occurred while loading the content.

45961Re: [mh] Festival on Audreys & local

Expand Messages
  • Stephen Switzer
    Oct 31, 2013
    • 0 Attachment

      Yes it does. The newest version also includes clash code, so the separate speech_clash.pl isn't needed when using pa_control.pl. All examples you need are in code comments, but let me know if there's something I didn't cover, or you find a typo...

      On Oct 31, 2013 1:02 PM, "H Plato" <hplato@...> wrote:
      Thats exactly the issue. The audrey pre_speech hook seems to call the voice_text generation, and after it's done voice_text generation runs again and the socket hasn't closed yet so it falls back to running the festival program ... Which takes a few seconds to startup.

      I'll try out your pa module, does that fully replace audreyspeak.pl?

      Sent from my mobile device.

      > On Oct 31, 2013, at 10:04 AM, Steve Switzer <steve@...> wrote:
      > Howard,
      >   Yes, I was working on the PA object and speaking to various devices.
      > I think there's a chance that you'll have a single voice with the PA
      > object, but I don't fully understand your environment. You have a PA
      > system, and Audreys?
      >   It appears that the audreys are being sent a file, and the MH server
      > is sending audio out the sound card as well. This might be the issue.
      > Either way, check out the PA object with the Audrey declarations in your
      > MHT file. You'll need to use the version that I have in my pull request.
      > Best regards,
      > Steve Switzer
      > ---
      > Get world-class business I.T. services and a phone system with awesome features that won't challenge your budget!
      > http://www.SwitzerBusinessSolutions.com
      >> On 10/30/2013 11:05 PM, H Plato wrote:
      >> Now that I have some house speakers, I’ve finally connected the local audio output. I’ve been using festival with multisyn voices on 4 Audreys with good results (other than they are naturally out of sync so a speech event is off by a second or so on the different floors).
      >> So I connected the main out to an amp and generated speech. Audrey started with the familar voice, and then about 2 seconds later the local audio played and it was an entirely different voice!  It looks like Voice_Text isn’t using the festival client, but rather is spawning a brand new instance of festival because the socket hasn’t closed yet. This is the line in Voice_Text.pm
      >>            if ($VTxt_festival and active $VTxt_festival)
      >> I’ve also tried to disable audreyspeak.pl, but get silence.In a perfect scenario, the speech gets generated once and then sent to the receivers (local, audrey, tablets). I think Steve might be workign on this with his PA object.
      >> Here are the speak/voice settings from my mh.private.ini
      >> speak_engine=festival
      >> voice_text=festival
      >> voice_text_festival=/usr/bin/festival
      >> festival_host=localhost
      >> festival_port=1314
      >> voice_names = male=>male1, mike=>male1, sam=>male2, elder_male =>male3, mary=>female
      >> speak_apps_default =   email => voice=male rooms=all, deep_thought => voice=male rooms=all,
      >>                     tagline => voice=male rooms=all,      chatbot => voice=male rooms=all,
      >>                       top10 => voice=male rooms=all,        timer => voice=male rooms=all,
      >>                          tv => voice=male rooms=all,      phone => rooms=all mode=unmuted,
      >>                      router => voice=mail no_chime=1, system => mode=mute,
      >>             dvd => mode=mute, frog => voice=mail,
      >>             cashier => chime=cash_register voice=mail
      >> speak_apps_default =   email => voice=male rooms=all, deep_thought => voice=male rooms=all,
      >> Here is the log when I send speech.
      >> Speaking eliza data with voice=male, compression=
      >> Voice_Text: speak app chatbot card  voice male compression  text please work to the local server requestor
      >> normal: please work to the local server
      >> input  phrase is 'please work to the local server'
      >> output phrase is 'please work to the local server'
      >> 10/30/13 08:51:25 PM pre_speak_to_audrey  rooms: all
      >> 10/30/13 08:51:25 PM pre_speak_to_audrey rooms: basement office bedroom kitchen
      >> 10/30/13 08:51:25 PM Clash control stub called!
      >> 10/30/13 08:51:25 PM speak_text: parms are  'app'='chatbot' 'async'='0' 'audreySpeakRooms'='ARRAY(0xa100054)' 'card'='' 'compression'='' 'engine'='festival' 'length'='500' 'mode'='' 'raw_text'='please work to the local server' 'requestor'='' 'rooms'='all' 'text'='please work to the local server' 'to_file'='' 'voice'='male' 'volume'='' 'web_file'='web_file' 'web_hook'='ARRAY(0xb0c4bec)'
      >> 10/30/13 08:51:25 PM speak_text: parms are  'app'='chatbot' 'async'='0' 'audreySpeakRooms'='ARRAY(0xa100054)' 'card'='' 'compression'='' 'engine'='festival' 'length'='500' 'mode'='' 'raw_text'='please work to the local server' 'requestor'='' 'rooms'='all' 'text'='please work to the local server' 'to_file'='/usr/local/misterhouse/instances/main/data/web/speakToWeb0.wav' 'voice'='male' 'volume'='' 'web_file'='web/speakToWeb0.wav' 'web_hook'='ARRAY(0xb0c4bec)'
      >> Voice_Text volume=, mh_volume=0 vc=
      >> Voice_Text new volume=0
      >> Data sent to festival: (Parameter.set 'Audio_Method 'Audio_Command)(Parameter.set 'Audio_Required_Rate 16000)(Parameter.set 'Audio_Required_Format 'au)
      >> 10/30/13 08:51:25 PM Voice_text TTS:  Festival saving to file via server: /usr/local/misterhouse/instances/main/data/web/speakToWeb0.wav
      >> -10/30/13 08:51:25 PM web_hook_callback: web/speakToWeb0.wav
      >> 10/30/13 08:51:25 PM file ready for audrey web/speakToWeb0.wav
      >> Running: ./get_url -quiet /dev/null &
      >> Running: ./get_url -quiet /dev/null &
      >> Running: ./get_url -quiet /dev/null &
      >> Running: ./get_url -quiet /dev/null &
      >> Voice_Text volume=, mh_volume=0 vc=
      >> Voice_Text new volume=0
      >> Voice_text TTS: /usr/bin/festival --tts /usr/local/misterhouse/instances/main/data/mh_temp.festival.205.sable
      >> 10/30/13 08:51:25 PM web_hook_callback: web_file
      >> Please wait: Initialising multisyn voice.
      >> Voice loaded successfully!
      >> 10/30/13 08:52:00 PM: Saving object states ... done
      >> ------------------------------------------------------------------------------
      >> Android is increasing in popularity, but the open development platform that
      >> developers love is also attractive to malware creators. Download this white
      >> paper to learn more about secure code signing practices that can help keep
      >> Android apps secure.
      >> http://pubads.g.doubleclick.net/gampad/clk?id=65839951&iu=/4140/ostg.clktrk
      >> ________________________________________________________
      >> To unsubscribe from this list, go to: http://sourceforge.net/mail/?group_id=1365
    • Show all 7 messages in this topic