Home of the Squeezebox™ & Transporter® network music players.
Page 6 of 6 FirstFirst ... 456
Results 51 to 60 of 60
  1. #51
    Junior Member
    Join Date
    May 2021
    Posts
    2

    Announcement only working after l restart squeezeboxserver service on rpi4

    Announcement only working after manual restart of squeezeboxserver service on rpi4
    Hi, I am using this plugin on my rpi4 with raspbian.

    Installed from Debian package.

    Code:
    Logitech Media Server Version: 8.1.1 - 1610364019 @ Thu Jan 14 06:24:07 CET 2021
    Betriebssystem: Debian - DE - utf8
    Plattformarchitektur: armv7l-linux
    Perl-Version: 5.28.1 - arm-linux-gnueabihf-thread-multi-64int
    Audio::Scan: 1.02
    IO::Socket::SSL: 2.060
    Datenbankversion: DBD::SQLite 1.58 (sqlite 3.22.0)
    I use gTTS python cli for tts. Because it outputs mp3 I had to adjust the announceTTS.wav in the Plugin.pm in InstalledPluginsFolder.

    Thats the current state.

    Now my problem:

    When my rpi is started and all services running I can use the squeezeboxserver to play on all my players without a problem. But when I trigger the LMSAnnounce API the music just stops and no sound is coming, the resume to playlist seems also buggy.

    The Logfile of LMS shows this:
    Code:
    [21-05-06 17:32:21.5945] Plugins::LMSannounce::Plugin::handleJSONP (194) volume specified: 60
    [21-05-06 17:32:21.5954] Plugins::LMSannounce::Plugin::handleJSONP (201) handleJSONP cmd=announceTTS pid=ab:cd:ef:12:34:02 forcewake=
    [21-05-06 17:32:21.5962] Plugins::LMSannounce::Plugin::handleJSONP (210) handleJSONP targetClient=ab:cd:ef:12:34:02
    [21-05-06 17:32:21.5969] Plugins::LMSannounce::Plugin::savePlaylistState (374) savePlaylistState called
    [21-05-06 17:32:21.5997] Plugins::LMSannounce::Plugin::announceTTS (323) announceTTS called: ttspreset=gtts text=Ansage: Test-Ansage
    [21-05-06 17:32:21.6010] Plugins::LMSannounce::Plugin::announceTTS (325) announceTTS: preset={'options' => '-l de','output' => '-o','text' => '','cmd' => '/home/pi/.local/bin/gtts-cli'}
    [21-05-06 17:32:21.6018] Plugins::LMSannounce::Plugin::announceTTS (330) announceTTS: command: /home/pi/.local/bin/gtts-cli -o /tmp/annTTS.mp3 -l de "Ansage: Test-Ansage"
    [21-05-06 17:32:22.6038] Plugins::LMSannounce::Plugin::announceFile (342) announceFile called: file=/tmp/annTTS.mp3 targetvolume=60
    [21-05-06 17:32:22.6233] Plugins::LMSannounce::Plugin::__ANON__ (350) announceFile: plid=1379
    [21-05-06 17:32:22.6554] Plugins::LMSannounce::Plugin::__ANON__ (357) announceFile: time=85.8651892604828 repeat:0 targetvolume: 60
    [21-05-06 17:32:22.7832] Slim::Formats::Playlists::Base:laylistEntryIsValid (125) Warning: file:///tmp/annTTS.mp3 found in playlist:
    file:///media/announceTTS.m3u doesn't exist on disk - skipping!
    [21-05-06 17:32:22.7838] Slim::Formats::Playlists::Base:laylistEntryIsValid (125) Warning: file:///tmp/annTTS.mp3 found in playlist:
    file:///media/announceTTS.m3u doesn't exist on disk - skipping!
    [21-05-06 17:32:22.8153] Plugins::LMSannounce::Plugin:laylistStoppedCallback (262) playlistStoppedCallback called for cmd playlist stop
    [21-05-06 17:32:22.8160] Plugins::LMSannounce::Plugin:laylistStoppedCallback (267) playlistStoppedCallback: playerid=ab:cd:ef:12:34:02 tmppls=tempplaylist_abcdef123402 time=85.8651892604828 power=1 mode=play
    [21-05-06 17:32:22.8166] Plugins::LMSannounce::Plugin:laylistStoppedCallback (276) playlistStoppedCallback: playlist resume parms: ['playlist','resume','tempplaylist_abcdef123402','n oplay:0']
    [21-05-06 17:32:23.0027] Plugins::LMSannounce::Plugin::__ANON__ (278) playlistStoppedCallback: resume done
    [21-05-06 17:32:23.0429] Plugins::LMSannounce::Plugin::__ANON__ (285) playlistStoppedCallback: deleting /tmp/annTTS.mp3
    [21-05-06 17:32:23.0488] Plugins::LMSannounce::Plugin:laylistLoadedCallback (298) playlistLoadedCallback called for cmd playlist load_done

    The warning in-between are the problematic point obviously. I don't know how to fix this the good way...
    My investigation showed that, only if I trigger the LMSAnnounce API once and after that I have to restart the squeezeboxserver services manually again via systemctl, the problem is gone. No warnings nothing, everything works as intended afterwards.


    I already tried to disable the service and start the server by crontab @reboot, but problem remains. After a manual restart of the service is works again... This confirmes also that I have to call the Announce API Once before I restart the service.

    After this it runs stable until I reboot the system.

    Can you help me to figure out what to do to fix this?

    Thank you very much

    Best regards

    Toni
    Last edited by Gozumuli; 2021-05-06 at 09:47. Reason: Formatting

  2. #52
    Senior Member
    Join Date
    Aug 2012
    Location
    Austria
    Posts
    1,213
    Quote Originally Posted by Gozumuli View Post
    The Logfile of LMS shows this:
    Code:
    [21-05-06 17:32:21.5962] Plugins::LMSannounce::Plugin::handleJSONP (210) handleJSONP targetClient=ab:cd:ef:12:34:02
    [21-05-06 17:32:21.6018] Plugins::LMSannounce::Plugin::announceTTS (330) announceTTS: command: /home/pi/.local/bin/gtts-cli -o /tmp/annTTS.mp3 -l de "Ansage: Test-Ansage"
    - verify that the command
    /home/pi/.local/bin/gtts-cli -o /tmp/annTTS.mp3 -l de "Ansage: Test-Ansage"
    can be successfully execute by the user LMS runs as (usually squeezeboxserver)
    - verify that the player ID (ab:cd:ef:12:34:02) is correct
    Various SW: Web Interface | TUI | Playlist Editor / Generator | Music Classification | Similar Music | Announce | EventTrigger | DB Optimizer | Image Enhancer | Chiptunes | LMSlib2go | ...
    Various HowTos: build a self-contained LMS | Bluetooth/ALSA | Control LMS with any device | ...

  3. #53
    Junior Member
    Join Date
    May 2021
    Posts
    2
    Quote Originally Posted by Roland0 View Post
    - verify that the command
    /home/pi/.local/bin/gtts-cli -o /tmp/annTTS.mp3 -l de "Ansage: Test-Ansage"
    can be successfully execute by the user LMS runs as (usually squeezeboxserver)
    - verify that the player ID (ab:cd:ef:12:34:02) is correct
    I investigated a bit further.

    When I execute the Announce request the first time after a reboot the annTTS.mp3 file gets created immediately. I also checked if it is a valid mp3. That means the gTTS-Cli command works as expected.

    I also can confirm that it is the correct MAC address.

    I changed the path for mp3 file from /tmp to /media, so it won't be deleted during reboot.

    Now it seems to work, but it still seems to be somehow a bug for which I now found a workaround.

    I will investigate this further and report back if something is still buggy.

  4. #54
    Junior Member
    Join Date
    Jun 2021
    Posts
    1
    Quote Originally Posted by Cadster View Post
    Thanks for creating a great plugin!
    Works perfectly with Home Assistant.

    Thanks!
    Would you be able to share how you use this plugin with Home Assistant? I've installed it in LMS, but haven't figured out how to trigger it from Home Assistant. Big thanks in advance!

  5. #55
    Junior Member
    Join Date
    Aug 2011
    Posts
    24
    Quote Originally Posted by kassai View Post
    Would you be able to share how you use this plugin with Home Assistant? I've installed it in LMS, but haven't figured out how to trigger it from Home Assistant. Big thanks in advance!
    Here is a link to homeassistant-forum where you can find your answer.

  6. #56
    Junior Member
    Join Date
    Aug 2011
    Posts
    24
    I've updated to the latest nightly and now the saved playlist (before the announcement) does not resume.
    Does anybody, using the latest lms 8.2 nightly, have the same problem?

    the fix: i've restored a backup from januari 2021, updated plugins and lms to the latest nightly and the problem is gone, happy times!
    Last edited by Cadster; 2021-06-22 at 14:49. Reason: solved the problem

  7. #57
    Junior Member
    Join Date
    Jul 2021
    Posts
    12

    TTS engines

    I played around with the plugin. Very nice!

    I would be interested regarding your experience with different tts engines.

    I started with espeak - very fast, but also very robotic sound. German voice support seems ok.
    I then tested the German mbrola files for espeak, which give better results, especially mb-de-4 to 7. I think the output here is acceptable (not perfect of course).

    After this I compiled mimic. Not as fast as espeak. Main problem currently for me is, that I cannot find a German voice. Using the built-in voice (I used default, rms & slt) is inferior to the German mbrola voices with espeak.

    All together, espeak with mbrola is easy to install (already included in the repos) and gives good results. Mimic needs to be compiled and gives bad results for me because of lack of German voice.


    Does anyone know if there is a German voice anywhere for mimic?
    Which engines do you use?

  8. #58
    Junior Member
    Join Date
    Jul 2021
    Posts
    12
    I also tried out larynx.

    Great quality of the generated audio file, lots of languages and voices - but very slow (too slow for my Raspberry...).

  9. #59
    Senior Member
    Join Date
    Aug 2012
    Location
    Austria
    Posts
    1,213
    Quote Originally Posted by Rothaar View Post
    Does anyone know if there is a German voice anywhere for mimic?
    you can find a HTS voice here: https://github.com/m-toman/SALB/blob...a/leo.htsvoice
    Supposedly, mimic can use this format. My install, however, can't use this one - maybe one has to compile mimic with an external HTS engine (and not the bundled one)
    Otherwise, try SALB or flite+hts_engine
    Which engines do you use?
    If larynx (which claims "Faster than realtime performance on a Raspberry Pi 4") is too slow on your Pi, your options will be very limited, as most (all?) modern TTSs use neural networks, which need more processing power than statistical speech synthesis (mimic1, espeak etc.)
    Have you tried using a low quality vocoder with larynx?

    There also is MARY, which I haven't tried myself - it does support German, so might be worth looking into.
    Various SW: Web Interface | TUI | Playlist Editor / Generator | Music Classification | Similar Music | Announce | EventTrigger | DB Optimizer | Image Enhancer | Chiptunes | LMSlib2go | ...
    Various HowTos: build a self-contained LMS | Bluetooth/ALSA | Control LMS with any device | ...

  10. #60
    Senior Member carsten_h's Avatar
    Join Date
    Apr 2019
    Location
    Wunstorf, Germany
    Posts
    631
    Quote Originally Posted by Rothaar View Post
    Does anyone know if there is a German voice anywhere for mimic?
    Which engines do you use?
    I am using Reverso TTS inside my Home Assistant installation and the voice "Klaus-German" and this sound really good.
    Pi4 4GB piCorePlayer with LMS and Squeezelite for USB inside an Argon One case
    Pi3B+ (7" Display, Hifiberry DAC+ Pro) piCorePlayer with Squeezlite/Jivelite for Hifiberry and Bluetooth headphone inside a SmartiPi touch case
    Two Airport Express
    Ikea Symfonisk

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •