American Male voice - fallback to Allan Pope

andlo · October 20, 2020, 5:46pm

On a newly installed Picroft on a Pi4 he voice falls back to Allan Pope mos of the times. If using femaile (google) voice it never does.
I cant figure out why…Did look into network and power issues bu nothing to see and nothing seems to help.
Anyone having same issue or have any ideas for wha to do ?

Dominik · October 20, 2020, 6:18pm

Hi Andlo, how does “tts” section of your mycroft.conf does look like?

andlo · October 20, 2020, 6:23pm

It is a plain vanilla picroft…
/etc/mycroft/mycroft.conf
´´```
{
“play_wav_cmdline”: “aplay %1”,
“play_mp3_cmdline”: “mpg123 %1”,
“enclosure”: {
“platform”: “picroft”
},
“tts”: {
“mimic”: {
“path”: “/home/pi/mycroft-core/mimic/bin/mimic”
}
},
“ipc_path”: “/ramdisk/mycroft/ipc/”
}

andlo · October 20, 2020, 6:26pm

He do uses american voice, but sometimes, most times he falls back to Allan Pope voice.
Like
Hey Mycroft - who is abaraham linkoln ?
(American voice) " Abrahim Linkoln is…"
Hey Mycroft - what time is it
(Allan Pope) “It is seven twenty three”
Hey Mycroft - Where is the ISS ?
(American voice) “The international spacestation is located at…”
Hey Mycroft - how is the weather ?
(Allan Pope) “todays forecast is…”

j1nx · October 20, 2020, 6:47pm

More inportantly! Welcome back😉

andlo · October 20, 2020, 6:49pm

Glad to be home again but still not enouh time But winter is comming and then hopefully I can have some fun again

Dominik · October 20, 2020, 6:56pm

Maybe you can try so set Google-TTS explicitly:

"tts": {
    "google": {
      "lang": "en"
    },
    "module": "google"
}

andlo · October 20, 2020, 6:58pm

Well - he works when using google voice set from home.mycroft.ai. The problem is when using the american voice (mycroft mimic2). And I do like a male voice. But I dont like Allan Pope and not like that he uses some of the one and some of the other

andlo · October 20, 2020, 7:01pm

But maybe I should give OpenVoiceOS try

j1nx · October 20, 2020, 7:19pm

Try deleting the TTS cache, if then you only get the mimic1 voice perhaps the issue is as simple as a dns problem not be able to reach te servers.

Btw, by the time you are fully back we should be able to give you a fancy OVOS build😜

andlo · October 20, 2020, 7:52pm

Where is the tts cache located ?

I did look at networkissues, but didnt find anything that cold be problem.

I will try my Pi3 board and see if it could be the Pi4 tht has some hardware issues.

baconator · October 21, 2020, 7:35am

cache is under /tmp/mycroft/
Sounds like it’s timing out somehow and coming back at odd times. But audio.log should have something about that somewhere if so?

andlo · October 21, 2020, 7:39am

Ohhh yes - my audiolog tells me an error 500

Cleaing the tts temp didnt change anything.

andlo · October 21, 2020, 7:43am

But it isnt consistant - right after it didnt error 500…

baconator · October 21, 2020, 7:53am

Hence the back and forth. Odd.

J_Montgomery_Mycroft · October 23, 2020, 6:27pm

Could be our TTS server is borking and giving out 500 errors. Anyone else having this trouble?

andlo · October 23, 2020, 7:43pm

I just set up one more Picroft on a Pi3 board to make sure it wassnt he Pi4 board that had some issues. But same behavior. So if it isnt my network which I have checked in and out it points to the Mycroft TTS server which also is the one giving the 500 error.

J_Montgomery_Mycroft · October 23, 2020, 8:00pm

@gez-mycroft - can you take a peek at our TTS server and see if it has fallen on it’s head or if someone has scripted a TTS transcription of The Holy Bible or something?

baconator · October 23, 2020, 8:40pm

And if so replace the text with something from Lovecraft or Herbert.

gamersat678 · October 25, 2020, 4:11am

Not sure if this helps or even related but I am using the American male voice as well and it loves to switch back to Allan pope. `Setup was working fine before I did the latest pull. Was using code back in February before that pull.

2020-10-23 19:06:04.641 | DEBUG    |  3890 | urllib3.connectionpool | https://mimic-api.mycroft.ai:443 "GET /synthesize?text=How%20long%20of%20a%20timer%3F&visimes=True HTTP/1.1" 200 70680
19:06:04.641 - urllib3.connectionpool - DEBUG - https://mimic-api.mycroft.ai:443 "GET /synthesize?text=How%20long%20of%20a%20timer%3F&visimes=True HTTP/1.1" 200 70680
2020-10-23 19:06:08.141 | DEBUG    |  3890 | mycroft.audio.audioservice:_restore_volume_after_record:342 | No audio service to restore volume of
2020-10-23 19:06:08.879 | INFO     |  3890 | mycroft.audio.speech:mute_and_speak:127 | Speak: I'm starting a timer for twenty five minutes
2020-10-23 19:06:08.880 | DEBUG    |  3890 | mycroft.tts.mimic2_tts:get_tts:232 | Generating Mimic2 TSS for: I'm starting a timer for twenty five minutes.
2020-10-23 19:06:08.956 | DEBUG    |  3890 | urllib3.connectionpool | https://mimic-api.mycroft.ai:443 "GET /synthesize?text=I%27m%20starting%20a%20timer%20for%20twenty%20five%20minutes.&visimes=True HTTP/1.1" 200 139201
19:06:08.956 - urllib3.connectionpool - DEBUG - https://mimic-api.mycroft.ai:443 "GET /synthesize?text=I%27m%20starting%20a%20timer%20for%20twenty%20five%20minutes.&visimes=True HTTP/1.1" 200 139201
2020-10-23 19:06:16.163 | INFO     |  3890 | mycroft.audio.speech:mute_and_speak:127 | Speak: An error occurred while processing a request in Timer Skill
2020-10-23 19:06:16.164 | DEBUG    |  3890 | mycroft.tts.mimic2_tts:get_tts:232 | Generating Mimic2 TSS for: An error occurred while processing a request in Timer Skill.
2020-10-23 19:06:16.257 | DEBUG    |  3890 | urllib3.connectionpool | https://mimic-api.mycroft.ai:443 "GET /synthesize?text=An%20error%20occurred%20while%20processing%20a%20request%20in%20Timer%20Skill.&visimes=True HTTP/1.1" 500 290
19:06:16.257 - urllib3.connectionpool - DEBUG - https://mimic-api.mycroft.ai:443 "GET /synthesize?text=An%20error%20occurred%20while%20processing%20a%20request%20in%20Timer%20Skill.&visimes=True HTTP/1.1" 500 290
2020-10-23 19:06:16.260 | ERROR    |  3890 | mycroft.audio.speech:mute_and_speak:131 | Backend returned HTTP status 500
2020-10-23 19:06:16.261 | DEBUG    |  3890 | mycroft.audio.speech:mimic_fallback_tts:146 | Mimic fallback, utterance : An error occurred while processing a request in Timer Skill
2020-10-23 19:23:28.223 | DEBUG    |  3890 | mycroft.audio.audioservice:_restore_volume_after_record:342 | No audio service to restore volume of
2020-10-23 19:23:29.250 | INFO     |  3890 | mycroft.audio.speech:mute_and_speak:127 | Speak: The timer for twenty five minutes has seven minutes thirty nine seconds remaining
2020-10-23 19:23:29.251 | DEBUG    |  3890 | mycroft.tts.mimic2_tts:get_tts:232 | Generating Mimic2 TSS for: The timer for twenty five minutes has seven minutes thirty nine seconds remaining.
2020-10-23 19:23:29.254 | DEBUG    |  3890 | urllib3.connectionpool | Resetting dropped connection: mimic-api.mycroft.ai
19:23:29.254 - urllib3.connectionpool - DEBUG - Resetting dropped connection: mimic-api.mycroft.ai
2020-10-23 19:23:29.588 | DEBUG    |  3890 | urllib3.connectionpool | https://mimic-api.mycroft.ai:443 "GET /synthesize?text=The%20timer%20for%20twenty%20five%20minutes%20has%20seven%20minutes%20thirty%20nine%20seconds%20remaining.&visimes=True HTTP/1.1" 500 290
19:23:29.588 - urllib3.connectionpool - DEBUG - https://mimic-api.mycroft.ai:443 "GET /synthesize?text=The%20timer%20for%20twenty%20five%20minutes%20has%20seven%20minutes%20thirty%20nine%20seconds%20remaining.&visimes=True HTTP/1.1" 500 290
2020-10-23 19:23:29.591 | ERROR    |  3890 | mycroft.audio.speech:mute_and_speak:131 | Backend returned HTTP status 500
2020-10-23 19:23:29.592 | DEBUG    |  3890 | mycroft.audio.speech:mimic_fallback_tts:146 | Mimic fallback, utterance : The timer for twenty five minutes has seven minutes thirty nine seconds remaining