Very bad quality of voices on mimic github repository compared to voice of shipping Mycroft device


i cloned the mimic github repository and tried all voices there but those are far away from the one presented in the commercial device Mycroft Mark 1.

I thought this is about open source and open hardware but what about publishing this voice or are there some secret commands to get the existing ones better?

No comments? Is the used voiced for Mark 1 proprietary?

Hi, the same voice is available in mimic as the mark one. In the mimic repo, the one called mycroft_voice_4.0.flitevox is the default mycroft voice.

If you are unsure if it sounds correct you can dump it to a wav-file “mimic -voice PATH_TO_VOICE -t “This is the text that mimic will output” -o out.wav” and upload it somewhere and I can see if it sounds correct.

1 Like

Hello forslund,

i used that voice, sounds too robotic for me while the one from official videos sounds more natural. I will upload a file.

Thanks for the proposal.

The video was made before mimic, instead a remote service was used. For speed and portability mimic was introduced. Higher quality local voices are in the making, not sure if all will be freely available or for subscribers.

1 Like

I got it but to be on the safe side:

I am referring to ‘Mycroft Alpha v0.6 Software Demo’ published on 11 Apr 2016 to be found here

The mentioned voice mycroft_voice_4.0.flitevox for mimic was published on github a year ago:

Same voices, one more robotic:-)

Though mimic is based on festival it can synthesize natural sounding voices? Can you point me to a demo?


there’s another voice you receive when becoming a supporter for 19$/year. Maybe it’s the one you’ve mentioned.

In any case i would like to know / hear what i buy and when supporting. A one time payment would also be appreciated.

At open source projects it’s more supporting than buying IMO. Also I only suggest that, I doesn’t can proof it because I’m not a supporter till now. I think about this option but for this I’ve to dig deeper and get a better understanding for this system/framework.

1 Like

Similar situation here. At least some of my questions get answered and the aim to be an opensource company is visible to some sense.

Without an active community i will not follow any support journey.

I might be misunderstanding you but the voice in the demo you’re pointing at sounds exactly like the one I mentioned above. what mimic command line are you executing?

If the voice isn’t found it will fall back to a voice that’s available. some are more robotic than others. kal for instance is very fast to synthesize but sounds robotic.

I use it this way

./mimic -t “Mozilla is a free-software community created in 1998 by members of Netscape. The Mozilla community uses, develops, spreads and supports Mozilla products, thereby promoting exclusively free software and open standards, with only minor exceptions.” -voice voices/mycroft_voice_4.0.flitevox

Need to check it again, maybe cut corresponding Youtube snippet 1:50 - 2:07 and compare.