mimic1
mimic1 copied to clipboard
MycroftAI
Demo voice
If you compile the sources as described in your manual, the resulting voice not as good as the demo voice here: https://www.youtube.com/watch?v=gvnhcNdXJsk
How to get such a brilliant voice?
The videos from that time period were created using another online Text to Speech tool. You'll notice the text is all forward speaking "we are going to..." type statements. We will be continuing work on creating more natural sounding TTS tools, but in the short term the local engine running on hardware like a Pi will not be able to produce sound like that in real-time. For the time being, that will need to come from a cloud-based voice service.
I had a conversation with RyanLeeSipes a while ago, about the same time that video was released. He told me that the voice in this particular promotional video was a non synthesized recording of Alan Pope (the person whose voice was sampled and used to create the mimic voice). Yeah, what the promotional video portrays and what we actually have don't match. That is advertising for you. Personally I would have made it more clear that this was simulated enactment of the technology and not a demo of the actual technology itself. I wasn't going to say anything since I am not on the Mycroft team, but I do think Spenrod's reply is inaccurate and may cause confusion. (no offense intended, PR is hard)
@Spenrod I think your comment is referring to the old voice like shown in this video?: https://www.youtube.com/watch?v=-c8kfupIbO4 I am a bit confused by your comment because this voice is not actually used in the "Introducing Mimic" promotional video in question. Also, no offense to that other TTS, but the old voice doesn't sound very "brilliant" to me, nor do I see how the old voice can be confused with being identical to the voice actor recording. I am not going to assume any nefarious motive. What likely happened is you looked at the date of the video, saw that it was an old video made during the time when the old voice was used, and then repeated a similar PR message, because I am sure you get the question, "why does the voice sound different in this video", a lot.
TLDR: That brilliant voice is not a demo, PR is hard.
On Tue, Jun 20, 2017 at 1:33 PM, Steve Penrod [email protected] wrote:
The videos from that time period was created using another online Text to Speech tool. You'll notice the text is all forward speaking "we are going to..." type statements. We will be continuing work on creating more natural sounding TTS tools, but in the short term the local engine running on hardware like a Pi will not be able to produce sound like that in real-time. For the time being, that will need to come from a cloud-based voice service.
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/MycroftAI/mimic/issues/126#issuecomment-309848530, or mute the thread https://github.com/notifications/unsubscribe-auth/ACp7HMfWXVdEN2pmd2Fz2HM4HVw-1F_Jks5sGBCKgaJpZM4N-kUG .
