NaturalReader

Hi everybody,

The service is HS.

Adding coyright And empty sound file

 

 

 

salut tous le monde,

il semble que le service est HS .

Ajout de coyright

et fichier son vide

Mats's picture

NaturalReader

That was just a matter of time. The advice is to use MarySpeech that is Open Source.

It's at least the third time that non Open source services has stopped working.

First Google, then Acapela and now Natural ( same as Acapela )

moz4r's picture

What do you think about it

What do you think about it :

http://www.voicerss.org/pricing/

java sdk, free to use. and we can pay if we want more.
I will test it and do report

 

kwatters's picture

Mimic Speech from MyCroft AI

Sad, but not un-expected.

So,  I found another open source speech system called "Mimic" from an open source project called MyCroft AI.  It's a native binary, but it does what we want it to.  I started the MimicSpeech service a while back, it's not complete, it currently only has a build for windows  (we can build for other platforms.)  the audio quality is decent.  

I think MarySpeech still is probably the best choice going forward.  If people want to pay for something, we could consider adding Amazon Polly API or some other commercial version.  It would be up to the end users to get their own license for it.

https://github.com/MyRobotLab/myrobotlab/blob/develop/src/org/myrobotlab...

 

 

steveh110's picture

The problem with MarySpeech

The problem with MarySpeech is that for French the voice is really very bad. I think it's good to have a commercial version like "http://www.voicerss.org/pricing/" that will be usable for users who feel like it.

 

Le problème avec MarySpeech est que pour le français la voix est vraiment très mauvaise. Je pense qu’il serai bon d’avoir une version commercial comme « http://www.voicerss.org/pricing/ » qui serai utilisable pour les utilisateur qui ont envie.

moz4r's picture

Hi Kevin ! amazon Polly web

Hi Kevin !

amazon Polly web service is a good idea I think

Doki's picture

I'm working on a Service for MS/Azure

Im already working on MS/Azure (https://www.microsoft.com/cognitive-services/en-us/speech-api)

because i also dont like the quality of MarySpeech.

Its free for 5k calls per month.

I have a prototype version for Text to Speech ...

and when i have time i will also include speech recognition.

When testing speech recognition at MS has a lot better

recognition then the current one from inmoov. 

moz4r's picture

Doki, great idea about speech

Doki, great idea about speech recognition.!!! It can replace webkitspeech the day after shutdown :)
With azure we can repair too translator service.

That great , we will have a lot of tts soon with voicerss + azure + amazon ...

 

 

 

kwatters's picture

Offline and Open Source is better...

I think we should focus on services that don't require being connected to the internet and services that are open source.  

This doesn't mean we can't include other services like google, amazon, or microsoft...  It's just a caution that if you rely on services that require the internet, your robot won't work properly if you're not connected to the internet.

 

 

Doki's picture

Offline and Open Source is

Offline and Open Source is better when the Quality can keep up with it, i fully agree. 

And sure if you rely on the internet it has to be availible. 

But like for most Builders Inmoov is part of my Home/Garage and there i have a fast and stable internet connection. Maybe this changes when Gael publish the legs :)

For me it is more annoying when he doesnt listen :)

steveh110's picture

service VoiceRSS

Hello friends

I realized a service for VoiceRSS

I can not add it to github. If someone can make a service thank you

The files are here: https://github.com/steveh110/myrobotlab/tree/master/java thank you in advance

 

 

hello les amis

j'ais réalisé un service pour VoiceRSS

je n'arrive pas a l'ajouter sur github. si quelqu'un pourais en faire un service merci.

les fichiers sont ici : https://github.com/steveh110/myrobotlab/tree/master/java

merci d'avance

moz4r's picture

Thank Steve I test to

Thank Steve I test to implement it , it will be a good exercise !