If anyone could record themselves saying "Astiga" here: https://vocaroo.com/ and either send me the MP3 (privately via info@asti.ga ) or the shared link I would be grateful 🙂
Thanks to everyone who has sent me recordings so far, hopefully this will make pronunciation detection better! I'm especially keen to hear from any Dutch/German based users with their recordings...
Ok, Google are going to accept these recordings into the appeal process to change the pronunciation, although they aren't promising that they will actually change the pronunciation.
To kick this off, I also had to resubmit the app with the original pronunciation of "Astiga" with the soft i. This means the action should be available again soon, once they have reviewed and accepted it.
What I think this means is that you will have to pronounce Astiga as in "Ass-tig-a", not as in "Ass-teeg-a". That said, using the actions console the pronunciation is currently being relayed as the latter. I'm not sure if this is some sort of caching problem - I'll check when the action goes live and update this thread.
Google will only start the pronunciation modification process once this has been in production for two weeks, so if anyone else wants to submit recordings please do! The more the merrier!
Does anyone have any feedback on how easy the action is to invoke right now?
Google have taken our recordings and used them to train the invocation of the action. However, I'm still experiencing a few failures. If anyone else is, I could really do with submitting voice recordings of "Talk to Astiga" - you could generate them here: https://vocaroo.com/
BTW, I have been able to successfully play a song on a Google Mini without saying "Ask Astiga" by setting up a custom Music Routine as follows:
When I say to my Assistant: "Rock Out" (substitute with your phrase)
Actions (Try adding your own/custom action): "Ask Astiga"
Google will then say, "What do you want to play?"
PS I just set this up today and could NOT get Astiga to work at all on my device at first, and it seems it was due to the library not being fully synced yet. It might be a good idea to set up some instructions on this for folks.
Yeah - the library has to be synced because we only have a record of song names etc after the sync process.
Thanks very much for this, it's an interesting approach. We definitely need some more docs around this, I'll have a think on the best way of integrating them.
Comments
If anyone could record themselves saying "Astiga" here: https://vocaroo.com/ and either send me the MP3 (privately via info@asti.ga ) or the shared link I would be grateful 🙂
Here's me https://voca.ro/1mx46otL2BWU 😀
Thanks to everyone who has sent me recordings so far, hopefully this will make pronunciation detection better! I'm especially keen to hear from any Dutch/German based users with their recordings...
Ok, Google are going to accept these recordings into the appeal process to change the pronunciation, although they aren't promising that they will actually change the pronunciation.
To kick this off, I also had to resubmit the app with the original pronunciation of "Astiga" with the soft i. This means the action should be available again soon, once they have reviewed and accepted it.
What I think this means is that you will have to pronounce Astiga as in "Ass-tig-a", not as in "Ass-teeg-a". That said, using the actions console the pronunciation is currently being relayed as the latter. I'm not sure if this is some sort of caching problem - I'll check when the action goes live and update this thread.
Google will only start the pronunciation modification process once this has been in production for two weeks, so if anyone else wants to submit recordings please do! The more the merrier!
The action was reinstated a couple of hours ago.
Does anyone have any feedback on how easy the action is to invoke right now?
Google have taken our recordings and used them to train the invocation of the action. However, I'm still experiencing a few failures. If anyone else is, I could really do with submitting voice recordings of "Talk to Astiga" - you could generate them here: https://vocaroo.com/
BTW, I have been able to successfully play a song on a Google Mini without saying "Ask Astiga" by setting up a custom Music Routine as follows:
Google will then say, "What do you want to play?"
PS I just set this up today and could NOT get Astiga to work at all on my device at first, and it seems it was due to the library not being fully synced yet. It might be a good idea to set up some instructions on this for folks.
Yeah - the library has to be synced because we only have a record of song names etc after the sync process.
Thanks very much for this, it's an interesting approach. We definitely need some more docs around this, I'll have a think on the best way of integrating them.