search

leeenglestone / UWP-AnimalNameSpeechRecognition

A fun game for young children that shows a photo and name of an animal when heard on the microphone
MIT License
6 stars 3 forks source link

Speech recognition not continous #2

Open CSharpTeoMan911 opened 3 years ago

CSharpTeoMan911 commented 3 years ago

Your speech recognition application does not have a totally continous recognition. MIT license.... really? I should have an Oxford license. You flow their servers with requests and your internal architecture and boolean logic doesn't cover any possible scenario. Try using the speech completed event in order to reactivate the engine. You nullify and dispose the speech recognition object without verifying if it is disposed or not. You create to many scenarios where memory leaks could occur. And you don't cover all the scenarios where the speech recognition can be re-enabled safely. You don't have a recursive loop where you re-enable the application window in order to ensure that the speech recognition engine stays on, even if the window is disabled(which cannot happen if you have a recursive loop that re-enables the application. ).

leeenglestone commented 3 years ago

Thanks for your feedback @CSharpTeoMan911

I look forward to your pull request should you feel the need to fix some of the shortcomings in my code you identified.

CSharpTeoMan911 commented 3 years ago

No worries..... I pointed that out because some people will use it as guidance and they'll remain stuck........ the current state of UWP continuous speech recognition is broken. I managed 1.5 or 2 months ago to make it continous. One month ago...... After I made a successful implementation of it in an an A. I. assistant reseted my computer and lost the project due to some technical issues and re-implemented the same technique in another A. I. assistant it didn't worked. I tried everything that it was in my power to make the A. I. recognise the words continously, without the use of a button, but it didn't worked. It's Microsoft's fault, not yours. The way I managed to make a semi-continous speech recognition system for my A. I. was made through a combination of in-proc and non-continous on-line speech recognition. This an friendly advice given from me that I hope will help you in your future projects. Good luck.

On Tue, 21 Sep 2021, 18:07 Lee Englestone, @.***> wrote:

Thanks for your feedback @CSharpTeoMan911 https://github.com/CSharpTeoMan911

I look forward to your pull request should you feel the need to fix some of the shortcomings in my code you identified.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/leeenglestone/UWP-AnimalNameSpeechRecognition/issues/2#issuecomment-924185152, or unsubscribe https://github.com/notifications/unsubscribe-auth/AUZUCHT6L6HCM4EDVU5IDM3UDC3WZANCNFSM5CPKAZIQ . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.