Issue with Sound Recognition feature not recording in Scratch
Thống Vũ Trần Nhất
Dear Machine Learning for Kids Support Team,
I am an elementary school teacher currently using your platform to introduce AI to my students. While working with the Sound Recognition feature, I successfully trained the model and tested it directly on the platform without any problems.
However, when I exported the model to Scratch and used the provided blocks as instructed, I noticed that the microphone icon does not appear and the program does not seem to receive any audio input. As a result, the sound detection does not function as expected in Scratch.
Could you please help clarify:
-
What might be causing the microphone not to activate in Scratch?
-
Are there any additional permissions or settings needed for this feature to work?
-
Is this a known issue with certain browsers or operating systems?
I would greatly appreciate any advice or troubleshooting steps you can provide.
Thank you very much for your support.
Best regards,
Nhat Thong
Dale Lane
Thống Vũ Trần Nhất
Hi,
I have a question regarding the sound recognition feature. I'm wondering whether it is capable of accurately recognizing spoken words.
I tested two different cases:
-
In the first case, I trained the model with sounds like clapping, desk tapping, and the spoken letter "A". The system recognized them very well.
-
In the second case, I trained it with spoken letters "A", "B", and "E". However, the recognition was not accurate — it often confused the letters.
Could you please help me understand the reason and suggest a way to improve the recognition of spoken words?
Best regards,
Nhat Thong