I’m using the ReSpeaker v2 4 mic array with pixel ring. I’m using the ROS implementation and Python 3 on a raspberry pi 4 with ubuntu impish and ROS Melotic. Localization coordinates show up just fine and it publishes messages when I’m speaking, but the speech to text data is always empty. I changed the language to en-US for English with no effect. Anyone have any suggestions?
You can give this method a try: Recognizing Speech with a Raspberry Pi - Hackster.io
Thanks! That’s a great suggestion, but I’m actually finding the issue is not with the speech recognition per say but with the raw audio. All the raw audio is noise. Any suggestions there?
Have you tried any software like ‘Audacity’ or any other online services that filters noise from audio files?