I am trying to get stretch to listen for a certain command, and then initiate a demo. I have it working perfectly fine when I have a monitor, mouse, and keyboard plugged into the robot, however, when I want to operate the robot remotely, it no longer works. I have checked my speaker settings and I am using the remote desktop speakers and output, but that does not work either. The speech is detected for the correct amount of time, the recognition software is just unable to translate to text.
Does anyone have any idea what could be going on? Is there a better way to operate remotely so that the speech recognition will still work? Thank you in advance!!
Hi @jgangemi, in your attached screenshot, there’s a warning about “Failed to find respeaker device by name. Using default input”. Do you see the same warning when the robot is tethered?
Hi @jgangemi, I am yet to reproduce this issue through a remote desktop. But generally remote desktop services tend to access the I/O drivers which might cause issues with the devices attached to the robot. For ROS applications I would recommend using the SSH method to access the robot and see if the respeaker speech recognition works by echoing the /speech_to_text topic. SSH has minimal overhead compared with a Remote desktop server.
Hello Mohamed, thank you for getting back to me! The SSH method does work better for getting speech recognition to work while operating untethered. I will use that for now
