> My guess is, if you consider proposing such a device to blind people, sound feedback would be an instant deal-breaker.
I think you're right if the audio takes over the environment. Beeping and buzzing would be especially irritating. But OTOH, echo location is audible, and it's the main way that some blind people "see", so there's already evidence that audio is not a deal-breaker, or doesn't have to be.
> It would be very interesting to try and encode a wider sensing in finer ways: more buzzers, different "tones" of vibration... giving simultaneously information about what's exactly in front of the stick, and what the wider surroundings are like. I wonder what kinds of encodings are easiest to learn.
I agree completely. If you think about echo location, the source sound, a tongue click or whatever, has a wide spread. The sound that goes out echoes off everything nearby, it's not just a point-sample of distance. Furthermore, the echo that comes back has all kinds of texture to it, and blind people often talk about how they can hear the shape and material of things. They're getting multi-dimensional information from the echoes, not just a guess of how far away something is. They get amplitude and reverb/decay and stereo to work with, just for a start. It would indeed be very interesting to explore how a sensor device could achieve this same subtlety and multi-dimensionality!
I think you're right if the audio takes over the environment. Beeping and buzzing would be especially irritating. But OTOH, echo location is audible, and it's the main way that some blind people "see", so there's already evidence that audio is not a deal-breaker, or doesn't have to be.
> It would be very interesting to try and encode a wider sensing in finer ways: more buzzers, different "tones" of vibration... giving simultaneously information about what's exactly in front of the stick, and what the wider surroundings are like. I wonder what kinds of encodings are easiest to learn.
I agree completely. If you think about echo location, the source sound, a tongue click or whatever, has a wide spread. The sound that goes out echoes off everything nearby, it's not just a point-sample of distance. Furthermore, the echo that comes back has all kinds of texture to it, and blind people often talk about how they can hear the shape and material of things. They're getting multi-dimensional information from the echoes, not just a guess of how far away something is. They get amplitude and reverb/decay and stereo to work with, just for a start. It would indeed be very interesting to explore how a sensor device could achieve this same subtlety and multi-dimensionality!