Hacker News Reader

Comment by coredog64 | Hacker News Reader

Several years ago I was trying to solve the problem of voice recognition in a noisy environment. After brainstorming with some colleagues, we came up with the idea of using a throat mike. The only problem was that we couldn't find reasonably priced hardware that would work with commodity computers.

View on HN

Jawbone Bluetooth headsets actually contact your jaw to use the conducted vibrations to cancel background noise, similar to your idea. Disclosure: I worked on their first generation Bluetooth headset. I'm not certain the current generations still work this way.

Noiseassassin - yup, the later gen ones do as well, but as I understand it, Jawbone is basically dead at this point.

Could get a second wind with voice recognition because the problem with comms is everyone sounds the same through that style of mic.

I had that one. It was really great in noisy environments... unless there were also wind. Wind outright killed the sound.

What the heck? I found cheap ones on Amazon designed for the Baofeng ham radios. Surely Shenzhen has figured it out.

What really needs to happen is for someone to figure out how to decode the electrical signals generated by subvocalization, train a DNN to interpret them, and use that.

I don't think that is practical without invasive surgery. Afaik the best we can do with external monitoring is roughly tell whether a person is trying to move or thinking about moving. Reading thoughts doesn't seem realistic with current technology.

If you're trying to tap the brain that's true, but if you could readout the muscular contractions around the vocal chords, you could translate that instead. It's definitely easier to understand the simple output at the level of muscles, than reading "thoughts" in the brain. There are already some prosthetic limbs which used implants to read muscular contractions to control the prosthetic, but those are far too crude to interpret speech.

Reply on news.ycombinator.com