4 ms·

The low WER numbers you've probably seen are for conversational telephone speech with constrained subjects. ASR is much harder when the audio source is farther

The low WER numbers you've probably seen are for conversational telephone speech with constrained subjects. ASR is much harder when the audio source is farther away from the microphone and when the topic isn't constrained.

dharma1 9y ago
Very true. Which is why all the home assistants (Google Home, Amazon Echo etc) use array mics and beamforming - they get a much cleaner speech signal from far field audio, and better WER as a result
- adrianbg 9y ago
  Exactly. This is also why Google sponsors the CHiME challenge, the existence of which is more proof that ASR is pretty far from solved. http://spandh.dcs.shef.ac.uk/chime_challenge/ http://spandh.dcs.shef.ac.uk/chime_challenge/
  - dharma1 9y ago
    Good stuff. Was looking into cheap array mics with linux drivers a few times in the past but not much is available. Speech separation - Mitsubishi Research has done some pretty impressive stuff on that - http://www.merl.com/demos/deep-clustering http://www.merl.com/demos/deep-clustering. Haven't seen equivalents of that in open source ASR
    - adrianbg 9y ago
      Totally. You may want to take a look at the papers from CHiME4 for more along those lines: http://spandh.dcs.shef.ac.uk/chime_challenge/chime2016/results.html http://spandh.dcs.shef.ac.uk/chime_challenge/chime2016/resul... I'm really fascinated by the whole idea of blind source separation and the fact that speech signals are "sparse" in frequency space. We've had a similar experience looking for hardware / open source beamforming. There's a package called beamformit, but I think it's pretty old.
    - j_s 9y ago
      This is the useful info I've seen hardware-wise recently: https://medium.com/snips-ai/benchmarking-microphone-arrays-respeaker-conexant-microsemi-acuedge-matrix-creator-minidsp-950de8876fda https://medium.com/snips-ai/benchmarking-microphone-arrays-r... https://developer.amazon.com/alexa-voice-service/dev-kits/conexant-4-mic/ https://developer.amazon.com/alexa-voice-service/dev-kits/co... Amazon's 7-mic hardware has its own OEM program.

dharma1 9y ago

Very true. Which is why all the home assistants (Google Home, Amazon Echo etc) use array mics and beamforming - they get a much cleaner speech signal from far field audio, and better WER as a result

adrianbg 9y ago
Exactly. This is also why Google sponsors the CHiME challenge, the existence of which is more proof that ASR is pretty far from solved. http://spandh.dcs.shef.ac.uk/chime_challenge/ http://spandh.dcs.shef.ac.uk/chime_challenge/
- dharma1 9y ago
  Good stuff. Was looking into cheap array mics with linux drivers a few times in the past but not much is available. Speech separation - Mitsubishi Research has done some pretty impressive stuff on that - http://www.merl.com/demos/deep-clustering http://www.merl.com/demos/deep-clustering. Haven't seen equivalents of that in open source ASR
  - adrianbg 9y ago
    Totally. You may want to take a look at the papers from CHiME4 for more along those lines: http://spandh.dcs.shef.ac.uk/chime_challenge/chime2016/results.html http://spandh.dcs.shef.ac.uk/chime_challenge/chime2016/resul... I'm really fascinated by the whole idea of blind source separation and the fact that speech signals are "sparse" in frequency space. We've had a similar experience looking for hardware / open source beamforming. There's a package called beamformit, but I think it's pretty old.
  - j_s 9y ago
    This is the useful info I've seen hardware-wise recently: https://medium.com/snips-ai/benchmarking-microphone-arrays-respeaker-conexant-microsemi-acuedge-matrix-creator-minidsp-950de8876fda https://medium.com/snips-ai/benchmarking-microphone-arrays-r... https://developer.amazon.com/alexa-voice-service/dev-kits/conexant-4-mic/ https://developer.amazon.com/alexa-voice-service/dev-kits/co... Amazon's 7-mic hardware has its own OEM program.