"listen" options

Common Voice
#1

In addition to a “yes” and “nope” option, it would be nice to have some of the following options:

  • spam/offensive
    • can be quickly verified by a few votes
    • the entire phrase can be discounted
  • missed/misspoken words
    • requires more verification
    • the entire phrase isn’t discounted
    • may also indicate cultural differences
      (e.g. “until” pronounced as “til”;
      “ask” pronounced as “axe”, etc.)
  • I don’t know if a pronunciation is correct
    (with an input field for the word(s) with the unknown pronunciation)
    • the entire phrase isn’t discounted

As well as some additional descriptive options:

  • native sounding
  • non-native sounding
  • accent (with an input field for more detail)
  • male sounding
  • female sounding
  • gender-neutral sounding
  • I want to hear this person narrate an audiobook!

These would still be used for general analytics, but if someone wanted to generate a voice, then they could select the appropriate subset of characteristics.

#2

There are two separate and interesting suggestions here:

1.) Allow more than just “Yes” or “No” answers to the listen options. Perhaps some of the problem here is that we aren’t clear about what to do all situations. See also:
https://github.com/mozilla/voice-web/issues/272
https://github.com/mozilla/voice-web/issues/273

2.) Allow people to semantically tag the data to the best of their ability.

^^ This is a really interesting idea, and I think it’s something we will want to explore as we develop the site further. Any chance you could file a github issue for this?

Thanks @metasean!