Pros and Cons of Voice Typing - osprey-voice/osprey GitHub Wiki

Pros

• It works for people who can't do regular typing
• It introduces a new way of thinking about issuing computer commands that is not restricted to a keyboard and mouse
  • provides a large command space
  • commands can be issued more naturally with language
  • allows for using voice macros for issuing repetitive commands
• Issuing commands can be relatively fast if commands are chained together
• Dictation can be very fast especially when using a good model

Cons

• It's different and takes some getting used to
• It's really only usable when you're in a private space
• You can't type and talk with somebody at the same time
• You need a decent microphone
• You need to be running a voice typing program for it to work
• You need to fluently speak one of the languages that an engine provides models for
• Speech recognition gives nondeterministic results compared to regular typing
• The speech recognition and voice typing communities and efforts are somewhat divided by language
  • speech data and models are divided by language
  • voice typing programs need to add support for each language
• Open source speech recognition is still somewhat in its early stages
  • Most engines only provide models for a few languages
  • Training a model can be difficult and can require a lot of computation and data
  • Only a few languages have sufficient speech data for training a model
  • Models can take up a lot of memory