Following on from yesterday’s post, I have used several speech-to-text systems. They all have good and bad points, but here is my experience with them.
The first one I tried was Dragon Naturally Speaking. It was version 9 that I had, it’s now on 15, so there may (should) have been some improvements in that time. I found this difficult to get on with, the training took so long it wasn’t worth the effort in the end. I have a friend that swears by this mind, so it could just be me. The big benefit of this is that you can use it to voice command just about all of your computer, not that I did. The downside is the cost.
I tried Google dictate – it worked, but not great, not well enough to make it acceptable to me. One attempt was enough to put me off.
Then I stumbled across this system called Speechnotes. I’ve used this on my tablet, my phone and via the internet. It works fine, it’s not perfect but good enough for me. There are only two problems I have with this, it doesn’t do speech marks, and “new paragraph” gives two lines for block paragraphs, so I have to remember to say, “new line”. Then of course come the hassle of having to transfer it all into whatever word processing tool I happen to be using. The big benefit is, it’s free.
Word now includes a dictate function. It has the same issue with speech marks, which gets to me because I tend to carry stories on conversations, but it works well enough and no moving into the word processing tool because I’m already in it. Unless I’m using Scrivener. The good thing is that it’s part of Word package, so no extra cost.
What I find is all work best where I’m reading out pre-written passages rather than the stop-start speech that comes from creating on the fly. I also find that if it can’t understand unusual names, so I tend to call characters ‘Bob’ or ‘Dave’ and then do a load of universal changes.
I’m only report my experience, I’d recommend others try a few to find what works for them.