iOS 10: Live Speech Recognition | Ray Wenderlich

🎓 Courses & Articles
1

Learn how to use the new speech framework to do live speech recognition in this iOS 10 screencast.

This is a companion discussion topic for the original entry at https://www.raywenderlich.com/4957-ios-10-live-speech-recognition
2

Hello Sam, great screencast. At the beginning of your tutorial, you mention about Gangstribe app that was done in a previous screencast. I was trying to find that specific screencast with no success, can you point me where it is? Thanks

3

Here you go:

https://videos.raywenderlich.com/screencasts/ios-10-audio-file-speech-transcription

You can see a list of all RW screencasts here:

https://videos.raywenderlich.com/screencasts

sam

4

Hi Sam, great video! Do you have a video podcast where you talking about building the face detect not the speech to text? Thanks!

5

I’m afraid not, no. But all the code is in there for you to take a look at. I don’t think it’s overly complicated?

I’ll put this on the list of suggestions for upcoming screencasts.

sam

6

Hi Sam,
thank you for this video, it’s a great one!
I’ve tried to compile and run the materials but I’ve got a lot of errors.
I use XCode Version 8.0 (8A218a), the last one, and I thing it uses Swift 3.
Should I go back to Swift 2? If yes how?
Thank you for your help
Best
Olivier

Screen Shot 2016-09-17 at 11.07.11.png762Ă—1088 189 KB

7

Hi @olive22

The sample project that accompanies this video was created in a previous beta of Xcode 8 - so had a beta version of Swift 3.

You have a couple of options:

  1. Fix each of the errors that the compiler is complaining about. They’re usually not that challenging to fix - often the automatic “fix-me” is along the right lines.
  2. This project also appears in the iOS 10 by Tutorials book, which as a subscriber you have access to. You can get early access versions of this right now, and when the completed book is launched, all the projects will have been updated to work with the released version of Xcode 8, and hence Swift 3.

Hope that helps

sam

1 Like
8

Hi Sam,
How should I do if I want the speech recognizer to recognize more than just one unique word ?
For example “a dog”, “some sugar”, “a small cat” etc?
Thank you for your help``
Olivier

9

Hi @olive22

If you know the exact phrase you’re after, then rather than checking the last words one-by-one, you’d instead have to search the array for a known subsequence. There are various approaches you could use to tackle this, the simplest being searching for a single word in the phrase and then checking either side of the array to see whether there’s a match.

This feels a bit unpleasant to me, but as far as I’m aware, this is the only real way of tackling this right now.

Hope that helps

sam

1 Like