Text To Speech as an MSP object?
Hey guys, so, random little project i'm researching, involving text to speech. The latest post I've seen about it suggests using the [shell] object, and surprisingly Masayuki Akamatsu's aka.speech module still works in max 7 - the problem is that I want to further process the audio of the speech using MSP, and the aka.speech (and I assume the [shell] "say whatever" method) literally just force the system to spit out the speech direct to the speakers.
I'd prefer to avoid having to do a bunch of weirdo routing using SoundFlower or Audio Hijack to get the speech back into Max (that seems roundabout, and I kind of need my routing - both virtual and physical - to stay as it is.).
This is for OSX - don't care if it works on Windows at all. any thoughts?
"say" is quite flexible. Type "man say" in the OSX terminal to see a short manual. You can output the text-to-speech to a file of your choosing and load it into your favorite Max object for further processing. Here is my first attempt at it. Have fun!
Hens Zimmerman
would you mind telling me what the 'clear, append ~/blah.m4a' message is doing?
I am having great success with aka.speech. It's 32-bit. I set Max to that and built a standalone. Then I set Max back to 64-bit and sent messages to/and from my app with udpsend/receive. Brought my apps audio back via SoundFlower included in an aggregate device. Weirdo has always worked for me! https://www.youtube.com/watch?v=B5ZvfwJdUCI