Text to Speech Synthesizer
April 8, 2004 4:01 PM Subscribe
I'm trying to locate a good text to speech synthesizer. I've looked at several but, I'm wondering if you know of any. (More inside).
I'd like to dump text files in a desktop (win32) app and have it spit out mp3/wav files.
Here's what I've tried so far: Oddcast, Freets, IBM TTS, Elanspeech, Realspeak, Oddcast has the best and most realistic but, licensing is remotely hosted and I would really like a desktop app... for now. Any help would be greatly appreciated.
I'd like to dump text files in a desktop (win32) app and have it spit out mp3/wav files.
Here's what I've tried so far: Oddcast, Freets, IBM TTS, Elanspeech, Realspeak, Oddcast has the best and most realistic but, licensing is remotely hosted and I would really like a desktop app... for now. Any help would be greatly appreciated.
Response by poster: I did check that one too... a great deal of the TTS products out there license the AT&T labs one.
posted by Dean_Paxton at 4:15 PM on April 8, 2004
posted by Dean_Paxton at 4:15 PM on April 8, 2004
festival claims to work in windows. I've used it in linux, the output is quite good. Plus its free.
posted by duckstab at 4:24 PM on April 8, 2004
posted by duckstab at 4:24 PM on April 8, 2004
I think that AT&T's Natural Speech is the best engine currently commercially available. I'm not an expert on this.
But I've used it, and I've installed a Trillian plugin that utilized it—and it was a little spooky in that it got the intonations mostly correct.
It's interesting to me that it seems to be the case that thay've given up on pure-phoneme artifical speech synthesis and are now obviously using manipulated sampled speech. It's reminiscent of the shift from synthesized instrumental music to sampled. It doesn't appeal to my sensibilities, it seems like a step backward. But it obviously produces superior verisimiltude.
posted by Ethereal Bligh at 4:59 PM on April 8, 2004 [1 favorite]
But I've used it, and I've installed a Trillian plugin that utilized it—and it was a little spooky in that it got the intonations mostly correct.
It's interesting to me that it seems to be the case that thay've given up on pure-phoneme artifical speech synthesis and are now obviously using manipulated sampled speech. It's reminiscent of the shift from synthesized instrumental music to sampled. It doesn't appeal to my sensibilities, it seems like a step backward. But it obviously produces superior verisimiltude.
posted by Ethereal Bligh at 4:59 PM on April 8, 2004 [1 favorite]
Response by poster: I think you are right in your music analogy and it is a step backwards. But, I'm trying various demos of it now. I'm almost intrigued enough to comitt to a developers license.
I'm really interersted in something from Russia called Sakrament. But, they haven't returned my calls or e-mails yet. Too bad because I've seen a lot and this one is awesome.
duckstab: Thanks for the link to festival, I hadn't seen it yet. This could go places too.
posted by Dean_Paxton at 5:15 PM on April 8, 2004
I'm really interersted in something from Russia called Sakrament. But, they haven't returned my calls or e-mails yet. Too bad because I've seen a lot and this one is awesome.
duckstab: Thanks for the link to festival, I hadn't seen it yet. This could go places too.
posted by Dean_Paxton at 5:15 PM on April 8, 2004
What about speech to text software?
Is there any out there at all?
What would be good would be like you describe - dumping a file into a program and spitting out text files. Even better wd be if it could handle microtape input somehow, like dictaphones.
Any thoughts?
posted by dash_slot- at 5:31 AM on April 9, 2004
Is there any out there at all?
What would be good would be like you describe - dumping a file into a program and spitting out text files. Even better wd be if it could handle microtape input somehow, like dictaphones.
Any thoughts?
posted by dash_slot- at 5:31 AM on April 9, 2004
Best answer: My expert on such matters says:
"I use Natural Voice Reader. It will read a text file to a WAV file and provides a couple o really outstanding voices... way, way better than the one's supplied by MS Agent."
posted by billsaysthis at 7:18 AM on April 9, 2004
"I use Natural Voice Reader. It will read a text file to a WAV file and provides a couple o really outstanding voices... way, way better than the one's supplied by MS Agent."
posted by billsaysthis at 7:18 AM on April 9, 2004
Response by poster: Dragon NaturallySpeaking will do this but, it's a $200 bill for the software.
posted by Dean_Paxton at 7:25 AM on April 9, 2004
posted by Dean_Paxton at 7:25 AM on April 9, 2004
Response by poster: I'm testing natrual voice reader now... the standard version. The enterprise version is only $40 and apparently includes a few voices. If they are like the samples I heard, they aren't too bad, I agree.
posted by Dean_Paxton at 7:41 AM on April 9, 2004
posted by Dean_Paxton at 7:41 AM on April 9, 2004
This thread is closed to new comments.
posted by m-bandy at 4:08 PM on April 8, 2004