IBM Dr Watson Text to Speech

Dolby

Honorary Master
Joined
Jan 31, 2005
Messages
39,177
Reaction score
6,150
I want to upload a YouTube video without my voice. So I started looking at text-to-speech and seems Dr Watson is best (open to other suggestions?).

But how do I add breaks or pauses? Commas don't seem to do the trick.

Wondering if anyone else uses it
 
Perhaps they have one of these sexy voices that people cannot listen to without going on an imaginary journey first?
 
Perhaps they have one of these sexy voices that people cannot listen to without going on an imaginary journey first?
That's it... It is just too sexy for most.

But jokes aside :

*I tend to do lots of 'er' and 'um'... So are plenty retakes
* I don't have the hardware or decent microphone. TTS means I don't need to buy anything additional
 
You can use SSML markup to indicate breaks. Here's an example (3s = 3 seconds):

String text = "<speak version='1.0'>This is some text <break time='3s'/>" +
"This is some more text</speak>";
 
You can use SSML markup to indicate breaks. Here's an example (3s = 3 seconds):

String text = "<speak version='1.0'>This is some text <break time='3s'/>" +
"This is some more text</speak>";
I don't even know what this is :/

Was hoping something I could just type - but I'll try read up
 
I did a Video to Text one (transcibe)


Out there stalks the evil that is Watson Node-Red.

With Watson, we built a bot for Platinum Wealth that takes user's post data and creates a "personality" around the user. Combining my love for psychology and tech.

D0Gy3-GWkAA7saf-1.jpg

This does not solve your issue though :P - so try Polly, she's amazing.

https://aws.amazon.com/polly/

Maybe I will do a blog post
 
I did a Video to Text one (transcibe)


Out there stalks the evil that is Watson Node-Red.

With Watson, we built a bot for Platinum Wealth that takes user's post data and creates a "personality" around the user. Combining my love for psychology and tech.

View attachment 1130714

This does not solve your issue though - so try Polly, she's amazing.

https://aws.amazon.com/polly/

Maybe I will do a blog post
+1 for Polly.

Or try these guys https://replicastudios.com/
 
That's it... It is just too sexy for most.

But jokes aside :

*I tend to do lots of 'er' and 'um'... So are plenty retakes
* I don't have the hardware or decent microphone. TTS means I don't need to buy anything additional
Write a script and read it. This is exactly what you're going to do with a text to speech thing anyway so not sure why ums and ahs and retakes are an issue. Decent mic under R1000.

Depends on your channel I guess, but my feeling is that people prefer listening to other people rather than digital voices unless the synthesis is really good.
 
Don't. HATE ai voice. Learn hoe to do without er and em. It's okay to pause. You get better at it with time.
Especially tutorials...if it's an obvious digital voice I don't even bother. There's just so much content choice right now that it's easy to click away to something that sounds more interesting / less robotic.
 
Especially tutorials...if it's an obvious digital voice I don't even bother. There's just so much content choice right now that it's easy to click away to something that sounds more interesting / less robotic.
If it is for work or something... Sure.. Shame though. But the moment a youtube video starts with an AI voice or those American advertising type voice.... Skip.
 
If it is for work or something... Sure.. Shame though. But the moment a youtube video starts with an AI voice or those American advertising type voice.... Skip.
I also watch the majority of tutorials at 1.3x - 2x speed and the digital voices sound kak / my ears aren't used to it at those speeds. Humans I can deal.
 
Top
Sign up to the MyBroadband newsletter
X