Text-to-Voice Services

Hello Emily and welcome to Heroes!

What makes you think TTS engines (and voices) don't change?

I would strongly recommend you think about this again, especially if the only reason against using an in house voice talent is consistency.

No TTS software is even remotely perfect IMHO. Most of them are wildly inconsistent to say the least. Even in the same sentence. It is distracting and a real disservice to what could otherwise be a nice elearning module.

Nothing in our world is consistent - even Apple changed Siri's voice (oh the monstrosity!)...

It such need arises, it will be much easier to replace the audio within a *.story file and adjust some timings on the slides, than doing the same with a computerized voice.

Just my 2c,

Alex

Alex,

Good point. I am running into that exact problem as I look into this more... all of the voices sound robotic and impersonal. Do you do your own voice overs for trainings? Has it worked for you?

I regularly use text-to-speech software for scratch audio. This provides an "ear sense" of how something will sound before we spend time and money producing human narration. 

For some things, robo-voice can work well. I think for most stuff, however, it's a little bit inhumane to read text to participants by machine. Voiceover is an opportunity to offer expression and explanation - it injects some humanity into the packaged machinations of an e-learning program. You loose the essence of both of these things by automating the output. Makes something that's already a little plastic feel... more plastic.

That said, voices are getting better. And TTS voices can be really useful in some cases and for some audiences. Here are a couple that I like:

1) https://www.vocalware.com/

Vocalware offers an API for generation of audio and a price that is relatively reasonable (quality TTS is really expensive). Using the JavaScript API you could do some neat things with dynamic personalized feedback right inside of Storyline. Opens some possibilities that just aren't there for pre-recorded audio. Vocalware is essentially the TTS engine behind the oddcast service (http://oddcast.com/demos/tts/tts_example.php?ttsapi)

2) http://www.ivona.com/us/

Ivona's voices are pretty nice compared to many others. Be aware that a commercial license for use of the voices is pretty spendy. I've used the reader ($60) for scratch audio. Since it's for a limited audience and it's not really public broadcast, it's fair to gray area. If you're broadcasting to many users, better get a commercial license. 

3) The voices built-in to the Mac if you're using Apple desktops or laptops. Using terminal, you can batch voice generation. I even wrote some PHP to generate via a web server at my last job. These work really well for scratch audio. I'd be careful about using these for commercial purposes as I don't think that would fit the licensing. But for temp voices, like the Ivona solution, it's not really public. 

Hi Emily and welcome to E-Learning Heroes! Looks like you are getting some fantastic assistance here from Alex and Steve

Thanks guys!

Thanks for the insight Steve. I will look into the sites and services you recommended!

Sounds great Emily. Thanks for the update and good luck in your search

You can actually use google translate to create english or foreign language audio files. I wrote about it here: Link

I had not seen that thread before Owen, thanks for sharing

Steve or anyone else with TTS experience:

In your experience with Vocalware, are you able to use their audio streams to generate files to be used in Presenter? Or are you using it solely as an ear check like you said? It seems that they provide single playback audio and are not really suited for producing audio files for use in training modules. We have some modules that are translated into Chinese. We'd like to add the Chinese audio and are looking for the cheapest way. Do you think a TTS vendor like Vocalware would work for my situation?