Just a quick question - has anyone had any good experiences with text to speech? Also wondering about multiple languages - any software that does well?
Context wise - I've had pretty terrible experiences, but it has been at least 2 years since I've looked at it closely.
Depends on what you'd like to use it for. I have really good luck using the built-in Mac text to speech engine and a BASH script to generate scratch audio files using terminal. It's relatively easy to setup. The voices are passable but I wouldn't use them for any production work for a couple of reasons:
1) Licensing. Pretty sure this is in the "no-no" area of terms of use.
2) It's impersonal / robotic / lazy. If we really have something to say to folks, I say it's worth saying with a human voice. Preferably a professional narrator.
This isn't to say that TTS isn't really useful to ear-test narration. I use it all of the time. I'm just hesitant to use soul-less synthetic voices to convey important concepts.
However, there is a use case that a human voice has a really tough time accommodating at scale. Let's say I wanted to create a dynamic audio presentation, personalized and updated using a script that referred to the learner by name. Tough / impossible to do with a narrator. This is a case where the synthetic voice can add to the experience. I've been eyeing vocalware for this purpose. The voices sound decent. The downside is you can't produce audio files. The terms of the service are you need to generate a new audio clip each time you use it.
Tools like Ivona do a pretty good job, especially with multiple languages. But the costs add up for these if you're using them for commercial work.
What services and tools have you looked at that delivered a terrible experience?
I've used the TTS built into Captivate in the past - I have my own issues with Captivate itself, but I didn't find the TTS great. Particularly where we were doing a Paramedic course and pronunciation became an issue.
I don't disagree with any of the points you made - I have a client who is concerned about scale, cost and multiple languages, and is concerned that using actual narration will be cost prohibitive.
I'm trying to convince them otherwise, but don't feel I'd be doing my due diligence if I didn't thoroughly examine the TTS options.
Personally, I vastly prefer using something like voices.com, but to generate audio across say 3 or 4 languages quickly might be challenging.
I wonder if reducing the amount of narration and sticking with a human set of narrators would get the job done? Voice is a great attention getting mechanism and can carry contrast into feedback or "hey, check this out" type audio callouts.
We've tended to get folks used to audio coverage for almost all concepts. But I think with the right audience, asking folks to read a bit before they DO isn't asking too much.
For what it's worth, the Ivona voices in languages other than English sound flawless to me:) The English voices are good but noticeably synthetic. When I was researching these a few years ago, Ivona came out on top. It cost around $700 for a studio license with 2 voices. I think voices were $300 or so each when bundled with the initial studio order. Costs could be more now.
Great points Steve. Using narration as an emphasis rather than for all content might be a great way to keep costs down (bundling recording across multiple modules).
I'll check out Ivona - again, a one time cost spread across multiple modules may work well. Thanks again for all your help.
5 Replies
Depends on what you'd like to use it for. I have really good luck using the built-in Mac text to speech engine and a BASH script to generate scratch audio files using terminal. It's relatively easy to setup. The voices are passable but I wouldn't use them for any production work for a couple of reasons:
1) Licensing. Pretty sure this is in the "no-no" area of terms of use.
2) It's impersonal / robotic / lazy. If we really have something to say to folks, I say it's worth saying with a human voice. Preferably a professional narrator.
This isn't to say that TTS isn't really useful to ear-test narration. I use it all of the time. I'm just hesitant to use soul-less synthetic voices to convey important concepts.
However, there is a use case that a human voice has a really tough time accommodating at scale. Let's say I wanted to create a dynamic audio presentation, personalized and updated using a script that referred to the learner by name. Tough / impossible to do with a narrator. This is a case where the synthetic voice can add to the experience. I've been eyeing vocalware for this purpose. The voices sound decent. The downside is you can't produce audio files. The terms of the service are you need to generate a new audio clip each time you use it.
Tools like Ivona do a pretty good job, especially with multiple languages. But the costs add up for these if you're using them for commercial work.
What services and tools have you looked at that delivered a terrible experience?
Thanks Steve!
I've used the TTS built into Captivate in the past - I have my own issues with Captivate itself, but I didn't find the TTS great. Particularly where we were doing a Paramedic course and pronunciation became an issue.
I don't disagree with any of the points you made - I have a client who is concerned about scale, cost and multiple languages, and is concerned that using actual narration will be cost prohibitive.
I'm trying to convince them otherwise, but don't feel I'd be doing my due diligence if I didn't thoroughly examine the TTS options.
Personally, I vastly prefer using something like voices.com, but to generate audio across say 3 or 4 languages quickly might be challenging.
I wonder if reducing the amount of narration and sticking with a human set of narrators would get the job done? Voice is a great attention getting mechanism and can carry contrast into feedback or "hey, check this out" type audio callouts.
We've tended to get folks used to audio coverage for almost all concepts. But I think with the right audience, asking folks to read a bit before they DO isn't asking too much.
For what it's worth, the Ivona voices in languages other than English sound flawless to me:) The English voices are good but noticeably synthetic. When I was researching these a few years ago, Ivona came out on top. It cost around $700 for a studio license with 2 voices. I think voices were $300 or so each when bundled with the initial studio order. Costs could be more now.
Great points Steve. Using narration as an emphasis rather than for all content might be a great way to keep costs down (bundling recording across multiple modules).
I'll check out Ivona - again, a one time cost spread across multiple modules may work well. Thanks again for all your help.
sean
This discussion is closed. You can start a new discussion or contact Articulate Support.