Text to Speech

[jc2006]

Has anyone used any text to speech programs? Any recommendations

[dmozealous]

I have not personally used it much, but I know TextAloud is pretty popular for text to speach.

[noah]

Hi,

We recommend Acapela Virtual Speaker, one of the best text to speech engine

http://demo.acapela-group.com/

Mor information in German:
http://www.partekk.ch/Produkte.asp?id=406

Tel us if you need price information.

Regards

[Alcoakid]

I'm looking into this also and found about 5 or 6 different software packages that can do this. Just search the web. What I've learned is that the voice is the determining factor not necessarily how the software works. I've just down loaded a package and 2 voices from two different suppliers.

I would suggest listening to different voices with a sample of your text and down loading any demos you can.

[andrewkay]

We have used TextAloud (NextUp.com) here for about a year now. It's not perfect by any means, but I doubt that any text-to-speech engine is.

It handles things like short bullet points well, and it's not until you get into long sentences or paragraphs that it starts to sound, well, mechanical.

TextAloud do offer a great range of voices and accents that are all pretty realistic and when we use them we alternate between male and female voices as that seems to be easier on the ear. One of the nice things in TextAloud is that if the engine has problems with pronounciation, which it always does with technical phrases, you can spell out individual words phonetically to get it to pronunce them correctly. I am pretty sure they offer a 30 day free trial but you will need to spend more $$ to get the "good" voices (the ones we use are the NeoSpeech voices, but the Acapela ones sound very good too).

[Spectre]

This is an innocent question: why would you use text-to-speech software for narration vs. a good "human" voiceover? Maintenance? At least in the samples I heard in a few presentations here, I found the text-to-speech narration to sound somewhat automated, while lacking the warmth and inflection inherent in a good human voiceover. Maybe I'm just old school, but I'm curious...

Thanks,
Phil

[andrewkay]

Originally Posted by Spectre

This is an innocent question: why would you use text-to-speech software for narration vs. a good "human" voiceover? Maintenance? At least in the samples I heard in a few presentations here, I found the text-to-speech narration to sound somewhat automated, while lacking the warmth and inflection inherent in a good human voiceover. Maybe I'm just old school, but I'm curious...

Thanks,
Phil

The short answer is that we don't use either one exclusively.

If there is plenty of time to spare and there is someone available to do a voiceover we'll use them. Unfortunately we do not have the $$ in our budget to go out and use professional voice talent. If we do voiceovers here it has to be somebody that works for the company and has the ability (and the time to spare, which is usually the more difficult thing to get) to sit down and do voiceovers in a totally non-professional environment. This can be very time consuming and if the end-client needs changes made then we have to try and duplicate the same setting with the same person all over again.

The Test-to-speech engine does work quite well with short bullet point type slides and if you vary the voices used it can be fairly effective. We probably wouldn't use it for a one hour online lesson, but at times for quick online presentations that we need to get completed fast it works (within its own limitations). The text engine never gets tired, never complains and never wants a break!

Having said all that, it definitely has lot of shortcomings although I am sure that in a couple of years it will be very close to human speech. It is also a fast way to make courses compliant for blind users.

[Spectre]

Andrew,

I appreciate the detailed explanation...very insightful! I certainly understand your challenges regarding recording...in addition to the challenge of finding an adequate recording space (I often have to do it at home!), it can be time-consuming and requires effort to get a top-notch recording. I can see one big advantage of text-to-speech is from a maintenance/update perspective: I don't have to schedule the same voice talent again and worry about the audio sounding different. I guess it's a matter of tradeoffs.

Thanks again for sharing!

Phil

[Alcoakid]

We just purchased TextAloud and will use it where appropriate. I've talked to other elearning developers who use text to speech exclusively. Their customers have agreed on this method of naration, citing some of the advantages you all have mentioned.

Just last month I wished Paul 16K had recorded the narration for a course that needed revisions. We ended up recording all the audio again because the original voice talent had move on to better things. We really didn't have time to re-record but had no choice.

I'm working on a new project now that I've used TextAloud for some samples to see how the customer likes the sound. Maybe they'll go for a computerized narrator.

[andrewkay]

Originally Posted by Alcoakid

Just last month I wished Paul 16K had recorded the narration for a course that needed revisions. We ended up recording all the audio again because the original voice talent had move on to better things. We really didn't have time to re-record but had no choice.

FWIW, we have found that the female voice sounds more realistic than the male.