Digital Workshop

Is there any way to add text to speech to my publication? I made a stupid assumption when I saw the Autonarration feature, that it was using tts, but when I went to try it out, I found it was just synching a sound file to the text. Does anyone know if there is a plugin or DLL or some way to do this?

I investigated this for a series of children's electronic books I'm building in Opus.

The best I could find were "Natural Voices" and Loquendo - you'll have to do a search on the web, as I have junked the information I had. I was impressed by the quality they could provide and the variety of male/female voices. I was not impressed by the price. I reckon I could have afforded to hire the Royal Shakespeare Company for a month and a recording studio for the same price!

Anything that's cheap - sound cheap, tinny and robotic. I would be very happy for anyone to point me in the direction that confounds my opinion!

I finally solved my problem with the aid of drama students and a lot of auditioning, and a professional studio. I didn't find my own voice very convincing!

Hope this helps.

John

Hi John

Thanks for your reply. Yes, I know they are expensive. I have researched a number of sources and also found that Natural Voices was the best for the price (runtime at $7.50 per copy, min 200). I also looked at Loquendo, even met with their chaps!

However, my problem is more technical. I would only use TTS to offer a cheap alternative to recording sound files. How do I use TTS within Opus? Any ideas?

Cheers

Eleonora

Look at Sayzme from http://sayzme.sourceforge.net.
It's a free (open source) text reader that will automatically read anything in the clipboard.
You can easily set a button to copy text in your talking book to the clipboard and use a button to speak it out.
We've been investigating it's use for creating our talking books in school. The added complication is that our books need to be switch /overlay board accessible and the text needs to be symbol supported.
But both are not too difficult with Opus.

Using text to speech does cut down the time and effort involved in creating accessible talking books; recording all those sound files, saving them under funny names and then losing them.
The quality of the voices although not up to the latest natural voices is generally ok. You cannot obviously use the text highlighting facility with text to speech, but there was a program I saw the other day demonstrated by Inclusive technology that allowed you to save synthesised speech as a wav file, which you could then use text highlighting on.
Perhaps it has posibilities but it does seem a long way round.

One thing I can't get Opus to do is to launch sayzeme from the start, I have to manually lauch it first (although I suppose it could easily be put into a run file -though not by me. I tend run a mile from anything like real programming that includes curly brackets, colons and odd squiggles)

Richard Walter

This is a bit dated. They say you need to install SAPI 4, but current technology uses SAPI 5 and I believe they are incompatible. I'm reluctant to install SAPI 4 and then find that TTS that comes standard with Windows XP no longer works.

Thanks anyway for the tip.

I too have done a lot of recent research on TTS. In relation to SAPI, this is what I found on a few web sites:

Some TTS engines use SAPI4 and some use SAPI5. Some web sites advise that a SAPI4 voice will only run on a SAPI4 TTS engine, while a SAPI5 voice requires a SAPI5 TTS engine.

I downloaded and tried several evaluation TTS engines, and settled on TextAloud for no reasons other than it was reasonably priced, was simple to use, came with an extensive printable manual, and fitted into the way I work. I made the mistake of ordering 1 Natural Voice before I learned of their expensive licencing costs. The TextAloud people refunded the cost of this voice without argument.

My comments are restricted to TextAloud because it is the only TTS with which I have experience. What follows is not a recommendation for TextAloud.

I agree that the assessment of how suitable a voice is for an application is very subjective. Over the past weeks I have listened to every FREE voice that I can find. Some are definitely better than others. Currently I have 15 voices that I have enabled for use with TextAloud, and I have a further 19 to add to these if I need to.

I also agree that from my perspective, Natural Voice voices seem to be better than the free voices, BUT the difference is not (IMHO) enough to warrant the costs of the AT&T voices.

I can't comment on other TTS engines, but TextAloud includes dozens of user-controlled options which can be used to fine tune the results. This includes a pronunciation editor. I am still learning how to optimise these settings.

As far as using TTS with Opus is concerned, my experience to date is minimal. TextAloud lets me save files in WAV or MP3 formats. I just use the required wav file as I would any other sound file.

HTH