14 DAY TRIAL // Did you ever fantasized about vocally interacting with your PC? that you would be able to speak to it and it would respond? Well, I know I did and I am waiting for that moment for many years now. This will definitely not happen in the near future because voice recognition software is far from being useful in real life situations. But what about the speech synthesizing software? This should be the voice of a PC. In this review we will see how useful it can be. The application chosen for reviewing is KTTS and all the rendering will be done by festival, which is one of the most popular engines for speech synthesis.
The name KTTS comes from KDE Text-to-Speech. This suggests that this application will be used for reading text in a somehow natural voice. This is actually a very complicated task but we are not going to get into details.
What's it good for?
To be as objective as possible and to play a little, I used KTTS to speak to my friends. When I was saying something in other language than English, they weren't able to understand anything. This is understandable, because the language supported by festival is English. Support for other languages is very poor or not present at all. When I used KTTS to speak in English they could understand most of the words. Actually some words where pronounced very well. The problem is that some bad apples really mess things up.
KTTS is an excellent subsystem for converting text to speech because the text is sent to it via DCOP. This means that it will use a daemon, KTTSD, which will take care of queuing the speech job and sending the text to the speech synthesis engine. Because of this design, virtually any application can be easily designed to use text to speech.
Using the KDE Text-to-Speech Manager (KTTSMgr) you can easily enable the daemon and make it speak notifications provided by KNotify.
A cool example of how you can use KTTS in other applications is the Kopete KTTS Plugin. This one can be used to announce Kopete's messages, read them and several other actions can be also configured.
The best feature of KTTSMgr is that it can be used to speak text from the KDE clipboard. With this feature you can speak text from virtually any document or website. All you have to do is select the text, use Ctrl+C and then open the speech manager and click the Speak Clipboard button from the jobs tab. As you can see, near it there is a Speak File button. You can try reading any file you wish. It is cool that you can also add jobs in KTTSMgr and that it is possible to put a job on hold, to delete it or to change the order of the cue. A very common practice can be to read long text. In this case, the manager knows to group it in sentences and paragraphs so you can backup by one of this two criteria. This is very useful because some sentences will not be understandable and replaying just that one is a lot more practical than replaying all the text.
KTTS supports several filters that can be used to make this program work better when it is really used. Built-in filters include a talk chooser, a string replacer and a XML transformer. Because I brought it up about the talk chooser you should know that several talkers can be used, but at the moment the only use of alternate talkers is to use different voices for different applications. For example, you can make KTTS read Kopete messages with a female voice, read system notifications with an English male voice and read everything else with an American male voice.
The audio can be outputted using aRts or using the Gstreamer sound engines. The speed can also be set from the audio tab and, if you like, the audio files can be saved in a folder.
The Good
KTTS is an excellent backend for different text to speech engines. It integrates very well in KDE and its functionality is limited by the speech synthesis. The way jobs are handled and support for several talkers, each one supporting any synthesis engine you want, are some excellent things this program has to offer.
The Bad
First of all, I didn't manage to read PDF documents with KTTS. Secondly, I think a little more functionality is still needed. Currently I believe that blind people can't really take advantage of it.
The Truth
KTTS can be a real help for handicapped persons perhaps with the help of KMouth. There are plans to make it better and when KDE4 will be available, a better KTTS should also be ready. Unfortunately, my dream of having a PC that speaks to me will have to wait some more years because the speech synthesis engines can't really generate a somehow natural voice. The conclusion is that if you really need it you should use it, otherwise just cool down.
Check out the screenshots below:
