The voice synthesizer in Okular always "speaks" with a robotic sound and an English accent that you can't understand at all (well, maybe the English do, but for most people it's a torture). Is so awful that using it for more than 3 minutes at a time is a torture, and the problem is that there's no tool to configure TTS in Okular's preferences, if you want to use TTS in Okular you are tied to that horrendous sound. So, could be added one simple interface to choose the voice, language, tone and speed, especially the voice and language? There are some open source projects that use Google's and IBM's even better voices: https://github.com/ken107/read-aloud https://github.com/rNeomy/reader-view Would be posible to make Okular do the same and use good voices even if not open source? Greetings.
Okular uses the QtSpeech library (https://doc.qt.io/qt-5/qtspeech-index.html) and its backends for the various platforms . On *nix it uses libspeechd/speech-dispatcher, so the proper way to configure this is to configure speech-dispatcher.
I understand, thanks for the hint. But I wasn't actually asking, I just wanted to suggest a feature request: if a graphical tool in Okular's preferences could be implemented so that the user does not have to mess around with configuration files, that from this tool can configure these things easily. I don't know if I have made myself clear. For example, the sound can be configured using Alsaconf and things like that, but obviously we prefer to do it using the tools provided by Plasma, which deal with Alsa and such from a graphical interface without bothering the user. So I ask if it would be possible that Okular, since it is a graphical application, include a graphical configuration tool for speech-dispatcher. Changing the mentioned paramenters is not an "exotic" setting either, I mean surely all users are going to want to change the horrible voices that come by default for ones that do not drill your ears, and let's not fool ourselves, neither my mother nor my father-in-law nor probably yours are going to know or want to waste time when they could have a graphical tool that with 4 clicks will solve the problem. Please, take a look at this video and you will see what I mean (it should start at minute 4:05). https://youtu.be/_JCKc1CCOR0?t=245 It doesn't look particularly complex, does it? Well, from my lack of knowledge of the programming world it doesn't seem so to me, maybe I'm wrong. Anyway, if Okular could have those controls to choose voice and 2 or 3 little things more and that is what is in charge of "negotiating" or whatever with speed-dispacher, it would be easily usable as a reader aloud by everyone not only by "geeks" (with all the due respect, don't get me wrong).
By the way, it also seems that configuring speech-dispatcher by hand doesn't work either. I don't know how to make SP use Google or IBM Watson voices, but I've tried editing the /etc/speech-dispatcher/speechd.conf file to use Espeak, and Okular still recites in a tinny voice with a gringo accent. xD. But I guess perhaps this should be discussed in another report, right?