Hi everyone, I'm more or less throwing this one out here for some end user feedback as I need it at this point. As many of you know STFC 1.0 and Montezuma's Revenge Alpha 4.4 are using prerecorded wav files for speech playback. The problem with this method is generally that It takes me lots of extra time to sample, record, trim, and edit every voice file for the game. In STFC I think there are a total of 400 wav files I had to do which took me something like three weeks to do. With the end result being massive over head. Needless to say using prerecorded speech has added allot of extra time to every production, added CPU over head, and at the end of the day I have to put in lots of special flags to get it to act something like a synth doing key interrupts etc so that the speech wouldn't hold up the application, etc... On the positive side though the voice, speech output, is generally better than most synths. I have the option to correct the way the synth says something which helps. For example Cardassian which might sound like car-dashen can be respelled during recorded to say car-dass-e-an. However, I am not sure the improved speech is all that worth it. What I would like to do is build a speech class into my engine.dll which should support all Sapi 5 synths, and through a good Sapi tts engine you could gain self voicing that way, and not have to wait an extra month or so as I edit voice samples to use for the game. The advantages here is less overhead, more rapid production time for games, smaller downloads, custom speech settings, and allot of other stuff you don't get with prerecorded speech. Since I was working on my screen reading project i realized how much faster it is when you can just do something like synth.Speak(message, SpeechVoiceSpeakFlags.SVSFDefault); and on to the next line of code without having to actually record the message, load and then play back a wave file. Which eats CPU power, not to mention lots more of my time to do the wav self-voicing option. The only thing I need to do with Sapi is construct a string of text, store it in the message string, and pass it to the synth.Speak method, and away I go. Another reason to go Sapi is there are allot of high quality speech engines out there that can be generically ssupported. The scansoft voices which ship with Jaws 8 are simply awesome. Since I have installed them I have no problems with using them with LW and Jim Kitchen's games, and they sound good for synths. The catch 22 is I don't know how many you have purchased the higher quality synths, and how many of you are using the older Microsoft ones Mike, Mary, Microsoft, etc which sound, well, terrible. So what do you as end users think? Should I go ahead and make Sapi a core part of future releases?
_______________________________________________ Gamers mailing list .. Gamers@audyssey.org To unsubscribe send E-mail to [EMAIL PROTECTED] You can visit http://audyssey.org/mailman/listinfo/gamers_audyssey.org to make any subscription changes via the web.