As a programmer, I feel almost compelled to provide my input on the subject of voice synthesis. If anyone more experienced finds any misinformation, please correct me as I don't have any direct experience working with speech synthesis software. I am providing an engineers outside perspective on the topic.
Speech synthesis can take up fair amounts of memory, as the system needs to know which sound to play for each syllable of the word. The sounds could either be generated on the fly, or stored in wav files then loaded when needed, such as is the case with a scripted in-game sequence or a cut-scene. In the former case, the speed at which the sound is generated depends on the speed of the hardware, and is pretty impractical in and of itself for realtime applications such as games, whereas in the latter, if the voice synthesis is purely for a situation such as saying a characters name, then yeah, it could work. But put it into the context of an RPG, where the amount of system resources used is already high, both storage and runtime resources, the added sound clips could push the team well beyond their storage budget. This is especially true if you store a sound bank for more than one character. In the case of console titles, developers are on very restrictive resource budgets. This is why you typically see speech synthesis used primarily on PC applications (macs are PC's, get over it) and in sound studios such as LMMS or Fruity Loops Studio.
I like the robot idea though... I may have to use that >=D
Speech synthesis can take up fair amounts of memory, as the system needs to know which sound to play for each syllable of the word. The sounds could either be generated on the fly, or stored in wav files then loaded when needed, such as is the case with a scripted in-game sequence or a cut-scene. In the former case, the speed at which the sound is generated depends on the speed of the hardware, and is pretty impractical in and of itself for realtime applications such as games, whereas in the latter, if the voice synthesis is purely for a situation such as saying a characters name, then yeah, it could work. But put it into the context of an RPG, where the amount of system resources used is already high, both storage and runtime resources, the added sound clips could push the team well beyond their storage budget. This is especially true if you store a sound bank for more than one character. In the case of console titles, developers are on very restrictive resource budgets. This is why you typically see speech synthesis used primarily on PC applications (macs are PC's, get over it) and in sound studios such as LMMS or Fruity Loops Studio.
I like the robot idea though... I may have to use that >=D