| ??? 01/23/09 10:28 Read: times |
#161705 - Synthesis vs construction? Responding to: ???'s previous message |
Per Westermark said:
Synthetic speech works best at home, for example if someone is listening to the morning paper and can stay focused on what is said. But in a PA system, you normally don't have any history - no expected context and you don't have time to get used to the synthetic voice before the message has ended. Very true. The automated anouncements at our local railway station are very good - but they are clearly not text-to-speech synthesis; rather, they are "constructed" from recorded words and phrases. Clearly, the recorded words and phrases were originally spoken by a well-trained person. I don't know if there are any recognised terms to distinguish between these different approaches? |
| Topic | Author | Date |
| RS 232 based text2speech Modules | 01/01/70 00:00 | |
| Unheppy | 01/01/70 00:00 | |
| Unhappy reasons | 01/01/70 00:00 | |
| Amount of material or update frequency? | 01/01/70 00:00 | |
| Thats reverberation Vs pronunciation | 01/01/70 00:00 | |
| Not the issue here? | 01/01/70 00:00 | |
| synthetic speech in itself harder to understand | 01/01/70 00:00 | |
| Synthesis vs construction? | 01/01/70 00:00 | |
Prerecorded words/sentences works quite well | 01/01/70 00:00 |