| ??? 05/19/07 22:47 Modified: 05/19/07 23:02 Read: times |
#139544 - That\'s the benefit of using low-bit-rate encoding Responding to: ???'s previous message |
Schemes that endeavor to reduce the transmitted bit rate associated with encoded voice tend to discriminate between vital and "nonessential" information contained within it. Hence, they "lose" much of the background noise and many other characteristics that make voice and its speaker recognizable, in favor of preserving the essentials that make it intelligible. It can be driven to to the limits where it's nearly impossible to recognize a speaker, yet understand what's being said. That's what probably ought to be the goal here, becuase it allows the data to be reduced to fewer, inherently repeatable and therefore recognizable bytes.
Careful choice of the optimal command words will do the rest. RE |
| Topic | Author | Date |
| speech recognizing via 8052??? | 01/01/70 00:00 | |
| Simplify! | 01/01/70 00:00 | |
| I doubt even that... | 01/01/70 00:00 | |
| Actually | 01/01/70 00:00 | |
| running that search machine... | 01/01/70 00:00 | |
| Me too! | 01/01/70 00:00 | |
| Yes. Simplify. | 01/01/70 00:00 | |
| Dedicated hardware? | 01/01/70 00:00 | |
| not necessarily! | 01/01/70 00:00 | |
| speach recognition isnt difficult | 01/01/70 00:00 | |
| I heard a story | 01/01/70 00:00 | |
| That\'s the benefit of using low-bit-rate encoding | 01/01/70 00:00 | |
| drop me an email | 01/01/70 00:00 | |
| Write a primer | 01/01/70 00:00 | |
| Sensory Inc has chips and modules... | 01/01/70 00:00 | |
| Jez smith | 01/01/70 00:00 | |
continue this thread further | 01/01/70 00:00 |