Monday, June 10, 2013

Current Systems in Speech Synthesis

Small Vocabulary (Can be used by many users)

We believe that this option is most suited for those who require services such as automated telephone answering. The most attractive feature in this system that you can have any kind of accent or different pattern to your speech and system will completely understand what you say!!! There are so many products and software out there that does not work on different variations of speech and this feature makes sure that is no longer a barrier for effective speech recognition. But the only downside to it is that there is limited usage and it can only answer to pre-set commands.


Large Vocabulary (Can be used by limited users)
If you’re running a business and need software that only a small number of people use, then look no further, as this is the perfect selection for you! Large vocabulary systems are known for their high level of accuracy of up to 85% (Or could even go higher depending on the user). Even though these systems have an impressive vocabulary which progresses well into the thousands, they are still unable to work with a small number of people without prior training. It is also important to mention that if it is being used by any other user than Primary users the accuracy rate is most likely to fall by leaps and bounds.
Discrete and Continuous speech. These are two speech recognition systems that were being used a decade ago. Discrete speech could be explained as a program which requires the speaker to speak words with regular intervals and much clarity. If the speaker were to speak in a normal manner with regular word flow and in an appropriate conversational manner then the best suited system, according to my research, would be the Continuous speech system. We believe that the latter option would be the best since the speaker does not have to spend a significant amount of time thinking and pronouncing their words so that the system can understand it. The good news is that the Continuous speech system is freely available in most modern systems so you’re good to go!

The healthcare industry is also benefiting quite heavily with this system. Using it in their front-end and back-end processes is proving to be quite helpful when it comes to documentation. All you have to do is speak into the speech recognition software and it instantly records everything and converts your speech into text. So whether you’re entering patient data or even doing some complicated and delicate medical research, this speech software is sure to make your life much easier and a plus point is that you can multi-task as well! So you’re actually getting twice the work done in half the time! Now that’s what we call advancements in technology!


Don’t forget to check in tomorrow where we will be discussing (In great detail) the impact speech recognition has on the Medical Industry.