hi,
i've tried both sphinx and dragon naturally speaking to recognize same set of continuous speech and found that naturally speaking has a very low word error rate while sphinx has very high (around 80% errors!) word error rate. However im not sure whether im using sphinx in most optimal way in this test. I've posted several threads in this forum to fine tune configurations, but never had the luck to achieve what i want. Therefore i would really like to know the maximum the sphinx could perform using a sample project or an unbiased comparison of these two products so that i can decide which path to take.
thank you
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
hi,
i've tried both sphinx and dragon naturally speaking to recognize same set of continuous speech and found that naturally speaking has a very low word error rate while sphinx has very high (around 80% errors!) word error rate. However im not sure whether im using sphinx in most optimal way in this test. I've posted several threads in this forum to fine tune configurations, but never had the luck to achieve what i want. Therefore i would really like to know the maximum the sphinx could perform using a sample project or an unbiased comparison of these two products so that i can decide which path to take.
thank you