Bob,
The fact that you're doing well with the generic Android application is very good news because it's a generic sampler that feeds up to 20 seconds of your voice through a high-speed speech recognition server at Nuance. However, a tremendous amount of changes have occurred between Ver. 9 and Ver. 11. Our
NaturallySpeaking 10 Review covers the changes between Ver. 10 and Ver. 9 and our
NaturallySpeaking 11.5 Review covers the changes between Ver. 11 and Ver. 10 but the short version is?
1. 80% of the code was rewritten from the ground up and Ver. 11 contains more changes than versions 1 through 10. Combined.
2. The sampling rate has been doubled from 11 kHz to 22 kHz.
3. NaturallySpeaking 11 is multithreaded, meaning that he uses 2 cores simultaneously.
However, NaturallySpeaking has been relatively accurate since Ver. 4 but there are other mitigating factors, such as the microphone and sound card. Your previous Ver. 9 software may have been much better than you think. Unfortunately, the manufacturer, OEM microphone barely qualifies as a backup microphone and if you combined that microphone with your computer?s integrated soundcard, it's equivalent to putting bicycle tires on a race car. Speech recognition can only be as strong as its weakest link.
If you still have your Dragon Medical 9, you're eligible for upgrade pricing but note that Dragon now includes a 30 day customer satisfaction no restock fee guarantee so if it fails, you could at least get your money back and it only takes about 4 min. to test our theory
