Speech Recognition: Software and Microphone

[ijames@mac.com]

Hello There Everyone,

I have a client who is looking to use his new installation of Parallels and Windows to run the latest and greatest speech recognition software. I have researched a bit but 10,000 heads are better than 1. Consequently, I have a few questions. Some questions may be outside the realm of the parallels team. Some questions should be within your sphere of concern. Any help that you can give will be most appreciated.

1. What is the best speech recognition software that works in the Mac parallels environment with Windows XP?

2. What is the best microphone designed for dictation that is functional with the current version of Parallels? (There is mention that isosynchronous mics do not function in parallels. Does this effect me?)

Customer Configuration
Mac Pro Quad Core 3.0GHz (Temporarily reduced to dual core to address Parallels kernel panics)
2GB's of RAM (we have more on hand)
Mac OS 10.4.7
Windows XP, Office XP


Thank you very much for your time,
James L. McMahan Jr.

 

[hacky]

I cannot address all of your questions due to my lack of expertise. I use Dragon 9 Preferred with the NON-USB microphone that came with it. There is a tolerable slight lag.

 

[billmccl]

This was supposed to be a quotattion from an earlier post:

1. My preference for Dragon over Ilisten is not based on the software, but on the microphone. I use voice recognition for hours a day as a busy internal medicine doctor, and I just can't use the boom microphone/headset required for Ilisten for extended periods of time. I find the workflow with Dragon a little superior to Ilisten, but accuracy seems quite good with both.

Now my question:

I have a question about your observation about the accuracy of the two programs. I've been hearing both views: from the iListen people and adherents I get the view that their voice software is as good as Dragon's, but their learning curve is steeper; it is capable of equal accuracy over time; from others I hear the opposite, including from some who are impartial. Now you sound impartial, and impartial to both software apps. Did you really get equal results (accuracy) from both? Or did you not get far enough with iListen to know? (I'm trying to decide which way to go.)

Thanks.

Sorry for the delay in getting back to you.

I am using DNS9.1 now with the Philip's Speech Mike Pro. All works great on Bootcamp and Windows XP. I have not used IListen in several months, since we concluded our evaluation process. Our sense at that time was that accuracy would be satisfactory with either if you used them. I don't know of any metrics that would measure and document this, though. I am waiting to upgrade to Vista in Bootcamp, but the Medical version of DNS9.1 is not yet available from Nuance. It is promised (as always!).

For us, the issue remains with the microphone and the workflow, not the software. If we were to try to go native OSX, we would then have to evaluate the use of Citrix, since our organization is rapidly moving to Citrix for all of our core business and clinical activities. Citrix support OSX just fine, and it works fine with DNS9.1 on Windows. I have not tested IListen on OSX Citrix applications, but that would be important.

Our views have not changed since October and my first post. If you are a fairly casual dictator, and committed to the OSX platform as I am, I think that IListen should work fine. For serious or industrial strength work, until a well thought out microphone for the Mac arrives, Dragon will remain the only reason I keep boot camp working in my otherwise all Mac environment. I really do not understand why some vendor doesn't address the Macintosh microphone issue, but there you have it.

I hope that was helpful.

 

[HenryB12]

I have been using Dragon NaturallySpeaking 9 Preferred with the non-USB microphone supplied with the software. Although the microphone is not USB-based, it works reasonably well for my needs. There is a slight delay during recognition, but it is generally tolerable.

I cannot answer all of the technical questions because my experience with speech-recognition hardware and software is limited, but a better-quality microphone or a USB audio device might help reduce latency and improve recognition performance, particularly when running Dragon in a virtual machine such as Parallels.