A good experience . . plus tips for v4 and NaturallySpeaking

Yes, this seems to be a common problem. By connecting the microphone as a USB attachment, my sound is terribly garbled, although the volume is fine. Using the VM sound panel to connect (whether or not the microphone is selected on the Mac side), the volume is very low. I don't know what the problem is, but I am hoping that it is in parallels since it works perfectly in version 3 (which is what I'm using to dictate this).

Also, I have not seen the 50% speed up.

It would be a marvelous Christmas present to have both of these, but may be too much to ask.
 
Hi - I am an enthusiast for DNS+Parallels v4 so at the risk of boring everyone stupid I will tell you again how it works for me.

1. Make sure at all times that you USB mike is NOT the selected mike in the Mac prefs. You should have "internal microphone" selecetd. This is important - you will have trouble if the USB mike is connecetd to the VM and the Mac at the same time.

2. Open the VM and in the menu Devices>USB choose your USB mike.

3. Then, in the menu Devices>sound choose the USB mike, or, if it doesn't appear, click "configure", make sure the box "connected" is ticked, then choose your USB mike in the pop-up "input" menu. (Note: it doesn't seem to matter what you have selected for "output").

4. In the Windows XP control panels (don't ask me about Vista) click on "sounds and audio devices" and in the "audio" and "voice" tabs choose your USB mike for input. Do NOT choose "intel integrated audio". Again, it doesn't seem to matter about output, but probably best to match it to what you have chosen for output in the devices menu, eg probably default).

5. Go back and check your USB mike is not the chosen input in your mac sound prefs!

6. Double check anything else you feel the need to check!

7. Set up a new voice user for your new configuration.

Good luck!

Tom
 
Also using the Plantronics calisto bluetooth USB headset myself.
It doesn't work as default USB Audio at all - very very soft, and seems to record the sound a little bit too fast when it does (rarely) work.
Can't get it to work when using it as an audio input device connected through Mac either.
Very frustrating when Dragon worked before in PD3 - now cannot use it at all in PD4.
Overall the only thing PD4 has given me is better DX9 support for things like Microsoft Live maps 3d but this is a trivial gimmick. I NEED the audio to work for Dragon - hope they make this a priority!
 
Sorry, but that's the whole point - DON'T select "default" as the input microphone in devices>sound or devices>USB. Select your actual USB mike! And select your USB mike in the Windows sound control panel as well.

OK I'll shut up now.
 
Hi tblack,
I did not miss your point. Just to be clear, when the usb audio device is connected for me to the VM itself, and I have it selected as the actual device under Windows XP, it does not work, and records either extremely softly or plays back too quickly. It may work for you with your particular USB device, but it seems not to do so with the bluetooth calisto connected through the VM.
 
Just an update -
Build 3810 did NOT fix the problem. Still getting overly soft, and very very sped up playback on recorded audio when using Plantronics calisto bluetooth USB headset (the kit bluetooth headset that Nuance sells with Dragon Naturally Speaking preferred).
 
I tried 3810 with high hopes. Dragon works smoothly with my headset connecting using VPN to the Mac side. I can dictate running DNS on the VM and have the output appear on the Mac side. The problem is, as before, the response lags dramatically using Parallels 4, whereas it was snappy in Parallels 3. I've tried single or multiple cores, various memory allocations of my 6 GB, but the bottom line is that Parallels 4 is just overall significantly slower than v3. I'm running a 2.66 GHz MacPro tower w/4 cores.
 
After creating a new installment of XP with PD 4, I found that Dragon (thought I only have 8) worked great. But along the line someone I lost something, and was only getting crackling recorded audio that was faster at playback than it was when recorded.

I went into help mode, uninstalled all my audio hardware, booted back up in normal mode, uninstalled Tools and reinstalled Tools, and now it works perfectly for me again.
 
dkwehe said:
Version 4 is just too slow for me. There are long periods in which I dictate and nothing happens. Then all of a sudden it all comes out, but the delay is unacceptable. I'm going back to version 3.

Thanks again for all your help Tom.

David
Click to expand...

Does anyone know how to speed up the delay? It looks like a "sound card delay" more than not enough ram/cpu power for me.
 
Delay amelioration

Just switch back to version 3 and you will find the delay is minimal. I have lost the pointer to the latest version of parallels 3, but probably the moderators can point you in the right direction.

dkw
 
Can I/how to run v3 and v4 at the same time? I don't want to downgrade all my other virtual machines.

thanks
 
Delay can be significant and dangerous

I think the delay depends on the part of the country you are from. Southerners that speak v e r y s l o w l y and use "y'all" a lot probably don't notice a delay. Northerners that have to speak quickly because their lips will freeze at lower speeds find the delays intolerable. I've heard rumors that some Northerners have had their frostbitten lips amputated because of this delay.
 
A small progress update

I installed the program Goldwave (a sound recording program for windows) and started to use its "monitor input" feature.

With only the GoldWave program running, and speaking one word at a time (pausing for let's say 2 seconds per word) the analog meter move instantaneously.

It is also instantaneous when I'm doing volume setting and snr tests.

The delay starts when I start to train the program... the VU meter will move after 3 seconds I speak, then for another 3 seconds after the VU meter has gone back to zero, then dragon type in the word.

The delay still happens after I quit the program. But when I quit the GoldWave program and start it again, the VU meter back to instantaneous.

Does dragon naturally speaking supports multicore CPU?
Should I upgrade to more cores, or a higher speed CPU instead?

Right now I am using 2.13 ghz x2 core. Maybe I will overclock the CPU and see if it helps.
 
Last edited:
I think I misunderstood the delay problem, please ignore my reply in relation to the delay, I have no experience with speech to text software under a virtual machine. :)
 
You must log in or register to reply here.
Back
Top