Sound input clipping workaround... in GA (no more robot sound)

Jun 15, 2006, 10:56 PM #1
unused_user_name Senior Member Join: Jun 2006 Posts: 501	Sound input clipping workaround... in GA (no more robot sound) Hello, I've posted this in anothe thread, but I figured I should move this to a new one so it gets more exposure... I'm trying to get Dragon Naturally Speaking working under a Windows XP VM, inside Parallels. I have a USB mic (Andrea AK5370). First I tried without success toget the mic working as a USB device under windows. After giving up on that, I moved to that apple sound system to discover the odd buzing (clipping) sound that others have described. I have found a reason and a workaround for this problem. The problem: Some programs (not all of them) have an odd buzzing noise (that sounds like the audio is being clipped in the analog world) in the sound input. Not all programs have this problem. My hypothsis: The Parallels virtual sound driver requires some kind of initilization that is in the spec for the Intel sound card it is emulating, but the real hardware does not actually require the initilization even though it is in the spec. Workaround: Install audacity (http://audacity.sourceforge.net). Open the program and "monitor input" by right clicking on the VU meter on the right. While Audacity is still open, fire up the MS sound recorder (or whatever audio program you want) and use the sound input. You will notice that the clipping noise is gone. Thoughts: Audacity will initilize sound hardware "to spec" as the devolopers of that program are very picky about their sound quality. Other programs will simply hope that MS's driver initilizes it (which my guess is that it does not). After this workaround, Dragon works correctly. (As another note: One thing Parallels does not like is suspending the machine with an external mic plugged in, and resuming on the internal mic. This breaks sound support for the VM until you restart PD with the external mic selected. If you use external mics, plug and unplug while PD is not running)

Jun 15, 2006, 10:56 PM

unused_user_name
Senior Member

Join: Jun 2006
Posts: 501

Sound input clipping workaround... in GA (no more robot sound)

Hello,

I've posted this in anothe thread, but I figured I should move this to a new one so it gets more exposure...

I'm trying to get Dragon Naturally Speaking working under a Windows XP VM, inside Parallels. I have a USB mic (Andrea AK5370).

First I tried without success toget the mic working as a USB device under windows. After giving up on that, I moved to that apple sound system to discover the odd buzing (clipping) sound that others have described.

I have found a reason and a workaround for this problem.

The problem: Some programs (not all of them) have an odd buzzing noise (that sounds like the audio is being clipped in the analog world) in the sound input. Not all programs have this problem.

My hypothsis: The Parallels virtual sound driver requires some kind of initilization that is in the spec for the Intel sound card it is emulating, but the real hardware does not actually require the initilization even though it is in the spec.

Workaround: Install audacity (http://audacity.sourceforge.net). Open the program and "monitor input" by right clicking on the VU meter on the right. While Audacity is still open, fire up the MS sound recorder (or whatever audio program you want) and use the sound input. You will notice that the clipping noise is gone.

Thoughts: Audacity will initilize sound hardware "to spec" as the devolopers of that program are very picky about their sound quality. Other programs will simply hope that MS's driver initilizes it (which my guess is that it does not).

After this workaround, Dragon works correctly.

(As another note: One thing Parallels does not like is suspending the machine with an external mic plugged in, and resuming on the internal mic. This breaks sound support for the VM until you restart PD with the external mic selected. If you use external mics, plug and unplug while PD is not running)

Jun 16, 2006, 09:20 AM #2
rohela Junior Member Join: Apr 2006 Posts: 19	Have you experimented with different size parallel installations to find the optimal RAM and hard disk allocation sizes fpr Dragon Naturally Speaking? How does the performance compare to Bootcamp? I assume you have to dictate i Dragon Pad and then cut and paste the results, correct? Your installation uses a OSX product to set the mic input. Did you use system preferences in OSX as well or just Win XP sound control panel?

Jun 16, 2006, 09:20 AM

rohela
Junior Member

Join: Apr 2006
Posts: 19

Have you experimented with different size parallel installations to find the optimal RAM and hard disk allocation sizes fpr Dragon Naturally Speaking?

How does the performance compare to Bootcamp? I assume you have to dictate i Dragon Pad and then cut and paste the results, correct?

Your installation uses a OSX product to set the mic input. Did you use system preferences in OSX as well or just Win XP sound control panel?

Jun 17, 2006, 12:18 AM #3
unused_user_name Senior Member Join: Jun 2006 Posts: 501	I am using a 10gb expanding partition, with 512Mb of RAM allocated to the VM. If I give it any less then 512mb of RAM the recognition accuracy gets really bad. Also If I give it more then 1GB of RAM (in my Macbook with 2GB of RAM) then it also loses accuracy, due to not enough RAM on the Mac side. I ahve not tried diffrent partition sizes, but I have found that as long as the partition does not change size much (i.e. creating large files) Dragon works as well as it does on my real hardware Windows box. (I am using the same Dragon user with the same USB mic to hold things as constant as I can) I changed the sound input in the Mac side. The audio driver inside Windows seems to mirror whatever the Mac gets as input, regardless of what is selected in it. I dictate to Dragon pad, and save the files in a SMB share so I can get at them directly on Mac. The only down side is that Windows likes to have ^M's at the end of each line, and that Dragon Pad does not add carrage returns at the end of lines when the wrap. (This is plain text files) I'm working on a little python script that will fix the second one. Fixing the first one is simply calling dos2unix on the file. Also, I do not get a CPU spike unless I tell Audacity to record. If I let it "monitor input" by clicking on the VU meter on the right, then minimize the window and ignore it then it does not eat my CPU for lunch. Sometimes if I suspend the machine while Dragon is running, then resume it without the mic plugged in (or even with the mic plugged in, but not selected as the input device), then Windows sound support is borked completely until I suspend/enable mic/resume again. This is annoying, but whatever, I can talk to my Mac now :)

Jun 17, 2006, 12:18 AM

unused_user_name
Senior Member

Join: Jun 2006
Posts: 501

I am using a 10gb expanding partition, with 512Mb of RAM allocated to the VM.

If I give it any less then 512mb of RAM the recognition accuracy gets really bad. Also If I give it more then 1GB of RAM (in my Macbook with 2GB of RAM) then it also loses accuracy, due to not enough RAM on the Mac side.

I ahve not tried diffrent partition sizes, but I have found that as long as the partition does not change size much (i.e. creating large files) Dragon works as well as it does on my real hardware Windows box. (I am using the same Dragon user with the same USB mic to hold things as constant as I can)

I changed the sound input in the Mac side. The audio driver inside Windows seems to mirror whatever the Mac gets as input, regardless of what is selected in it.

I dictate to Dragon pad, and save the files in a SMB share so I can get at them directly on Mac. The only down side is that Windows likes to have ^M's at the end of each line, and that Dragon Pad does not add carrage returns at the end of lines when the wrap. (This is plain text files)

I'm working on a little python script that will fix the second one. Fixing the first one is simply calling dos2unix on the file.

Also, I do not get a CPU spike unless I tell Audacity to record. If I let it "monitor input" by clicking on the VU meter on the right, then minimize the window and ignore it then it does not eat my CPU for lunch.

Sometimes if I suspend the machine while Dragon is running, then resume it without the mic plugged in (or even with the mic plugged in, but not selected as the input device), then Windows sound support is borked completely until I suspend/enable mic/resume again. This is annoying, but whatever, I can talk to my Mac now :)

Jul 5, 2006, 08:44 PM #4
ahmontgo Junior Member Join: Jul 2006 Posts: 6	Thanks for the hint, works fine now for me. I've been experimenting with Win2VNC + OSXvnc as a way of dictating directly into my Mac applications. Win2VNC allows you to use the VNC protocol to send keyboard and mouse commands to your Mac (running OSXvnc). I have a two-screen setup where Windows is running in one screen on the left and Mac OS X is running in the screen on the right. I set up Win2VNC to transfer control of the keyboard and mouse to the Mac when I run the cursor off the left side of the Windows screen. If Dragon is already running, you can then dictate into any application on the Mac side. Caveats: correction doesn't work properly in all applications, apparently since the keyboard controls for selecting and deleting text differ from application to application. For example, this dialog box in Safari works fine for correction, but OS X Mail doesn't work properly. Additionally, for whatever reason, control cannot be returned to the Windows part. Normally, I should be able to transfer control back (through VNC) by moving the cursor to the right side of the Mac screen, but this just bounces. I suspect that if we could turn off mouse synchronization in Parallels Tools, this would work properly. Synergy might also work, but it's much more difficult to configure properly, so I haven't tried it yet.

Jul 5, 2006, 08:44 PM

ahmontgo
Junior Member

Join: Jul 2006
Posts: 6

Thanks for the hint, works fine now for me.

I've been experimenting with Win2VNC + OSXvnc as a way of dictating directly into my Mac applications. Win2VNC allows you to use the VNC protocol to send keyboard and mouse commands to your Mac (running OSXvnc). I have a two-screen setup where Windows is running in one screen on the left and Mac OS X is running in the screen on the right. I set up Win2VNC to transfer control of the keyboard and mouse to the Mac when I run the cursor off the left side of the Windows screen. If Dragon is already running, you can then dictate into any application on the Mac side.

Caveats: correction doesn't work properly in all applications, apparently since the keyboard controls for selecting and deleting text differ from application to application. For example, this dialog box in Safari works fine for correction, but OS X Mail doesn't work properly. Additionally, for whatever reason, control cannot be returned to the Windows part. Normally, I should be able to transfer control back (through VNC) by moving the cursor to the right side of the Mac screen, but this just bounces. I suspect that if we could turn off mouse synchronization in Parallels Tools, this would work properly.

Synergy might also work, but it's much more difficult to configure properly, so I haven't tried it yet.

Aug 4, 2006, 10:03 AM #5
dkwehe Member Join: Aug 2006 Posts: 55	WinXP Dragon to Mac side Brilliant idea -- dictating into DNS with output to Mac. Can you give a bit more detail as to how you accomplished this? I've installed OSXvmc and RealVNC, but haven't had any luck making this connection. David.

Aug 6, 2006, 07:34 PM #6
ahmontgo Junior Member Join: Jul 2006 Posts: 6	Install OSXvnc on the OS X side, and Win2VNC on the Windows side. Win2VNC takes keyboard and mouse actions from the Windows side and just transmits them (and not video, unlike most VNC clients) to OSXvnc. RealVNC is a traditional VNC client; you could use for the same purpose, but you would get the additional lag of retransmitting the video feedback from the OSX side into Windows.

Jun 16, 2006, 10:02 PM #7
beg_ne Junior Member Join: May 2006 Posts: 5	You sir are a GENIUS!!! This fixed the crappy robot voice problem in my VOIP apps. Now I actually sound human. Hopefully the guys at parallels will implement a fix now that you've found something that fixes the bad sound. Running Audacity causes my Macbook to run at about 116% CPU and makes the fans kicks on so the situation isn't ideal. If only I knew about this fix a week ago, I could have had my friend sold on a Mac. Maybe she'll still listen to me though when she hears my clear voice online. :)

Jun 16, 2006, 10:02 PM

beg_ne
Junior Member

Join: May 2006
Posts: 5

You sir are a GENIUS!!! This fixed the crappy robot voice problem in my VOIP apps. Now I actually sound human. Hopefully the guys at parallels will implement a fix now that you've found something that fixes the bad sound.

Running Audacity causes my Macbook to run at about 116% CPU and makes the fans kicks on so the situation isn't ideal. If only I knew about this fix a week ago, I could have had my friend sold on a Mac. Maybe she'll still listen to me though when she hears my clear voice online. :)

Jun 23, 2006, 01:09 PM

MicroDev
Senior Member

Join: Apr 2006
Posts: 122

Quote:

Originally Posted by unused_user_name

Hello,
I'm trying to get Dragon Naturally Speaking working under a Windows XP VM, inside Parallels. I have a USB mic (Andrea AK5370)

Have you tried reducing the hardware acceleration and sound sample quality prior to doing this? In Win 2K this made a big difference for me. Win XP seemed unaffected...

Jun 24, 2006, 04:31 PM

unused_user_name
Senior Member

Join: Jun 2006
Posts: 501

Quote:

Originally Posted by MicroDev

Have you tried reducing the hardware acceleration and sound sample quality prior to doing this? In Win 2K this made a big difference for me. Win XP seemed unaffected...

Changing those settings have no effect under XP as far as I can tell...

Oct 15, 2006, 04:48 AM #10
stephtara Junior Member Join: Oct 2006 Posts: 6	When you say "inside parallels" is it parallels desktop or parallels workstation? I don't really understand the difference between the two.

Oct 15, 2006, 12:29 PM #11
joem Senior Member Join: Apr 2006 Posts: 1,275	Parallels Desktop is the name of the Mac version of the product. Parallels workstation is the name of the product for other hosts. Why there are two different names for what is essentially the same product escapes my feeble little brain. __________________ MacBook Pro 2.4G; core 2 duo; 4GB RAM, 500G HD 10.4.11 XP, 768 Meg RAM

Nov 8, 2006, 01:07 PM #12
brunchboy Junior Member Join: Nov 2006 Posts: 1	Direct support would be nicer I'd really like to be able to keep system-level audio coming out the Mac speakers, and within Cisco's IP Communicator under XP in Parallels, configure the phone audio to use the Logitech USB headset. For that to work, Parallels would need to be able to "take over" the headset as a Windows device. Is there any hope for that in the future?

« Previous Thread | Next Thread »

Search this thread	Forum jump
Advanced Search

Thread tools	Display modes
Printable Version Email this Page	Linear Mode Hybrid Mode Threaded Mode

Parallels Technology Network

Sound input clipping workaround... in GA (no more robot sound)