Note
Access to this page requires authorization. You can try signing in or changing directories.
Access to this page requires authorization. You can try changing directories.
In a VMware/Omnissa environment, the speech recognition app can be hosted on a VMware/Omnissa server or virtual desktop. The client end point can be a thick client running a Microsoft Windows operating system, a thin client running a Linux/Microsoft Windows Embedded operating system or a zero client with no operating system. Regardless of your virtualization technology and architecture, you must be able to deliver audio from the client end point to the hosted app; for more information, see Audio channel deployment and configuration.
App deployment
Apps or desktops that are hosted in a virtualized environment are displayed as a bit map image via a receiver or remote desktop app.
This has the following implications:
Your speech recognition app isn't installed on the client end point, but on the VMware/Omnissa server/virtual desktop.
Your speech recognition app isn't installed where the microphone is plugged in.
If your speech recognition app sends recognized text to a target app (for example, a clinical documentation program or word processor), the speech recognition app must be installed on the same server or virtual desktop image as the target app to be able to access the target app's text controls.
System configurations
The following system configurations are supported:
Single hop configuration (described in this guide): Your speech recognition app is hosted on a VMware/Omnissa virtual desktop or server and streamed to the client end point (for example, a Microsoft Windows PC or thin client).
Double hop configuration: Your speech recognition app is hosted on a Citrix Virtual App server (second hop) and delivered to a Citrix or VMware/Omnissa virtual desktop or Citrix Virtual App server (first hop), which is then streamed to the client end point (for example, a Microsoft Windows PC or thin client). For more information, see Dragon Medical One - double hop configuration for Citrix environments.
Audio channel deployment and configuration
To deliver audio from the client end point to the hosted app, the following options are available:
Recommended: Install the Dragon VMware/Omnissa Client Audio Extension on the client end point (for example, a Microsoft Windows PC). For more information, see Dragon virtual extensions.
Use the VMware/Omnissa native audio channel and configure automatic device splitting on the client end point (for example, zero clients or Linux thin clients).
Configure USB redirection on client end point (for example, zero clients). We don't recommend using USB redirection because it depends on the state of the network and is therefore unreliable. In addition, some devices require a high network throughput even when they're idle. Audio devices work better with the native audio channel.
Note
In Windows Server 2019 and higher, apps must be given explicit access to hardware devices via Group Policy. To allow the speech recognition app to access the microphone device that's passed into the VDI session, create a Group Policy Object (GPO) for all users to set the registry key HKEY_CURRENT\USER\Software\Microsoft\Windows\CurrentVersion\CapabilityAccessManager\ConsentStore\microphone to Allow; this can then be applied to all servers.
Dragon virtual extensions
The Dragon VMware/Omnissa Client Audio Extension and Dragon PowerMic VMware/Omnissa Client Extension provide custom audio and microphone button channels for the following products:
VMware/Omnissa Horizon
Microsoft Windows and Windows Embedded operating systems
The following speech recognition aps:
Dragon Medical One Desktop Application
Dragon Medical Direct
Dragon Case and Care
Apps based on Dragon Medical SpeechKit (.NET and COM editions)
Apps based on SpeechMagic SDK
For more information on hardware, software and network requirements, see: Requirements.
Audio channels
High quality audio is required for accurate speech recognition. Regardless of the virtualization technology and architecture, you must be able to deliver audio from the client end point to the app hosted on the server.
Native audio channels can require between 150 kbps and 1 Mbps bandwidth between the client end point and the hosted app. The Dragon VMware/Omnissa Client Audio Extension reduces the bandwidth requirement to 19.2-27.8 kbps (depending on the sound format).
The corresponding improvements in app responsiveness and performance are critical to the user experience.
Microphone control channels
Microphone buttons, sliders and other controls must be routed separately to the speech recognition app.
To enable this for the Nuance PowerMic in a VMware/Omnissa Horizon environment, the Dragon PowerMic VMware/Omnissa Client Extension provides a custom channel for Nuance PowerMic button controls.
For third-party devices, configure USB redirection or automatic device splitting. We recommend configuring device splitting on Linux thin clients; configure USB redirection only if the Linux thin client doesn't support device splitting.
Package contents
Dragon VMware/Omnissa Client Audio Extension
A virtual audio channel for VMware/Omnissa Horizon systems.
Client component:
Dragon VMware Audio and Button Extensions\Client\Dragon VMware Client Audio Extension.exeDragon PowerMic VMware/Omnissa Client Extension
A virtual channel for Nuance PowerMic button controls in a VMware/Omnissa Horizon system.
Client component:
Dragon VMware Audio and Button Extensions\Client\Dragon PowerMic VMware Client Extension.exe
Medical device disclaimer
Microsoft products and services (1) are not designed, intended or made available as medical devices, and (2) are not designed or intended to be a substitute for professional medical advice, diagnosis, treatment, or judgment and should not be used to replace or as a substitute for professional medical advice, diagnosis, treatment, or judgment.
Some information relates to pre-released products which may be substantially modified before they are commercially released. Microsoft makes no warranties, express or implied, with respect to the information provided here.