First, using VirtualBox you can run Windows and on an Ubuntu desktop. You can find more info to see if your favorite programs and devices are supported at https://www.virtualbox.org/.
Next, If you already have Windows on your computer, you should probably just consider installing Ubuntu on the same hard drive next to Windows. That way, if you ever want to use Windows, just reboot your PC.
The option to install Ubuntu "along side" or "along with" your current OS is suggested to the user during a regular installation from a LiveDVD or USB. Ubuntu installs the Grub2 bootloader that allows you to select operating systems when you boot your computer after installation.
There are dozens of youtube videos that show this process step by step.
Update I did some checking on dragon and found some info here (http://thenerdshow.com/platypus.html)[http://thenerdshow.com/platypus.html] I haven't used it but platypus is supposed to allow dragon give keyboard commands when running in a virtual environment and even Wine as well.
source: (http://en.wikipedia.org/wiki/Speech_recognition_software_for_Linux#cite_note-2)[http://en.wikipedia.org/wiki/Speech_recognition_software_for_Linux#cite_note-2]
Simon is also a linux alternative for speech recognition to check out.
Finally if you feel up to it you can code your own software using the dragonfly python module. Dragonfly appears to be a promising tool that supports D.N.S. and many other voice rec. technologies. (https://pypi.python.org/pypi/dragonfly/0.6.5)[https://pypi.python.org/pypi/dragonfly/0.6.5]