Models in the database not used in speaker recognition under Windows 8

[GoogleCodeExporter]

What steps will reproduce the problem?
In the Windows Command Prompt, enter the following commands:
1. python vid -s speaker1 -g training.wav
2. python vid -i testing.wav -f json

What is the expected output? What do you see instead?

In testing.json, "speaker1" should show up as one of the speaker candidates. 
However, the recognition was done as if there was no model in the database. The 
same data and procedure generate correct results under Mac OS X.

What version of the product are you using? On what operating system?

Both v0.3 and the latest from the trunk. OS is Win 8.1.

Please provide any additional information below.

Please let me know if there is way to make it work. Thanks!

Original issue reported on code.google.com by dong...@umich.edu on 22 May 2014 at 1:35