The fact that I could program something that actually does “something” in the real world was fascinating.
When the time to go to a university came about, I was also dabbling in some other specializations, but automatization won in the end.
In the second experimental part of this thesis, the developed version of the system is assessed for speaker data sets of different size.
Clean speech is used for the training phase while speech in the presence of babble noise is used for speaker testing.
During my first year as a Ph D student, I met Honza Černocký, who had supervised Petr Schwarz’s thesis, and I started co-operating with his group fully.
That resulted in a year-long internship together with Peter at the Oregon Graduate Institute (OGI), US, which was supervised by world-renowned Hynek Heřmanský.
After all, all you need is to understand the core principle and the rest is piece of cake.
Later, at high school of electric engineering, I learned about automatization, and I was riveted by it.
What we managed to do within that one year was almost unbelievable.
And then, in 2003 and still under the auspices of OGI, we first introduced ourselves to the worldwide speech-recognition community.