Theme : Experiment
Content :
I am trying to reproduce the experiment of the previous article, but I found that I am not very familiar with the relevant fields, so I read the relevant books and did some basic experiments.
This are my progress
In the last two weeks, my progress are mostly on the above, and then the details are like this.
I read two books and learned about the application of matlab in speech synthesis.
According to the instructions in the book, the following two experiments were completed.
The first one is the spectrogram. I learned about the sampling and preprocessing of speech signals.
The second one is 0-9 word recognition, I learned mfcc and simple pattern matching ideas
I have learned about the development of speech recognition and found that speech recognition is divided into isolated word recognition and continuous word recognition. There are specific people and non-specific person recognition points.
I also Understand the development of identification methods, from the traditional GNN-HMM to the current DNN-HMM, I have received some Inspiration about my future job from the history of speech recognition.
Then I learned some more professional phonetics software, which can provide the convenience of my experiment.
Finally, in some papers, I saw the commonly used open source phonetic software HTK, so I also tried it and made a small experiment of isolated words.
I want to deepen my understanding and use of HTK first. I want to use htk to do continuous speech recognition. Here are some articles and experiments that I am interested in.
After my current job is over, this is some of my future plans.