Missing WAV files

[cmlooi]

I failed to execute modeltraining.py because the WAV files are missing. Would you be able to share them?

[Atul-Anand-Jha]

Brother, thanks for showing interest in my repo. If you like it, give it a star.

There are few samples given in the repo. You can edit the number of times loop runs for the speker voice samples. Otherwise, you can put the terminator while number of speaker samples end.

Sorry, I can not share all of the voice samples; due to privacy.

Try to play with the code.

[cmlooi]

I failed to execute modeltraining.py because the WAV files are missing. Would you be able to share them?

After many trial and error, I managed to figure out how to use the modeltraining.py to generate the gmm out of the WAV files downloaded from VoxForge. Also, have to spend time understand the file naming that look for the "-" and the relationship of the txt files. Anyway, speaker identification is running. Lot more to learn. Well done Atul.

[Atul-Anand-Jha]

I am glad that you liked it.Please give it a star. So, it ranks higher,and others could also get benefit from this. Do experimentations from this project. There is a lot more scope. In case you come up with something great out of this project.Let me know.

[cmlooi]

Hi Atul Anand,

First of all, thanks for sharing your work. I read the related thesis too. It is an opportunity for me to learn new stuffs and understand how you make use of gmm for speaker recognition.

For your codes, there are plenty of room for improvement. Some enhancement and simplification is required for other to run your code without having to spend time trying to figure out the many details. Most people will just stop trying when they see error popping up.

I cannot access to my office computer now. I will send you a minor improvement coming Monday. Just two lines of code to exit lopping and prevent an error message after finished processing the gmm files. My key point to you is keep making improvement; and do not stop working on it. Getting more people interested in the subject is more important. An example is Pydub (https://github.com/jiaaro/pydub ) and Speech Recognition ( https://github.com/Uberi/speech_recognition ) in Github.

Are you aware of a company GoVivace (originated from India) who is commercialising their speech solution. Commercialisation follows a different path.

I am an EE engineer specialises in designing voice remote control, far-field voice and application involving speech recognition. I have some understand about the speech feature extraction and HMM kind of stuff, but did not really go into that detail. I graduated more than thirty years ago. I have to pick up new stuffs from people like you.

Last but not least, I will give you a star coming Monday.

Best regards, CM Looi

On Fri, 2 Nov 2018 at 23:51, Atul Anand notifications@github.com wrote:

I am that you liked it.Please give it a star. So, it ranks higher,and others could also get advantage from this. Do experimentations from this project. There is a lot more scope. In case you come up with something great out of this project.Let me know.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/Atul-Anand-Jha/Speaker-Identification-Python/issues/2#issuecomment-435425126, or mute the thread https://github.com/notifications/unsubscribe-auth/AEjObdZh5ts8jn9oyzv_DH3GiZ6SRi1Vks5urGoJgaJpZM4YGC22 .

[cmlooi]

Improvement on modeltraining.py so that it can exit the loop without throwing up an error.

Extracting features for each speaker (5 files per speakers)

features = np.asarray(()) for path in file_paths:
path = path.strip()

# Detection to exit loop when no more file for processing
if path == "":
    print "Exit upon completion!"
    break

print path

# read the audio
sr,audio = read(source + path)

[Atul-Anand-Jha]

Thanks a lot @cmlooi for showing such immense interest into my project. And, also for suggesting improvement. I will embed it.

Actually, this was just a college minor project. SO, There are ample scopes of improvement into this. Currently, I am in my final year of graduation. So, I dont have much time to invest on this project. But, sooner I will start adding improvements into this. ANd, Once, i get to learn REST APIs. I will convert this project's architecture. SO , everyone could use it by simply installing/embedding.

Regards, -Atul.