It is required to write a program that will segment the speech signal by ponemone (it is not essential to determine what kind of letter it is), whether it is an audio recording or a speech signal with a micro. The result can be reflected on the spectrogram with labels or, for example, broken into small pieces of audio, each of which will contain a letter, or in some other way. Tell me where to start and where to go?
- Do you need this for a thesis? Well, that is, such work will undoubtedly be another small ugly bike, because large and not very development teams have already managed to work on this task for several years. And you alone on your knee want to plan this. Can it be easier to immediately use the ready-made speech recognition library? - Bulson
- @Bulson For graduation thesis) Already wrote a program for recognition, but I was scolded that I was engaged in segmentation and classification, but only the first is needed, they say we leave it for an emergency case your program, young man) Even an ugly bike will do, t. to. there the water will be more important about the segmentation as a whole, and the bike will be an excellent example. - TonnyPeperonny
- Python is now actively used in the academic environment, such a library is being downloaded, source codes are open, if you really need C #, you can peek and rewrite, well, articles can be found, for example, it’s funny that the list of references at the end of the article consists entirely of domestic authors :) For .Net there is a good library NAudio c using it you can read sound files and work with them. I wish you success. - Bulson
- @Bulson thank you very much - TonnyPeperonny
|