Answer the question
In order to leave comments, you need to log in
How to determine who is speaking at the moment?
Hello everyone, such a question. There is a conference where people talk among themselves, you need to somehow determine who is the active speaker at the moment. Getting the data is not the problem, the problem is in the calculation. Those. there are many variations here: one may constantly speak loudly, one may constantly speak softly, which means that the average calculation does not quite fit here. I receive data every n ms, I can change the period. Thank you all in advance.
Answer the question
In order to leave comments, you need to log in
1. First you need to write down all the data flow criteria that you can get for one participant (it can be not only audio data!).
2. Form a model for each of the participants based on these criteria (pre-learn).
3. Generate a hash according to the model (hashing).
4. Search for such a hash in the stream (stream->hash==hash->participant) and return the participant by hash (analyze).
This will work even if several are talking at the same time!
Didn't find what you were looking for?
Ask your questionAsk a Question
731 491 924 answers to any question