How to measure the similarity between the two audios on the iOS application is to check any open source or simple implementation.
To calculate the distance between just 1d vector, audio can be represented by 1-D vector. But the audio length will be different, so some pre-processing is needed.
Go ahead to get some clues here, thanks
variable length Equality between two scenes can be counted efficiently with DTW:
This algorithm is easy to implement and there are many existing implementations related to the wiki page. .
Just speak, audio can be presented by 1-D vector,
It is advisable to split on audio Frames and it has 2-D vector features , Where for each frame you have an array of values for different frequency bands. If you want to deal with music, for every frame is a good idea for an FFT speech, it is better to calculate
Again, you can use several existing libraries for mail frequency features. One of those speech recognition toolkits
No comments:
Post a Comment