I am new to AudioKit and nearly know nothing about DSP, I currently work a project which needs to record speech, which needs to reduce the noise.
I google this topic and find some ways to implement it and read some articles.
On the internet, I found there are two ways to do it, first is get noise profile, then use it to reduce noise; second, to build a noise gate.
I can not figure out what is better and what is easier to implement.
Basically, I think I need to do follow things to implement a very basic noise reduction.
1. record some silent audio, which will contain the noise.
2. use FFT to find out the frequency of the noise.
3. use a filter, high pass maybe, to remove the noise from actual speech.
Is this correct way to do with AudioKit?
If this is an OK or correct way to do, then I have another question with FFT.
I use AKFFTTap to get fftData, but what does the data mean? how should I use these data to figure out the frequency of noise ?
Thanks.