You need to resample the incoming audio in the time domain first to 16 kHz and then use our frequency domain class. You can get a 2:1 overlap ratio with the advance() method, and setting up the fft size to 512 can get you near 400 (fft can be power of two only). But really, this is too complex and reduces audio quality without any significant performance benefit over the Superpowered frequency domain example.
Articles in this section
- Are there any resources you can share about how to approach an auto mixer with Superpowered?
- Can I use Superpowered to detect DTMF tones?
- Can I use the offline analyzer for real-time pitch detection?
- Can your tool detect and export more fine-grained pitch data in Hz/msec using MIDI pitch bend messages ?
- Does the Superpowered API do pitch detection based on a monophonic vocal recording in mp3 format or compare and return Hz/msec for each detected pitch?
- I don't know how to return most of the analysis results from Offline Analyzer to my java code so that I can use and display them. Do you have any tips?
- I'm analyzing audio at the rate of 400 sample frames which has a 200-sample overlap at 16 KHz. I would be grateful if you could provide me a setup on how to do that?