package audiofingerprinter;

import java.io.File;

import java.util.List;

/**

* Created by bjackson on 11/17/2015.

*/

public interface AudioFingerprinter {

/**

* Returns the database of songs that this fingerprinter uses to recognize.

* @return

*/

SongDatabase getSongDB();

/**

* Given an array of bytes representing a song, this method will return a list of song names with matching fingerprints.

* The algorithm is as follows:

* - Convert the audio to the frequency domain

* - For each slice of time, determine key points, and create a hash fingerprint

* - Find matching datapoints from the song database.

* - Calculate the number of matching points for each candidate song

* - return a list of song names in order of most matches to least matches.

* @param audioData array of bytes representing a song

* @return A list of song names with matching fingerprints, sorted in order from most likely match to least likely match.

*/

List<String> recognize(byte[] audioData);

/**

* Overloaded method given a file object to recognize.

* Hint: get the raw audio data from the file and call the other overloaded recognize method.

* @param fileIn

* @return

*/

List<String> recognize(File fileIn);

/**

* Given a 2D array of frequency information over time, returns the keypoints.

* @param results, an array of frequency data. The first index corresponds with a slice of time, the second with the frequency.

* The data is represented as complex numbers with interleaved real and imaginary parts. For example, to get the

* magnitude of a specific frequency:

* double re = results[time][2*freq];

* double im = results[time][2*freq+1];

* double mag = Math.log(Math.sqrt(re * re + im * im) + 1);

* @return a 2D array where the first index represents the time slice, and the second index contains the highest frequencies

* for the following ranges with that time slice: 30 Hz - 40 Hz, 40 Hz - 80 Hz and 80 Hz - 120 Hz for the low tones (covering bass guitar,

* for example), and 120 Hz - 180 Hz and 180 Hz - 300 Hz for the middle and higher tones (covering vocals and most other instruments).

*/

long[][] determineKeyPoints(double[][] results);

/**

* Returns a hash combining information of several keypoints.

* @param points array of key points for a particular slice of time. Must be at least length 4.

* @return

*/

long hash(long[] points);

}