私はまだアンドロイドとFFTの初心者です。Frequency AnalyzerのようなAndroidアプリを作りたいです。垂直側に周波数と振幅があります。http://therandomlab.blogspot.com/2013/05/fft-audio-frequency-analysis-with.htmlからこのコードを試します
プロセッシングを使用しています。
import processing.core.*;
import processing.data.*;
import processing.event.*;
import processing.opengl.*;
import android.media.AudioRecord;
import android.media.AudioFormat;
import android.media.MediaRecorder;
import android.widget.Toast;
import java.util.HashMap;
import java.util.ArrayList;
import java.util.logging.Logger;
import java.io.File;
import java.io.BufferedReader;
import java.io.PrintWriter;
import java.io.InputStream;
import java.io.OutputStream;
import java.io.IOException;
public class android_fft_minim extends PApplet {
int RECORDER_SAMPLERATE = 44100;
int MAX_FREQ = RECORDER_SAMPLERATE / 2;
final int RECORDER_CHANNELS = AudioFormat.CHANNEL_IN_MONO;
final int RECORDER_AUDIO_ENCODING = AudioFormat.ENCODING_PCM_16BIT;
final int PEAK_THRESH = 20;
short[] buffer = null;
int bufferReadResult = 0;
AudioRecord audioRecord = null;
boolean aRecStarted = false;
int bufferSize = 2048;
int minBufferSize = 0;
float volume = 0;
FFT fft = null;
float[] fftRealArray = null;
int mainFreq = 0;
float drawScaleH = (float) 1.5; // TODO: calculate the drawing scales
float drawScaleW = (float) 1.0; // TODO: calculate the drawing scales
int drawStepW = 2; // display only every Nth freq value
float maxFreqToDraw = 2500; // max frequency to represent graphically
int drawBaseLine = 0;
public void setup() {
//Toast.makeText(android_fft_minim.this, "lalag", Toast.LENGTH_LONG).show();
drawBaseLine = displayHeight - 150;
minBufferSize = AudioRecord.getMinBufferSize(RECORDER_SAMPLERATE,
RECORDER_CHANNELS, RECORDER_AUDIO_ENCODING);
// if we are working with the android emulator, getMinBufferSize() does
// not work
// and the only samplig rate we can use is 8000Hz
if (minBufferSize == AudioRecord.ERROR_BAD_VALUE) {
RECORDER_SAMPLERATE = 8000; // forced by the android emulator
MAX_FREQ = RECORDER_SAMPLERATE / 2;
bufferSize = 2 << (int) (log(RECORDER_SAMPLERATE) / log(2) - 1);// buffer
// size
// must
// be
// power
// of
// 2!!!
// the buffer size determines the analysis frequency at:
// RECORDER_SAMPLERATE/bufferSize
// this might make trouble if there is not enough computation power
// to record and analyze
// a frequency. In the other hand, if the buffer size is too small
// AudioRecord will not initialize
} else
bufferSize = minBufferSize;
buffer = new short[bufferSize];
// use the mic with Auto Gain Control turned off!
audioRecord = new AudioRecord(
MediaRecorder.AudioSource.VOICE_RECOGNITION,
RECORDER_SAMPLERATE, RECORDER_CHANNELS,
RECORDER_AUDIO_ENCODING, bufferSize);
// audioRecord = new AudioRecord( MediaRecorder.AudioSource.MIC,
// RECORDER_SAMPLERATE,
// RECORDER_CHANNELS,RECORDER_AUDIO_ENCODING, bufferSize);
if ((audioRecord != null)
&& (audioRecord.getState() == AudioRecord.STATE_INITIALIZED)) {
try {
// this throws an exception with some combinations
// of RECORDER_SAMPLERATE and bufferSize
try {
audioRecord.startRecording();
} catch (Exception e) {
Toast.makeText(android_fft_minim.this, "error recording", Toast.LENGTH_LONG).show();
}
aRecStarted = true;
} catch (Exception e) {
aRecStarted = false;
}
if (aRecStarted) {
bufferReadResult = audioRecord.read(buffer, 0, bufferSize);
// verify that is power of two
if (bufferReadResult % 2 != 0)
bufferReadResult = 2 << (int) (log(bufferReadResult) / log(2));
fft = new FFT(bufferReadResult, RECORDER_SAMPLERATE);
fftRealArray = new float[bufferReadResult];
drawScaleW = drawScaleW * (float) displayWidth
/ (float) fft.freqToIndex(maxFreqToDraw);
}
}
fill(0);
noStroke();
}
public void draw() {
background(128);
fill(0);
noStroke();
if (aRecStarted) {
bufferReadResult = audioRecord.read(buffer, 0, bufferSize);
// After we read the data from the AudioRecord object, we loop
// through
// and translate it from short values to double values. We can't do
// this
// directly by casting, as the values expected should be between
// -1.0 and 1.0
// rather than the full range. Dividing the short by 32768.0 will do
// that,
// as that value is the maximum value of short.
volume = 0;
for (int i = 0; i < bufferReadResult; i++) {
fftRealArray[i] = (float) buffer[i] / Short.MAX_VALUE;// 32768.0;
volume += Math.abs(fftRealArray[i]);
}
volume = (float) Math.log10(volume / bufferReadResult);
// apply windowing
for (int i = 0; i < bufferReadResult / 2; ++i) {
// Calculate & apply window symmetrically around center point
// Hanning (raised cosine) window
float winval = (float) (0.5f + 0.5f * Math.cos(Math.PI
* (float) i / (float) (bufferReadResult / 2)));
if (i > bufferReadResult / 2)
winval = 0;
fftRealArray[bufferReadResult / 2 + i] *= winval;
fftRealArray[bufferReadResult / 2 - i] *= winval;
}
// zero out first point (not touched by odd-length window)
fftRealArray[0] = 0;
fft.forward(fftRealArray);
//
fill(255);
stroke(100);
pushMatrix();
rotate(radians(90));
translate(drawBaseLine - 3, 0);
textAlign(LEFT, CENTER);
for (float freq = RECORDER_SAMPLERATE / 2 - 1; freq > 0.0f; freq -= 150.0f) {
int y = -(int) (fft.freqToIndex(freq) * drawScaleW); // which
// bin
// holds
// this
// frequency?
line(-displayHeight, y, 0, y); // add tick mark
text(Math.round(freq) + " Hz", 10, y); // add text label
}
popMatrix();
noStroke();
float lastVal = 0;
float val = 0;
float maxVal = 0; // index of the bin with highest value
int maxValIndex = 0; // index of the bin with highest value
for (int i = 0; i < fft.specSize(); i++) {
val += fft.getBand(i);
if (i % drawStepW == 0) {
val /= drawStepW; // average volume value
int prev_i = i - drawStepW;
stroke(255);
// draw the line for frequency band i, scaling it up a bit
// so we can see it
line(prev_i * drawScaleW, drawBaseLine,
prev_i * drawScaleW, drawBaseLine - lastVal
* drawScaleH);
if (val - lastVal > PEAK_THRESH) {
stroke(255, 0, 0);
fill(255, 128, 128);
ellipse(i * drawScaleW,
drawBaseLine - val * drawScaleH, 20, 20);
stroke(255);
fill(255);
if (val > maxVal) {
maxVal = val;
maxValIndex = i;
}
}
line(prev_i * drawScaleW, drawBaseLine - lastVal
* drawScaleH, i * drawScaleW, drawBaseLine - val
* drawScaleH);
lastVal = val;
val = 0;
}
}
if (maxValIndex - drawStepW > 0) {
fill(255, 0, 0);
ellipse(maxValIndex * drawScaleW, drawBaseLine - maxVal
* drawScaleH, 20, 20);
fill(0, 0, 255);
text(" " + fft.indexToFreq(maxValIndex - drawStepW / 2) + "Hz",
25 + maxValIndex * drawScaleW, drawBaseLine - maxVal
* drawScaleH);
}
fill(255);
pushMatrix();
translate(displayWidth / 2, drawBaseLine);
text("buffer readed: " + bufferReadResult, 20, 80);
text("fft spec size: " + fft.specSize(), 20, 100);
text("volume: " + volume, 20, 120);
popMatrix();
} else {
fill(255, 0, 0);
text("AUDIO RECORD NOT INITIALIZED!!!", 100, height / 2);
}
fill(255);
pushMatrix();
translate(0, drawBaseLine);
text("sample rate: " + RECORDER_SAMPLERATE + " Hz", 20, 80);
text("displaying freq: 0 Hz to " + maxFreqToDraw + " Hz", 20, 100);
text("buffer size: " + bufferSize, 20, 120);
popMatrix();
}
public void stop() {
audioRecord.stop();
audioRecord.release();
}
public int sketchWidth() {
return displayWidth;
}
public int sketchHeight() {
return displayHeight;
}}
そして、fftにはこのクラスを使用しています。 http://code.compartmental.net/minim/javadoc/ddf/minim/analysis/package-summary.html
私には2つの問題があります：
- エミュレータで実行すると正常に動作しますが、デバイスで実行するとエラーが発生しました。誰か私に提案してもらえますか？
- このコードは垂直線の振幅を返しますか?