Question

我正在使用musicg库来比较两个音频声音。其中一个保存之前，另一个通过记录检测。 musicg库有whistle演示;我编辑它是为了我想要的工作。问题是当我想将检测到的声音[这是一个缓冲区]转换为wav并获得它的fingurprint时，它返回nullpointerexption。

这是在MainAcivity.java中的的

的

public static Wave w1;

InputStream sw1 = getBaseContext（）。getResources（）。openRawResource（R.raw.top_of_the_world_rec）;

w1 = new Wave（sw1）;

的

这是在DetectorThread.java中的的

的
InputStream obj = new ByteArrayInputStream(buffer); Wave woo = new Wave(obj); if (woo.toString() == null) Log.i("", "recoded sound is null"); if (MainActivity.w1.toString() == null) Log.i("", "the detected sound is null"); try { similarity = woo.getFingerprintSimilarity(MainActivity.w1);/// this couse null pointer exption Log.i("", "1-" + similarity); // not appear if (similarity == null) { // not enter this if at all ~! Log.i("", "lolol" + similarity); TheSemilarityISnull = true; } } catch (NullPointerException n) { Log.i("", "first error"); TheSemilarityISnull = true; } // EnD of addition Log.i("", "2-" + similarity); //appear 2-null // byte[] io = MainActivity.w1.getFingerprint(); // good no problem // Log.i("", "00-"+io); // apear bk09823 like this try { Log.i("", "0-"); byte[] po = woo.getFingerprint(); // ~> this couse null pointer exption Log.i("", "0-" + po); if (po == null) Log.i("", "fingurprint for detected sound is null ");//not appear byte[] io = MainActivity.w1.getFingerprint(); Log.i("", "00-" + io); if (io == null) Log.i("", "fingurprint for recorded sound is null");//not appear double scoure = new FingerprintSimilarityComputer(po, io).getFingerprintsSimilarity().getSimilarity(); } catch (Exception e) { Log.i("", "second error "); }

的

所以;

1 - 如何将Byte []转换为wav？

2 - 我可以在线程中调用class或Mothed吗？我不知道每当我打电话给其中一个时它都没有发现任何东西！

3-我根本不编辑记录线程＆gt;这会是问题吗？请帮我。

Answer 1

读缓冲区不包含任何波形标头。这些只是音频字节。我已经开始实现一个可能有效的类。但它还不是最终的

public class WavApi extends DetectionApi {
  Wave waveToMatch;

  public WavApi(Wave wave) {
    super(wave.getWaveHeader());
    this.waveToMatch = wave;
  }

  public boolean matches(byte[] audioBytes) {
    Wave wave = new Wave(waveToMatch.getWaveHeader(), audioBytes);

    FingerprintSimilarityComputer fpc = new FingerprintSimilarityComputer(waveToMatch.getFingerprint(), wave.getFingerprint());
    FingerprintSimilarity similarity = fpc.getFingerprintsSimilarity();

    //--- TODO: find optimal condition to decide whether {audioBytes} matches {waveToMatch} or not
    float sim = similarity.getScore() * 100f;
    int pos   = similarity.getMostSimilarFramePosition();

    if (pos >= 0 && sim > 0)
        return true;
    //----

    return false;
  }
}

比较android中的两个声音

1 个答案: