之前写过了如何将speex与wav格式的音频互相转换,如果没有看过的请看一下连接
http://www.cnblogs.com/dongweiq/p/4515186.html
虽然自己实现了相关的压缩算法,但是发现还是与gauss的压缩比例差了一些,一部分是参数设置的问题,另外一部分是没有使用ogg的问题。
本来想研究一下gauss的ogg算法,然后将他录制的音频转为wav格式,再继续进行后面的频谱绘制之类的。
在后续的研究gauss的解码过程,他是先解了ogg的格式,然后分段,然后去掉speex的头,然后把一段段的speex数据再解码成pcm的原数据,最后使用audiotrack一段段播放出来了。audiotrack播放的时候是阻塞的,所以播了多久就解了多久。
既然他得到了一段段pcm的原数据,那我就可以去将这一段段的原数据拼起来,最后得到解码完成的整个的pcm数据,最后加上wav的头不就可以直接转换成wav格式的音频了么???
前天的时候想到这里,立马就去改了。
SpeexDecoder是gauss的demo里主要的解码类,我们复制一份,改名为SpeexFileDecoder
去掉里面跟播放相关的audiotrack变量,因为我们要得到的是解码数据,跟播放无关。
修改后代码如下
1 package com.sixin.speex; 2 3 import java.io.File; 4 import java.io.FileOutputStream; 5 import java.io.IOException; 6 import java.io.RandomAccessFile; 7 import java.util.ArrayList; 8 import java.util.List; 9 10 import android.media.AudioFormat; 11 import android.media.AudioManager; 12 import android.media.AudioTrack; 13 import android.os.RecoverySystem.ProgressListener; 14 import android.util.Log; 15 16 /** 17 * 采用Jspeex方案,首先解包,从ogg里面接出来,然后使用speex decode将speex转为wav数据并进行播放 18 * 19 * @author Honghe 20 */ 21 public class SpeexFileDecoder { 22 23 protected Speex speexDecoder; 24 private String errmsg = null; 25 private List<ProgressListener> listenerList = new ArrayList<ProgressListener>(); 26 private File srcPath; 27 private File dstPath; 28 29 public SpeexFileDecoder(File srcPath, File dstPath) throws Exception { 30 this.srcPath = srcPath; 31 this.dstPath = dstPath; 32 } 33 34 private void initializeAndroidAudio(int sampleRate) throws Exception { 35 int minBufferSize = AudioTrack.getMinBufferSize(sampleRate, AudioFormat.CHANNEL_OUT_MONO, AudioFormat.ENCODING_PCM_16BIT); 36 37 if (minBufferSize < 0) { 38 throw new Exception("Failed to get minimum buffer size: " + Integer.toString(minBufferSize)); 39 } 40 } 41 42 public void addOnMetadataListener(ProgressListener l) { 43 listenerList.add(l); 44 } 45 46 public String getErrmsg() { 47 return errmsg; 48 } 49 50 public void decode() throws Exception { 51 errmsg = null; 52 byte[] header = new byte[2048]; 53 byte[] payload = new byte[65536]; 54 final int OGG_HEADERSIZE = 27; 55 final int OGG_SEGOFFSET = 26; 56 final String OGGID = "OggS"; 57 int segments = 0; 58 int curseg = 0; 59 int bodybytes = 0; 60 int decsize = 0; 61 int packetNo = 0; 62 // construct a new decoder 63 speexDecoder = new Speex(); 64 speexDecoder.init(); 65 // open the input stream 66 RandomAccessFile dis = new RandomAccessFile(srcPath, "r"); 67 FileOutputStream fos = new FileOutputStream(dstPath); 68 69 int origchksum; 70 int chksum; 71 try { 72 73 // read until we get to EOF 74 while (true) { 75 if (Thread.interrupted()) { 76 dis.close(); 77 return; 78 } 79 80 // read the OGG header 81 dis.readFully(header, 0, OGG_HEADERSIZE); 82 origchksum = readInt(header, 22); 83 readLong(header, 6); 84 header[22] = 0; 85 header[23] = 0; 86 header[24] = 0; 87 header[25] = 0; 88 chksum = OggCrc.checksum(0, header, 0, OGG_HEADERSIZE); 89 90 // make sure its a OGG header 91 if (!OGGID.equals(new String(header, 0, 4))) { 92 System.err.println("missing ogg id!"); 93 errmsg = "missing ogg id!"; 94 return; 95 } 96 97 /* how many segments are there? */ 98 segments = header[OGG_SEGOFFSET] & 0xFF; 99 dis.readFully(header, OGG_HEADERSIZE, segments); 100 chksum = OggCrc.checksum(chksum, header, OGG_HEADERSIZE, segments); 101 102 /* decode each segment, writing output to wav */ 103 for (curseg = 0; curseg < segments; curseg++) { 104 105 if (Thread.interrupted()) { 106 dis.close(); 107 return; 108 } 109 110 /* get the number of bytes in the segment */ 111 bodybytes = header[OGG_HEADERSIZE + curseg] & 0xFF; 112 if (bodybytes == 255) { 113 System.err.println("sorry, don't handle 255 sizes!"); 114 return; 115 } 116 dis.readFully(payload, 0, bodybytes); 117 chksum = OggCrc.checksum(chksum, payload, 0, bodybytes); 118 119 /* decode the segment */ 120 /* if first packet, read the Speex header */ 121 if (packetNo == 0) { 122 if (readSpeexHeader(payload, 0, bodybytes, true)) { 123 124 packetNo++; 125 } else { 126 packetNo = 0; 127 } 128 } else if (packetNo == 1) { // Ogg Comment packet 129 packetNo++; 130 } else { 131 132 /* get the amount of decoded data */ 133 short[] decoded = new short[160]; 134 if ((decsize = speexDecoder.decode(payload, decoded, 160)) > 0) { 135 //把边解边播改为写文件 136 fos.write(ShortAndByte.shortArray2ByteArray(decoded), 0, decsize * 2); 137 } 138 packetNo++; 139 } 140 } 141 if (chksum != origchksum) 142 throw new IOException("Ogg CheckSums do not match"); 143 } 144 } catch (Exception e) { 145 e.printStackTrace(); 146 } 147 fos.close(); 148 dis.close(); 149 } 150 151 /** 152 * Reads the header packet. 153 * 154 * <pre> 155 * 0 - 7: speex_string: "Speex " 156 * 8 - 27: speex_version: "speex-1.0" 157 * 28 - 31: speex_version_id: 1 158 * 32 - 35: header_size: 80 159 * 36 - 39: rate 160 * 40 - 43: mode: 0=narrowband, 1=wb, 2=uwb 161 * 44 - 47: mode_bitstream_version: 4 162 * 48 - 51: nb_channels 163 * 52 - 55: bitrate: -1 164 * 56 - 59: frame_size: 160 165 * 60 - 63: vbr 166 * 64 - 67: frames_per_packet 167 * 68 - 71: extra_headers: 0 168 * 72 - 75: reserved1 169 * 76 - 79: reserved2 170 * </pre> 171 * 172 * @param packet 173 * @param offset 174 * @param bytes 175 * @return 176 * @throws Exception 177 */ 178 private boolean readSpeexHeader(final byte[] packet, final int offset, final int bytes, boolean init) throws Exception { 179 if (bytes != 80) { 180 return false; 181 } 182 if (!"Speex ".equals(new String(packet, offset, 8))) { 183 return false; 184 } 185 // int mode = packet[40 + offset] & 0xFF; 186 int sampleRate = readInt(packet, offset + 36); 187 // int channels = readInt(packet, offset + 48); 188 // int nframes = readInt(packet, offset + 64); 189 // int frameSize = readInt(packet, offset + 56); 190 // RongXinLog.SystemOut("mode=" + mode + " sampleRate==" + sampleRate + " channels=" + channels 191 // + "nframes=" + nframes + "framesize=" + frameSize); 192 initializeAndroidAudio(sampleRate); 193 194 if (init) { 195 // return speexDecoder.init(mode, sampleRate, channels, enhanced); 196 return true; 197 } else { 198 return true; 199 } 200 } 201 202 protected static int readInt(final byte[] data, final int offset) { 203 /* 204 * no 0xff on the last one to keep the sign 205 */ 206 return (data[offset] & 0xff) | ((data[offset + 1] & 0xff) << 8) | ((data[offset + 2] & 0xff) << 16) | (data[offset + 3] << 24); 207 } 208 209 protected static long readLong(final byte[] data, final int offset) { 210 /* 211 * no 0xff on the last one to keep the sign 212 */ 213 return (data[offset] & 0xff) | ((data[offset + 1] & 0xff) << 8) | ((data[offset + 2] & 0xff) << 16) | ((data[offset + 3] & 0xff) << 24) | ((data[offset + 4] & 0xff) << 32) 214 | ((data[offset + 5] & 0xff) << 40) | ((data[offset + 6] & 0xff) << 48) | (data[offset + 7] << 56); 215 } 216 217 protected static int readShort(final byte[] data, final int offset) { 218 /* 219 * no 0xff on the last one to keep the sign 220 */ 221 return (data[offset] & 0xff) | (data[offset + 1] << 8); 222 } 223 224 }
注意上面因为speex解码出来的是160个short类型的数组,而java写文件要求写入的是byte数组,所以我们还是用到了short转byte数组的方法shortArray2ByteArray,我封装了一个类。也贴在下边
1 package com.sixin.speex; 2 3 public class ShortAndByte { 4 /** 5 * @功能 短整型与字节的转换 6 * @param 短整型 7 * @return 两位的字节数组 8 */ 9 public static byte[] shortToByte(short number) { 10 int temp = number; 11 byte[] b = new byte[2]; 12 for (int i = 0; i < b.length; i++) { 13 b[i] = new Integer(temp & 0xff).byteValue();// 将最低位保存在最低位 14 temp = temp >> 8; // 向右移8位 15 } 16 return b; 17 } 18 19 /** 20 * @功能 字节的转换与短整型 21 * @param 两位的字节数组 22 * @return 短整型 23 */ 24 public static short byteToShort(byte[] b) { 25 short s = 0; 26 short s0 = (short) (b[0] & 0xff);// 最低位 27 short s1 = (short) (b[1] & 0xff); 28 s1 <<= 8; 29 s = (short) (s0 | s1); 30 return s; 31 } 32 33 /** 34 * @说明 主要是为解析静态数据包,将一个字节数组转换为short数组 35 * @param b 36 */ 37 public static short[] byteArray2ShortArray(byte[] b) { 38 int len = b.length / 2; 39 int index = 0; 40 short[] re = new short[len]; 41 byte[] buf = new byte[2]; 42 for (int i = 0; i < b.length;) { 43 buf[0] = b[i]; 44 buf[1] = b[i + 1]; 45 short st = byteToShort(buf); 46 re[index] = st; 47 index++; 48 i += 2; 49 } 50 return re; 51 } 52 53 /** 54 * @说明 主要是为解析静态数据包,将一个short数组反转为字节数组 55 * @param b 56 */ 57 public static byte[] shortArray2ByteArray(short[] b) { 58 byte[] rebt = new byte[b.length * 2]; 59 int index = 0; 60 for (int i = 0; i < b.length; i++) { 61 short st = b[i]; 62 byte[] bt = shortToByte(st); 63 rebt[index] = bt[0]; 64 rebt[index + 1] = bt[1]; 65 index += 2; 66 } 67 return rebt; 68 } 69 }
读出来的原数据我们放入到了dstPath的文件,再来看看是怎么操作的呢?其中我还是修改了gauss的speexplayer方法。
我们复制speexPlayer方法,改名为SpeexFileDecoderHelper,按照如下方法修改
1 /** 2 * 3 */ 4 package com.sixin.speex; 5 6 import java.io.File; 7 8 import android.os.Handler; 9 10 /** 11 * @author honghe 12 * 13 */ 14 public class SpeexFileDecoderHelper { 15 private String srcName = null; 16 private String dstName = null; 17 private SpeexFileDecoder speexdec = null; 18 private OnSpeexFileCompletionListener speexListener = null; 19 private static final int speexdecode_completion = 1001; 20 private static final int speexdecode_error = 1002; 21 22 public Handler handler = new Handler() { 23 public void handleMessage(android.os.Message msg) { 24 int what = msg.what; 25 switch (what) { 26 case speexdecode_completion: 27 if (speexListener != null) { 28 speexListener.onCompletion(speexdec); 29 } else { 30 System.out.println("司信---------null===speexListener"); 31 } 32 break; 33 case speexdecode_error: 34 if (speexListener != null) { 35 File file = new File(SpeexFileDecoderHelper.this.srcName); 36 if (null != file && file.exists()) { 37 file.delete(); 38 } 39 speexListener.onError(null); 40 } 41 break; 42 default: 43 break; 44 } 45 }; 46 }; 47 48 public SpeexFileDecoderHelper(String fileName,String dstName, OnSpeexFileCompletionListener splistener) { 49 this.speexListener = splistener; 50 this.srcName = fileName; 51 this.dstName = dstName; 52 try { 53 speexdec = new SpeexFileDecoder(new File(this.srcName),new File(this.dstName)); 54 } catch (Exception e) { 55 e.printStackTrace(); 56 File file = new File(SpeexFileDecoderHelper.this.srcName); 57 if (null != file && file.exists()) { 58 file.delete(); 59 } 60 } 61 } 62 63 public void startDecode() { 64 RecordDecodeThread rpt = new RecordDecodeThread(); 65 Thread th = new Thread(rpt); 66 th.start(); 67 } 68 69 public boolean isDecoding = false; 70 71 class RecordDecodeThread extends Thread { 72 73 public void run() { 74 try { 75 if (speexdec != null) { 76 isDecoding = true; 77 speexdec.decode(); 78 if (null != speexdec.getErrmsg()) { 79 throw new Exception(speexdec.getErrmsg()); 80 } 81 } 82 System.out.println("RecordPlayThread 文件转换完成"); 83 if (isDecoding) { 84 handler.sendEmptyMessage(speexdecode_completion); 85 } 86 isDecoding = false; 87 } catch (Exception t) { 88 t.printStackTrace(); 89 System.out.println("RecordPlayThread 文件转换出错"); 90 handler.sendEmptyMessage(speexdecode_error); 91 isDecoding = false; 92 } 93 } 94 } 95 96 /** 97 * 结束播放 98 */ 99 public void stopDecode() { 100 isDecoding = false; 101 } 102 103 public String getSpxFileName() { 104 return this.srcName; 105 }; 106 }
这个方法是开启了一个现成去解码,然后解码完成后会发送handler,调用回调方法,通知解码失败还是成功。OnSpeexFileCompletionListener这个很简单,我需要贴吗?还是贴上吧,省的被骂娘。
1 package com.sixin.speex; 2 3 /** 4 * Speex音频解码完成监听 5 * @author honghe 6 * 7 */ 8 public interface OnSpeexFileCompletionListener { 9 void onCompletion(SpeexFileDecoder speexdecoder); 10 void onError(Exception ex); 11 }
到此代码都贴出来了。什么?!还不会用?哦,我还没写怎么加wav头呢,那再写个方法吧
1 /** 2 * 语音转换 3 * 4 * @param name 5 * @param srcFileName spx文件名 6 * @param dstFileName 转换后得到文件的文件名 7 */ 8 public static void decodeSpx(Context context, String srcFileName, final String dstFileName) { 9 final String temppath = AudioFileFunc.getFilePathByName("temp.raw"); 10 try { 11 // 如果是speex录音 12 if (srcFileName != null && srcFileName.endsWith(".spx")) { 13 if (mSpeexFileDecoderHelper != null && mSpeexFileDecoderHelper.isDecoding) { 14 stopMusic(context); 15 } else { 16 muteAudioFocus(context, true); 17 mSpeexFileDecoderHelper = new SpeexFileDecoderHelper(srcFileName, temppath, new OnSpeexFileCompletionListener() { 18 19 @Override 20 public void onError(Exception ex) { 21 System.out.println("转换错误"); 22 } 23 24 @Override 25 public void onCompletion(SpeexFileDecoder speexdecoder) { 26 System.out.println("转换完成"); 27 WaveJoin.copyWaveFile(temppath, dstFileName); 28 } 29 }); 30 mSpeexFileDecoderHelper.startDecode(); 31 } 32 } else { 33 System.out.println("音频文件格式不正确"); 34 } 35 36 } catch (Exception e) { 37 e.printStackTrace(); 38 } 39 }
copyWaveFile这个方法在哪里?去看开头的那个链接吧,上面有。不过在加wav头的时候要注意,加的头要和你录制音频的时候设置的参数一致,比如samplerate,声道数,framesize这些,我就设置错了一次,gauss录制音频的时候使用的是单声道,我加入的wav头的channel设置成了2,结果放出来的声音老搞笑了,就跟快放一样。有兴趣你可以试试。
代码已更新
代码链接如下:
https://github.com/dongweiq/study/tree/master/Record
我的github地址:https://github.com/dongweiq/study
欢迎关注,欢迎star o(∩_∩)o 。有什么问题请邮箱联系 dongweiqmail@gmail.com qq714094450