ios多音频混合

ios多音频混合

from:  http://blog.csdn.net/zengconggen/article/details/7842076

ios上音频文件合成有多种方法，一是使用AVAssetExportSession结合AVMutableComposition，二是使用AVAssetReader和AVAssetWriter. 三是转换成pcm数据后处理pcm数据合成。

方法一和方法二在处理速度上相对来说要慢。一般3分半左右音频需要30s左右，特别是合成不同采样率的音频，速度比较慢。下文提供本人自己写的一个类，采用第三种方法合成音频，采用该算法合成的音频，比网络上一般代码提供的音频在合成效果上好。并且处理速度快，ipod touch4上处理4分半音频仅需6s左右。

[cpp] view plain copy

1. //  
2. //  ExtAudioFileMixer.h  
3. //  AudioManager  
4. //  
5. //  Created by zeng songgen on 12-8-7.  
6. //  Copyright (c) 2012年 Rumtel. All rights reserved.  
7. //  
8.   
9. #import <Foundation/Foundation.h>  
10.   
11. @interface ExtAudioFileMixer : NSObject  
12.   
13. + (OSStatus)mixAudio:(NSString *)audioPath1  
14.             andAudio:(NSString *)audioPath2  
15.               toFile:(NSString *)outputPath  
16.   preferedSampleRate:(float)sampleRate;  
17.   
18. @end  

[cpp] view plain copy

1. //  
2. //  ExtAudioFileMixer.m  
3. //  AudioManager  
4. //  
5. //  Created by zeng songgen on 12-8-7.  
6. //  Copyright (c) 2012年 Rumtel. All rights reserved.  
7. //  
8.   
9. #import "ExtAudioFileMixer.h"  
10. #import <AudioToolbox/AudioToolbox.h>  
11.   
12. @implementation ExtAudioFileMixer  
13.   
14.   
15.   
16. + (OSStatus)mixAudio:(NSString *)audioPath1  
17.             andAudio:(NSString *)audioPath2  
18.               toFile:(NSString *)outputPath  
19.   preferedSampleRate:(float)sampleRate  
20. {  
21.     OSStatus                            err = noErr;  
22.     AudioStreamBasicDescription         inputFileFormat1;  
23.     AudioStreamBasicDescription         inputFileFormat2;  
24.     AudioStreamBasicDescription         converterFormat;  
25.     UInt32                              thePropertySize = sizeof(inputFileFormat1);  
26.     ExtAudioFileRef                     inputAudioFileRef1 = NULL;  
27.     ExtAudioFileRef                     inputAudioFileRef2 = NULL;  
28.     ExtAudioFileRef                     outputAudioFileRef = NULL;  
29.     AudioStreamBasicDescription         outputFileFormat;  
30.       
31.     NSURL *inURL1 = [NSURL fileURLWithPath:audioPath1];  
32.     NSURL *inURL2 = [NSURL fileURLWithPath:audioPath2];  
33.     NSURL *outURL = [NSURL fileURLWithPath:outputPath];  
34.       
35.     // Open input audio file  
36.       
37.     err = ExtAudioFileOpenURL((CFURLRef)inURL1, &inputAudioFileRef1);  
38.     if (err)  
39.     {  
40.         goto reterr;  
41.     }  
42.     assert(inputAudioFileRef1);  
43.       
44.     err = ExtAudioFileOpenURL((CFURLRef)inURL2, &inputAudioFileRef2);  
45.     if (err)  
46.     {  
47.         goto reterr;  
48.     }  
49.     assert(inputAudioFileRef2);  
50.       
51.     // Get input audio format  
52.       
53.     bzero(&inputFileFormat1, sizeof(inputFileFormat1));  
54.     err = ExtAudioFileGetProperty(inputAudioFileRef1, kExtAudioFileProperty_FileDataFormat,  
55.                                   &thePropertySize, &inputFileFormat1);  
56.     if (err)  
57.     {  
58.         goto reterr;  
59.     }  
60.       
61.     // only mono or stereo audio files are supported  
62.       
63.     if (inputFileFormat1.mChannelsPerFrame > 2)  
64.     {  
65.         err = kExtAudioFileError_InvalidDataFormat;  
66.         goto reterr;  
67.     }  
68.       
69.     bzero(&inputFileFormat2, sizeof(inputFileFormat2));  
70.     err = ExtAudioFileGetProperty(inputAudioFileRef2, kExtAudioFileProperty_FileDataFormat,  
71.                                   &thePropertySize, &inputFileFormat2);  
72.     if (err)  
73.     {  
74.         goto reterr;  
75.     }  
76.       
77.     // only mono or stereo audio files are supported  
78.       
79.     if (inputFileFormat2.mChannelsPerFrame > 2)  
80.     {  
81.         err = kExtAudioFileError_InvalidDataFormat;  
82.         goto reterr;  
83.     }  
84.       
85.     int numChannels = MAX(inputFileFormat1.mChannelsPerFrame, inputFileFormat2.mChannelsPerFrame);  
86.       
87.     // Enable an audio converter on the input audio data by setting  
88.     // the kExtAudioFileProperty_ClientDataFormat property. Each  
89.     // read from the input file returns data in linear pcm format.  
90.       
91.     AudioFileTypeID audioFileTypeID = kAudioFileCAFType;  
92.       
93.     Float64 mSampleRate = sampleRate? sampleRate : MAX(inputFileFormat1.mSampleRate, inputFileFormat2.mSampleRate);  
94.       
95.     [self _setDefaultAudioFormatFlags:&converterFormat sampleRate:mSampleRate numChannels:inputFileFormat1.mChannelsPerFrame];  
96.       
97.     err = ExtAudioFileSetProperty(inputAudioFileRef1, kExtAudioFileProperty_ClientDataFormat,  
98.                                   sizeof(converterFormat), &converterFormat);  
99.     if (err)  
100.     {  
101.         goto reterr;  
102.     }  
103.     [self _setDefaultAudioFormatFlags:&converterFormat sampleRate:mSampleRate numChannels:inputFileFormat2.mChannelsPerFrame];  
104.     err = ExtAudioFileSetProperty(inputAudioFileRef2, kExtAudioFileProperty_ClientDataFormat,  
105.                                   sizeof(converterFormat), &converterFormat);  
106.     if (err)  
107.     {  
108.         goto reterr;  
109.     }  
110.     // Handle the case of reading from a mono input file and writing to a stereo  
111.     // output file by setting up a channel map. The mono output is duplicated  
112.     // in the left and right channel.  
113.       
114.     if (inputFileFormat1.mChannelsPerFrame == 1 && numChannels == 2) {  
115.         SInt32 channelMap[2] = { 0, 0 };  
116.           
117.         // Get the underlying AudioConverterRef  
118.           
119.         AudioConverterRef convRef = NULL;  
120.         UInt32 size = sizeof(AudioConverterRef);  
121.           
122.         err = ExtAudioFileGetProperty(inputAudioFileRef1, kExtAudioFileProperty_AudioConverter, &size, &convRef);  
123.           
124.         if (err)  
125.         {  
126.             goto reterr;  
127.         }  
128.           
129.         assert(convRef);  
130.           
131.         err = AudioConverterSetProperty(convRef, kAudioConverterChannelMap, sizeof(channelMap), channelMap);  
132.           
133.         if (err)  
134.         {  
135.             goto reterr;  
136.         }  
137.     }  
138.     if (inputFileFormat2.mChannelsPerFrame == 1 && numChannels == 2) {  
139.         SInt32 channelMap[2] = { 0, 0 };  
140.           
141.         // Get the underlying AudioConverterRef  
142.           
143.         AudioConverterRef convRef = NULL;  
144.         UInt32 size = sizeof(AudioConverterRef);  
145.           
146.         err = ExtAudioFileGetProperty(inputAudioFileRef2, kExtAudioFileProperty_AudioConverter, &size, &convRef);  
147.           
148.         if (err)  
149.         {  
150.             goto reterr;  
151.         }  
152.           
153.         assert(convRef);  
154.           
155.         err = AudioConverterSetProperty(convRef, kAudioConverterChannelMap, sizeof(channelMap), channelMap);  
156.           
157.         if (err)  
158.         {  
159.             goto reterr;  
160.         }  
161.     }  
162.     // Output file is typically a caff file, but the user could emit some other  
163.     // common file types. If a file exists already, it is deleted before writing  
164.     // the new audio file.  
165.       
166.     [self _setDefaultAudioFormatFlags:&outputFileFormat sampleRate:mSampleRate numChannels:numChannels];  
167.       
168.     UInt32 flags = kAudioFileFlags_EraseFile;  
169.       
170.     err = ExtAudioFileCreateWithURL((CFURLRef)outURL, audioFileTypeID, &outputFileFormat,  
171.                                     NULL, flags, &outputAudioFileRef);  
172.     if (err)  
173.     {  
174.         // -48 means the file exists already  
175.         goto reterr;  
176.     }  
177.     assert(outputAudioFileRef);  
178.       
179.     // Enable converter when writing to the output file by setting the client  
180.     // data format to the pcm converter we created earlier.  
181.       
182.     err = ExtAudioFileSetProperty(outputAudioFileRef, kExtAudioFileProperty_ClientDataFormat,  
183.                                   sizeof(outputFileFormat), &outputFileFormat);  
184.     if (err)  
185.     {  
186.         goto reterr;  
187.     }  
188.       
189.     // Buffer to read from source file and write to dest file  
190.       
191.     UInt16 bufferSize = 8192;  
192.       
193.     AudioSampleType * buffer1 = malloc(bufferSize);  
194.     AudioSampleType * buffer2 = malloc(bufferSize);  
195.     AudioSampleType * outBuffer = malloc(bufferSize);  
196.       
197.     AudioBufferList conversionBuffer1;  
198.     conversionBuffer1.mNumberBuffers = 1;  
199.     conversionBuffer1.mBuffers[0].mNumberChannels = inputFileFormat1.mChannelsPerFrame;  
200.     conversionBuffer1.mBuffers[0].mDataByteSize = bufferSize;  
201.     conversionBuffer1.mBuffers[0].mData = buffer1;  
202.   
203.     AudioBufferList conversionBuffer2;  
204.     conversionBuffer2.mNumberBuffers = 1;  
205.     conversionBuffer2.mBuffers[0].mNumberChannels = inputFileFormat2.mChannelsPerFrame;  
206.     conversionBuffer2.mBuffers[0].mDataByteSize = bufferSize;  
207.     conversionBuffer2.mBuffers[0].mData = buffer2;  
208.       
209.     //  
210.     AudioBufferList outBufferList;  
211.     outBufferList.mNumberBuffers = 1;  
212.     outBufferList.mBuffers[0].mNumberChannels = outputFileFormat.mChannelsPerFrame;  
213.     outBufferList.mBuffers[0].mDataByteSize = bufferSize;  
214.     outBufferList.mBuffers[0].mData = outBuffer;  
215.       
216.     UInt32 numFramesToReadPerTime = INT_MAX;  
217.     UInt8 bitOffset = 8 * sizeof(AudioSampleType);  
218.     UInt64 bitMax = (UInt64) (pow(2, bitOffset));  
219.     UInt64 bitMid = bitMax/2;  
220.       
221.       
222.     while (TRUE) {  
223.         conversionBuffer1.mBuffers[0].mDataByteSize = bufferSize;  
224.         conversionBuffer2.mBuffers[0].mDataByteSize = bufferSize;  
225.         outBufferList.mBuffers[0].mDataByteSize = bufferSize;  
226.           
227.         UInt32 frameCount1 = numFramesToReadPerTime;  
228.         UInt32 frameCount2 = numFramesToReadPerTime;  
229.           
230.         if (inputFileFormat1.mBytesPerFrame)  
231.         {  
232.             frameCount1 = bufferSize/inputFileFormat1.mBytesPerFrame;  
233.         }  
234.         if (inputFileFormat2.mBytesPerFrame)  
235.         {  
236.             frameCount2 = bufferSize/inputFileFormat2.mBytesPerFrame;  
237.         }  
238.         // Read a chunk of input  
239.           
240.         err = ExtAudioFileRead(inputAudioFileRef1, &frameCount1, &conversionBuffer1);  
241.           
242.         if (err) {  
243.             goto reterr;  
244.         }  
245.           
246.         err = ExtAudioFileRead(inputAudioFileRef2, &frameCount2, &conversionBuffer2);  
247.           
248.         if (err) {  
249.             goto reterr;  
250.         }  
251.         // If no frames were returned, conversion is finished  
252.           
253.         if (frameCount1 == 0 && frameCount2 == 0)  
254.             break;  
255.           
256.         UInt32 frameCount = MAX(frameCount1, frameCount2);  
257.         UInt32 minFrames = MIN(frameCount1, frameCount2);  
258.           
259.         outBufferList.mBuffers[0].mDataByteSize = frameCount * outputFileFormat.mBytesPerFrame;  
260.           
261.         UInt32 length = frameCount * 2;  
262.         for (int j =0; j < length; j++)  
263.         {  
264.             if (j/2 < minFrames)  
265.             {  
266.                 SInt32 sValue =0;  
267.                   
268.                 SInt16 value1 = (SInt16)*(buffer1+j);   //-32768 ~ 32767  
269.                 SInt16 value2 = (SInt16)*(buffer2+j);   //-32768 ~ 32767  
270.                   
271.                 SInt8 sign1 = (value1 == 0)? 0 : abs(value1)/value1;  
272.                 SInt8 sign2 = (value2== 0)? 0 : abs(value2)/value2;  
273.                   
274.                 if (sign1 == sign2)  
275.                 {  
276.                     UInt32 tmp = ((value1 * value2) >> (bitOffset -1));  
277.                       
278.                     sValue = value1 + value2 - sign1 * tmp;  
279.                       
280.                     if (abs(sValue) >= bitMid)  
281.                     {  
282.                         sValue = sign1 * (bitMid -  1);  
283.                     }  
284.                 }  
285.                 else  
286.                 {  
287.                     SInt32 tmpValue1 = value1 + bitMid;  
288.                     SInt32 tmpValue2 = value2 + bitMid;  
289.                       
290.                     UInt32 tmp = ((tmpValue1 * tmpValue2) >> (bitOffset -1));  
291.                       
292.                     if (tmpValue1 < bitMid && tmpValue2 < bitMid)  
293.                     {  
294.                         sValue = tmp;  
295.                     }  
296.                     else  
297.                     {  
298.                         sValue = 2 * (tmpValue1  + tmpValue2 ) - tmp - bitMax;  
299.                     }  
300.                     sValue -= bitMid;  
301.                 }  
302.                   
303.                 if (abs(sValue) >= bitMid)  
304.                 {  
305.                     SInt8 sign = abs(sValue)/sValue;  
306.                       
307.                     sValue = sign * (bitMid -  1);  
308.                 }  
309.                   
310.                 *(outBuffer +j) = sValue;  
311.             }  
312.             else{  
313.                 if (frameCount == frameCount1)  
314.                 {  
315.                     //将buffer1中的剩余数据添加到outbuffer  
316.                     *(outBuffer +j) = *(buffer1 + j);  
317.                 }  
318.                 else  
319.                 {  
320.                     //将buffer1中的剩余数据添加到outbuffer  
321.                     *(outBuffer +j) = *(buffer2 + j);  
322.                 }  
323.             }  
324.         }  
325.           
326.         // Write pcm data to output file  
327.         NSLog(@"frame count (%ld, %ld, %ld)", frameCount, frameCount1, frameCount2);  
328.         err = ExtAudioFileWrite(outputAudioFileRef, frameCount, &outBufferList);  
329.           
330.         if (err) {  
331.             goto reterr;  
332.         }  
333.     }  
334.       
335. reterr:  
336.     if (buffer1)  
337.         free(buffer1);  
338.       
339.     if (buffer2)  
340.         free(buffer2);  
341.       
342.     if (outBuffer)  
343.         free(outBuffer);  
344.       
345.     if (inputAudioFileRef1)  
346.         ExtAudioFileDispose(inputAudioFileRef1);  
347.       
348.     if (inputAudioFileRef2)  
349.         ExtAudioFileDispose(inputAudioFileRef2);  
350.       
351.     if (outputAudioFileRef)  
352.         ExtAudioFileDispose(outputAudioFileRef);  
353.       
354.     return err;  
355. }  
356.   
357. // Set flags for default audio format on iPhone OS  
358.   
359. + (void) _setDefaultAudioFormatFlags:(AudioStreamBasicDescription*)audioFormatPtr  
360.                           sampleRate:(Float64)sampleRate  
361.                          numChannels:(NSUInteger)numChannels  
362. {  
363.     bzero(audioFormatPtr, sizeof(AudioStreamBasicDescription));  
364.       
365.     audioFormatPtr->mFormatID = kAudioFormatLinearPCM;  
366.     audioFormatPtr->mSampleRate = sampleRate;  
367.     audioFormatPtr->mChannelsPerFrame = numChannels;  
368.     audioFormatPtr->mBytesPerPacket = 2 * numChannels;  
369.     audioFormatPtr->mFramesPerPacket = 1;  
370.     audioFormatPtr->mBytesPerFrame = 2 * numChannels;  
371.     audioFormatPtr->mBitsPerChannel = 16;  
372.     audioFormatPtr->mFormatFlags = kAudioFormatFlagsNativeEndian |  
373.     kAudioFormatFlagIsPacked | kAudioFormatFlagIsSignedInteger;  
374. }  
375.   
376. @end