WAV文件结构和PCM数据转存WAV文件

Alsa相关命令

arecord 录制

可以使用arecord录制音频输入，保存为wav文件

arecord -f <格式> -r <采样率> -c <声道数> -d <时长> output.wav例如：
arecord -f S16_LE -r 16000 -c 1 -d 5 output.wav

常用参数说明

• -f：指定音频格式（常用 S16_LE 表示 16 位 PCM，WAV 常用格式）
• -r：采样率（如 44100 Hz，CD 音质）
• -c：声道数（1 单声道，2 立体声）
• -d：录制时长（单位：秒，省略则手动停止）

aplay 播放

再使用aplay来播放

aplay output.wav

实验环境：

下面这个我是直接在Linux虚拟机上做的实验

WAV文件结构

可以参考这一篇：【音视频 | wav】wav音频文件格式详解——包含RIFF规范、完整的各个块解析、PCM转wav代码

我有一个wav文件，除了LIST段的Chuck，我都做了简单的对照

WAV文件头结构体：

除音频数据外的信息都属于WAV文件头的部分

先定义一下相关结构体，用于解析头部数据

#include <stdio.h>
#include <stdlib.h>
#include <string>
#include <vector>
#include <algorithm> 
#include <stddef.h>
#include <fcntl.h>
#include <unistd.h>
#include <sys/stat.h>
#include <sys/types.h>
#include <string.h>#include <alsa/asoundlib.h>
#include <stdint.h>/*
编译命令: gcc wav_tool.c -o AlsaTool -lasound -lstdc++
使用方法:录音: ./AlsaTool播放: ./AlsaTool 文件名.wav
*/// 固定宽度类型定义
typedef uint16_t  WORD;
typedef uint32_t  DWORD;
typedef int8_t    CHAR;// 结构体定义（按字节对齐）
typedef struct 
{WORD    wFormatTag;         // 编码格式（1=PCM）WORD    nChannels;          // 声道数DWORD   nSamplesPerSec;     // 采样率DWORD   nAvgBytesPerSec;    // 平均字节率WORD    nBlockAlign;        // 块对齐
} __attribute__((packed)) WAVEFORMAT;typedef struct 
{WAVEFORMAT wf;WORD       wBitsPerSample;  // 采样位数
} __attribute__((packed)) PCMWAVEFORMAT;typedef struct 
{CHAR    SubchunkID[4];      // 块标识（"fact"）DWORD   SubchunkSize;       // 块大小DWORD   dwSampleLength;     // 采样帧数
} __attribute__((packed)) FACT_CHUNK;typedef struct 
{CHAR    ChunkID[4];         // 标识（"RIFF"）DWORD   ChunkSize;          // 总大小（不含自身ID和大小）CHAR    Format[4];          // 格式（"WAVE"）
} __attribute__((packed)) RIFF_CHUNK;typedef struct 
{CHAR    SubchunkID[4];      // 标识（"data"）DWORD   SubchunkSize;       // 数据大小
} __attribute__((packed)) DATA_CHUNK;// 完整WAV头结构体（RIFF + fmt块 + data块）
typedef struct 
{RIFF_CHUNK    riff;           // RIFF块（12字节）char          fmt_id[4];      // fmt块ID（"fmt "，4字节）DWORD         fmt_size;       // fmt块大小（4字节）PCMWAVEFORMAT fmt;            // fmt块内容（24字节）DATA_CHUNK    data;           // data块（8字节）
} __attribute__((packed)) WAV_HEADER;  // 总大小：12+4+4+24+8=52字节

解析头部信息的函数

根据上面的结构体，一个一个chuck来解析，适应有额外元素和LIST Chuck的情况

通过参数传出解析出来的采样率，通道数，位深

// 解析WAV文件头
int ParseWav(const char* filename, unsigned int* sample_rate, unsigned short* channels, unsigned short* bits_per_sample) 
{if (!sample_rate || !channels || !bits_per_sample) {fprintf(stderr, "参数错误：指针不能为空\n");return -1;}FILE* fp = fopen(filename, "rb");if (!fp) {perror("文件打开失败");return -1;}// 解析RIFF块RIFF_CHUNK riff;size_t read_size = fread(&riff, 1, sizeof(RIFF_CHUNK), fp);if (read_size != sizeof(RIFF_CHUNK)) {fprintf(stderr, "RIFF块读取失败（实际%zu字节，预期%zu字节）\n", read_size, sizeof(RIFF_CHUNK));fclose(fp);return -1;}printf("检测到的标识: ChunkID=%.4s, Format=%.4s\n", riff.ChunkID, riff.Format);if (memcmp(riff.ChunkID, "RIFF", 4) != 0 || memcmp(riff.Format, "WAVE", 4) != 0) {fprintf(stderr, "不是有效的WAV文件\n");fclose(fp);return -1;}// 查找并解析fmt块PCMWAVEFORMAT pcmFmt = {0};FACT_CHUNK fact = {0};int foundFmt = 0, foundFact = 0;CHAR subchunkID[4];DWORD subchunkSize;while (1) {if (fread(subchunkID, 1, 4, fp) != 4 || fread(&subchunkSize, sizeof(DWORD), 1, fp) != 1) {perror("子块信息读取失败");fclose(fp);return -1;}if (memcmp(subchunkID, "fmt ", 4) == 0) {if (subchunkSize < sizeof(PCMWAVEFORMAT)) {fprintf(stderr, "fmt块大小异常（%u字节 < %zu字节）\n", subchunkSize, sizeof(PCMWAVEFORMAT));fclose(fp);return -1;}if (fread(&pcmFmt, sizeof(PCMWAVEFORMAT), 1, fp) != 1) {perror("fmt块内容读取失败");fclose(fp);return -1;}if (subchunkSize > sizeof(PCMWAVEFORMAT)) {fseek(fp, subchunkSize - sizeof(PCMWAVEFORMAT), SEEK_CUR);}// 提取参数*sample_rate = pcmFmt.wf.nSamplesPerSec;*channels = pcmFmt.wf.nChannels;*bits_per_sample = pcmFmt.wBitsPerSample;foundFmt = 1;break;}else if (memcmp(subchunkID, "fact", 4) == 0) {foundFact = 1;memcpy(fact.SubchunkID, subchunkID, 4);fact.SubchunkSize = subchunkSize;if (subchunkSize >= sizeof(DWORD)) {fread(&fact.dwSampleLength, sizeof(DWORD), 1, fp);}if (subchunkSize > sizeof(DWORD)) {fseek(fp, subchunkSize - sizeof(DWORD), SEEK_CUR);}}else {fseek(fp, subchunkSize, SEEK_CUR);}if (ftell(fp) > (long)(riff.ChunkSize + 8)) {fprintf(stderr, "未找到fmt块\n");fclose(fp);return -1;}}// 解析data块DATA_CHUNK dataChunk = {0};int foundData = 0;while (1) {if (fread(dataChunk.SubchunkID, 1, 4, fp) != 4 || fread(&dataChunk.SubchunkSize, sizeof(DWORD), 1, fp) != 1) {perror("data块信息读取失败");fclose(fp);return -1;}if (memcmp(dataChunk.SubchunkID, "data", 4) == 0) {foundData = 1;break;} else if (memcmp(dataChunk.SubchunkID, "fact", 4) == 0 && !foundFact) {foundFact = 1;fact.SubchunkSize = dataChunk.SubchunkSize;if (dataChunk.SubchunkSize >= sizeof(DWORD)) {fread(&fact.dwSampleLength, sizeof(DWORD), 1, fp);}if (dataChunk.SubchunkSize > sizeof(DWORD)) {fseek(fp, dataChunk.SubchunkSize - sizeof(DWORD), SEEK_CUR);}}else {fseek(fp, dataChunk.SubchunkSize, SEEK_CUR);}if (ftell(fp) > (long)(riff.ChunkSize + 8)) {fprintf(stderr, "未找到data块\n");fclose(fp);return -1;}}// 打印头信息printf("===== WAV 文件头信息 =====\n");printf("RIFF 块:\n");printf("  标识: %.4s\n", riff.ChunkID);printf("  总大小: %u 字节(不含 RIFF 头)\n", riff.ChunkSize);printf("  格式: %.4s\n", riff.Format);printf("\nWAVEFORMAT:\n");printf("  编码格式: %u (%s)\n", pcmFmt.wf.wFormatTag,pcmFmt.wf.wFormatTag == 1 ? "WAVE_FORMAT_PCM" : "非 PCM");printf("  声道数: %u\n", pcmFmt.wf.nChannels);printf("  采样率: %u Hz\n", pcmFmt.wf.nSamplesPerSec);printf("  平均字节率: %u 字节/秒\n", pcmFmt.wf.nAvgBytesPerSec);printf("  块对齐: %u 字节\n", pcmFmt.wf.nBlockAlign);printf("\nPCMWAVEFORMAT:\n");printf("  采样位数: %u bits\n", pcmFmt.wBitsPerSample);if (foundFact) {printf("\nfact 块:\n");printf("  标识: %.4s\n", fact.SubchunkID);printf("  块大小: %u 字节\n", fact.SubchunkSize);printf("  总采样帧数: %u\n", fact.dwSampleLength);printf("  音频时长: %.2f 秒（基于采样帧）\n",(float)fact.dwSampleLength / pcmFmt.wf.nSamplesPerSec);}printf("\ndata 块:\n");printf("  标识: %.4s\n", dataChunk.SubchunkID);printf("  数据大小: %u 字节\n", dataChunk.SubchunkSize);if (!foundFact) {printf("  音频时长: %.2f 秒(估算)\n",(float)dataChunk.SubchunkSize / pcmFmt.wf.nAvgBytesPerSec);}fclose(fp);return 0;
}

设置声卡参数

录制和播放都需要设置声卡

// 设置PCM硬件参数
static int set_hw_params(snd_pcm_t *pcmHandler, unsigned int sample_rate, unsigned short channels, unsigned short bits_per_sample) 
{snd_pcm_hw_params_t *hwParams;int ret;int dir;snd_pcm_format_t pcm_format;// 映射采样位数到ALSA格式if (bits_per_sample == 8)pcm_format = SND_PCM_FORMAT_U8;else if (bits_per_sample == 16)pcm_format = SND_PCM_FORMAT_S16_LE;else if (bits_per_sample == 24)pcm_format = SND_PCM_FORMAT_S24_LE;else if (bits_per_sample == 32)pcm_format = SND_PCM_FORMAT_S32_LE;else{fprintf(stderr, "不支持的采样位数: %u bits\n", bits_per_sample);return -1;}snd_pcm_hw_params_malloc(&hwParams);do {ret = snd_pcm_hw_params_any(pcmHandler, hwParams);if (ret < 0) {fprintf(stderr, "snd_pcm_hw_params_any error: %s\n", snd_strerror(ret));break;}ret = snd_pcm_hw_params_set_access(pcmHandler, hwParams, SND_PCM_ACCESS_RW_INTERLEAVED);if (ret < 0) {fprintf(stderr, "set_access error: %s\n", snd_strerror(ret));break;}ret = snd_pcm_hw_params_set_format(pcmHandler, hwParams, pcm_format);if (ret < 0) {fprintf(stderr, "set_format error: %s\n", snd_strerror(ret));break;}ret = snd_pcm_hw_params_set_channels(pcmHandler, hwParams, channels);if (ret < 0) {fprintf(stderr, "set_channels error: %s\n", snd_strerror(ret));break;}ret = snd_pcm_hw_params_set_rate_near(pcmHandler, hwParams, &sample_rate, &dir);if (ret < 0) {fprintf(stderr, "set_rate error: %s\n", snd_strerror(ret));break;}printf("实际设置的采样率: %u Hz\n", sample_rate);unsigned int period_time = 10000;  // 10ms周期ret = snd_pcm_hw_params_set_period_time_near(pcmHandler, hwParams, &period_time, &dir);if (ret < 0) {fprintf(stderr, "set_period_time error: %s\n", snd_strerror(ret));break;}ret = snd_pcm_hw_params(pcmHandler, hwParams);if (ret < 0) {fprintf(stderr, "apply hw_params error: %s\n", snd_strerror(ret));break;}snd_pcm_hw_params_free(hwParams);return 0;} while (0);snd_pcm_hw_params_free(hwParams);return -1;
}

Record 麦克风输入为WAV文件

因为需要保存在标准的WAV文件中，所以需要先写入头部数据。

写头部数据

写入各个字段

// 写入WAV文件头
void write_wav_header(int fd, unsigned int sample_rate, unsigned short channels, unsigned short bits_per_sample, int duration) 
{WAV_HEADER header = {0};DWORD byte_rate = sample_rate * channels * (bits_per_sample / 8);WORD block_align = channels * (bits_per_sample / 8);DWORD data_size = duration * byte_rate;  // 预估数据大小// 填充RIFF块memcpy(header.riff.ChunkID, "RIFF", 4);header.riff.ChunkSize = data_size + sizeof(WAV_HEADER) - 8;  // 总大小计算memcpy(header.riff.Format, "WAVE", 4);// 填充fmt块memcpy(header.fmt_id, "fmt ", 4);header.fmt_size = sizeof(PCMWAVEFORMAT);header.fmt.wf.wFormatTag = 1;  // PCM格式header.fmt.wf.nChannels = channels;header.fmt.wf.nSamplesPerSec = sample_rate;header.fmt.wf.nAvgBytesPerSec = byte_rate;header.fmt.wf.nBlockAlign = block_align;header.fmt.wBitsPerSample = bits_per_sample;// 填充data块memcpy(header.data.SubchunkID, "data", 4);header.data.SubchunkSize = data_size;// 写入完整头部write(fd, &header, sizeof(WAV_HEADER));
}

读声卡数据，存为WAV文件

// 录音函数
int RecordAudio(const char *filename, int duration) 
{printf("Calling RecordAudio() filename = %s, duration = %d\n", filename, duration);unsigned int sample_rate = 48000;unsigned short channels = 2;unsigned short bits_per_sample = 16;snd_pcm_t *pcm_handle = NULL;snd_pcm_uframes_t period_size;unsigned char *buff = NULL;unsigned int buf_bytes;int ret = -1;int fd = -1;snd_pcm_hw_params_t *temp_hw_params = NULL;do {// 打开PCM捕获设备ret = snd_pcm_open(&pcm_handle, "hw:0,0", SND_PCM_STREAM_CAPTURE, 0);if (ret < 0) {printf("Open PCM device error: %s\n", snd_strerror(ret));break;}// 设置硬件参数if (set_hw_params(pcm_handle, sample_rate, channels, bits_per_sample) < 0) {snd_pcm_close(pcm_handle);break;}// 获取周期大小snd_pcm_hw_params_malloc(&temp_hw_params);ret = snd_pcm_hw_params_current(pcm_handle, temp_hw_params);if (ret < 0) {fprintf(stderr, "get current hw_params error: %s\n", snd_strerror(ret));break;}snd_pcm_hw_params_get_period_size(temp_hw_params, &period_size, NULL);snd_pcm_hw_params_free(temp_hw_params);// 分配缓冲区buf_bytes = period_size * channels * (bits_per_sample / 8);buff = (unsigned char *)malloc(buf_bytes);if (buff == NULL) {printf("Malloc buffer failed\n");break;}// 打开文件fd = open(filename, O_WRONLY | O_CREAT | O_TRUNC, 0644);if (fd < 0) {perror("Open file failed");break;}// 写入WAV头write_wav_header(fd, sample_rate, channels, bits_per_sample, duration);// 录音循环unsigned int total_periods = (duration * sample_rate) / period_size;for (unsigned int i = 0; i < total_periods; i++) {ret = snd_pcm_readi(pcm_handle, buff, period_size);if (ret < 0) {fprintf(stderr, "snd_pcm_readi error: %s\n", snd_strerror(ret));ret = snd_pcm_recover(pcm_handle, ret, 0);if (ret < 0) {fprintf(stderr, "Recover PCM failed: %s\n", snd_strerror(ret));break;}continue;}int write_len = ret * channels * (bits_per_sample / 8);if (write(fd, buff, write_len) != write_len) {perror("Write file failed");break;}}// 修正文件头if (fd != -1 && ret >= 0) {off_t total_file_size = lseek(fd, 0, SEEK_END);DWORD real_data_size = total_file_size - sizeof(WAV_HEADER);DWORD real_riff_size = real_data_size + sizeof(WAV_HEADER) - 8;lseek(fd, offsetof(WAV_HEADER, data.SubchunkSize), SEEK_SET);write(fd, &real_data_size, sizeof(real_data_size));lseek(fd, offsetof(WAV_HEADER, riff.ChunkSize), SEEK_SET);write(fd, &real_riff_size, sizeof(real_riff_size));printf("录音完成，实际数据大小: %u 字节\n", real_data_size);}if (ret >= 0) printf("Recording completed successfully.\n");ret = 0;} while (0);// 资源清理if (fd != -1) close(fd);if (buff != NULL) free(buff);if (pcm_handle != NULL) snd_pcm_close(pcm_handle);return ret;
}

播放WAV文件

1. 先读取WAV文件头部信息，读出采样率，通道数和位深
2. 然后根据读出的数据，设置声卡
3. 然后读取文件数据，写到声卡中播放

// 播放函数
int PlayWavFile(const char * cstrFileName)
{printf("Start Play Wav File %s\n", cstrFileName);// 解析WAV参数unsigned int sample_rate;unsigned short channels, bits_per_sample;if (ParseWav(cstrFileName, &sample_rate, &channels, &bits_per_sample) != 0){fprintf(stderr, "WAV文件解析失败，无法播放\n");return -1;}// 打开WAV文件FILE *file = fopen(cstrFileName, "rb");if (!file) {perror("fopen");return 1;}// 定位到音频数据（跳过WAV头）fseek(file, sizeof(WAV_HEADER), SEEK_SET);snd_pcm_t *handle;snd_pcm_hw_params_t *params;snd_pcm_uframes_t period_size;int err;// 打开播放设备err = snd_pcm_open(&handle, "default", SND_PCM_STREAM_PLAYBACK, 0);if (err < 0) {fprintf(stderr, "cannot open audio device %s (%s)\n", "default", snd_strerror(err));fclose(file);return 1;}// 设置播放参数snd_pcm_hw_params_alloca(&params);err = snd_pcm_hw_params_any(handle, params);if (err < 0){fprintf(stderr, "snd_pcm_hw_params_any error: %s\n", snd_strerror(err));snd_pcm_close(handle);fclose(file);return 1;}if (set_hw_params(handle, sample_rate, channels, bits_per_sample) < 0){snd_pcm_close(handle);fclose(file);return 1;}// 获取周期大小snd_pcm_hw_params_get_period_size(params, &period_size, NULL);unsigned int buf_bytes = period_size * channels * (bits_per_sample / 8);unsigned char *buffer = (unsigned char *)malloc(buf_bytes);if (!buffer) {fprintf(stderr, "malloc failed\n");snd_pcm_close(handle);fclose(file);return 1;}// 播放循环while (1) {size_t read_bytes = fread(buffer, 1, buf_bytes, file);if (read_bytes == 0) {if (feof(file)) {printf("播放完成（已到文件末尾）\n");break;} else {fprintf(stderr, "fread failed\n");break;}}snd_pcm_uframes_t play_frames = read_bytes / (channels * (bits_per_sample / 8));err = snd_pcm_writei(handle, buffer, play_frames);if (err < 0) {fprintf(stderr, "write error: %s\n", snd_strerror(err));err = snd_pcm_recover(handle, err, 0);if (err < 0) {fprintf(stderr, "Recover PCM failed: %s\n", snd_strerror(err));break;}continue;}}// 清理资源free(buffer);snd_pcm_drain(handle);snd_pcm_close(handle);fclose(file);return 0;
}

调用函数

// 检查是否为WAV文件
bool IsAudioFile(const std::string &strFileName)
{size_t pos = strFileName.rfind('.');if (pos == std::string::npos) return false;std::string extension = strFileName.substr(pos);std::transform(extension.begin(), extension.end(), extension.begin(), ::tolower);return extension == ".wav";
}int main(int argc, char *argv[]) 
{if (argc == 1) {RecordAudio("Record.wav", 10);  // 录制10秒return 0;} else if (argc == 2) {char cstrArg[256] = {0};strncpy(cstrArg, argv[1], sizeof(cstrArg)-1);if (IsAudioFile(cstrArg)) {PlayWavFile(cstrArg);}else{fprintf(stderr, "仅支持WAV格式文件\n");ShowHelpInfo();}} else {ShowHelpInfo();}return 0;
}