绝大数人都知道mp3格式编码,以及aac,amr等压缩格式编码。
而在语音通信界有一个强悍的音频格式编码opus.
经过实测,压缩比最高可以达到1:10。
100KB 压缩后 10KB
虽然是有损压缩,
但是根据实际对比试听,
几乎听不出差别。
而且还原度还比mp3高,压缩比也比mp3高。
用来压缩传输音频,绝对是一大杀器。
项目官方地址:
https://opus-codec.org/
维基上的描述:
Opus是一个有损声音编码的格式,由Xiph.Org基金会开发,之后由互联网工程任务组(IETF)进行标准化,目标用希望用单一格式包含声音和语音,取代Speex和Vorbis,且适用于网络上低延迟的即时声音传输,标准格式定义于RFC 6716文件。Opus格式是一个开放格式,使用上没有任何专利或限制。
Opus集成了两种声音编码的技术:以语音编码为导向的SILK和低延迟的CELT。Opus可以无缝调节高低比特率。在编码器内部它在较低比特率时使用线性预测编码在高比特率时候使用变换编码(在高低比特率交界处也使用两者结合的编码方式)。Opus具有非常低的算法延迟(默认为22.5 ms),非常适合用于低延迟语音通话的编码,像是网络上的即时声音流、即时同步声音旁白等等,此外Opus也可以通过降低编码比特率,达成更低的算法延迟,最低可以到5 ms。在多个听觉盲测中,Opus都比MP3、AAC、HE-AAC等常见格式,有更低的延迟和更好的声音压缩率。
更重要的是要看这条:
Opus被提出用于在IETF上标准化新的音频格式,最终被IETF编解码器工作组接受和授予。它基于Xiph.Org基金会和Skype技术公司两项最初分开的标准提案。其主要开发人员包括Jean-Marc Valin(Xiph.Org,Octasic,Mozilla Corporation),Koen Vos(Skype)和Timothy B. Terriberry(Xiph.Org,Mozilla Corporation)。其中包括Juin-Hwey(Raymond)Chen(Broadcom),Gregory Maxwell(Xiph.Org,Wikimedia)和Christopher Montgomery(Xiph.Org)也参与其中。
这个项目被几大公司加持,也难怪能有如此出色的压缩比。
由于是纯C代码,比较好理解。
抽了点空,俺写个wav压缩解压的示例代码,
分享出来,权当抛砖引玉。
示例代码位置:
https://github.com/cpuimage/opus/blob/master/example/opus.cpp
贴上完整C++代码:
#include#include #include #include #include // https://github.com/mackron/dr_libs/blob/master/dr_wav.h #define DR_WAV_IMPLEMENTATION #include "dr_wav.h" #define FRAME_SIZE 480 #define MAX_FRAME_SIZE (6*FRAME_SIZE) #define MAX_CHANNELS 1 #define MAX_PACKET_SIZE (3*1276) #pragma pack(push) #pragma pack(1) struct WavInfo { uint16_t channels; uint32_t sampleRate; uint32_t bitsPerSample; }; #pragma pack(pop) #ifndef nullptr #define nullptr NULL #endif class FileStream { public: FileStream() { cur_pos = 0; } void Append(const char *data, size_t size) { if (cur_pos + size > Size()) { vec.resize(cur_pos + size); } memcpy(vec.data() + cur_pos, data, size); cur_pos += size; } void AppendU32(uint32_t val) { Append((char *) (&val), sizeof(val)); } char *Data() { return vec.data(); } size_t Size() { return vec.size(); } size_t Read(void *buff, size_t elemSize, size_t elemCount) { size_t readed = std::min((vec.size() - cur_pos), (elemCount * elemSize)) / elemSize; if (readed > 0) { memcpy(buff, vec.data() + cur_pos, readed * elemSize); cur_pos += readed * elemSize; } return readed; } bool SeekCur(int offset) { if (cur_pos + offset > vec.size()) { cur_pos = !vec.empty() ? (vec.size() - 1) : 0; return false; } else { cur_pos += offset; return true; } } bool SeekBeg(int offset = 0) { cur_pos = 0; return SeekCur(offset); } bool WriteToFile(const char *filename) { FILE *fin = fopen(filename, "wb"); if (!fin) { return false; } fseek(fin, 0, SEEK_SET); fwrite(vec.data(), sizeof(char), vec.size(), fin); fclose(fin); return true; } bool ReadFromFile(const char *filename) { FILE *fin = fopen(filename, "rb"); if (!fin) { return false; } fseek(fin, 0, SEEK_END); long fileSize = ftell(fin); vec.resize(static_cast long long int>(fileSize)); fseek(fin, 0, SEEK_SET); fread(vec.data(), sizeof(char), vec.size(), fin); fclose(fin); return true; } private: std::vector<char> vec; size_t cur_pos; }; bool Wav2Opus(FileStream *input, FileStream *output); bool Opus2Wav(FileStream *input, FileStream *output); bool wav2stream(char *input, FileStream *output); bool stream2wav(FileStream *input, char *output); bool wavWrite_int16(char *filename, int16_t *buffer, int sampleRate, uint32_t totalSampleCount) { drwav_data_format format = {}; format.container = drwav_container_riff; // <-- drwav_container_riff = normal WAV files, drwav_container_w64 = Sony Wave64. format.format = DR_WAVE_FORMAT_PCM; // <-- Any of the DR_WAVE_FORMAT_* codes. format.channels = 1; format.sampleRate = (drwav_uint32) sampleRate; format.bitsPerSample = 16; drwav *pWav = drwav_open_file_write(filename, &format); if (pWav) { drwav_uint64 samplesWritten = drwav_write(pWav, totalSampleCount, buffer); drwav_uninit(pWav); if (samplesWritten != totalSampleCount) { fprintf(stderr, "ERROR\n"); return false; } return true; } return false; } int16_t *wavRead_int16(char *filename, uint32_t *sampleRate, uint64_t *totalSampleCount) { unsigned int channels; int16_t *buffer = drwav_open_and_read_file_s16(filename, &channels, sampleRate, totalSampleCount); if (buffer == nullptr) { fprintf(stderr, "ERROR\n"); return nullptr; } if (channels != 1) { drwav_free(buffer); buffer = nullptr; *sampleRate = 0; *totalSampleCount = 0; } return buffer; } bool wav2stream(char *input, FileStream *output) { uint32_t sampleRate = 0; uint64_t totalSampleCount = 0; int16_t *wavBuffer = wavRead_int16(input, &sampleRate, &totalSampleCount); if (wavBuffer == nullptr) return false; WavInfo info = {}; info.bitsPerSample = 16; info.sampleRate = sampleRate; info.channels = 1; output->SeekBeg(); output->Append((char *) &info, sizeof(info)); output->Append((char *) wavBuffer, totalSampleCount * sizeof(int16_t)); free(wavBuffer); return true; } bool stream2wav(FileStream *input, char *output) { WavInfo info = {}; input->SeekBeg(); size_t read = input->Read(&info, sizeof(info), 1); if (read != 1) { return false; } size_t totalSampleCount = (input->Size() - sizeof(info)) / 2; return wavWrite_int16(output, (int16_t *) (input->Data() + sizeof(info)), info.sampleRate, static_cast (totalSampleCount)); } bool Wav2Opus(FileStream *input, FileStream *output) { WavInfo in_info = {}; input->SeekBeg(); size_t read = input->Read(&in_info, sizeof(in_info), 1); if (read != 1) { return false; } uint32_t bitsPerSample = in_info.bitsPerSample; uint32_t sampleRate = in_info.sampleRate; uint16_t channels = in_info.channels; int err = 0; if (channels > MAX_CHANNELS) { return false; } OpusEncoder *encoder = opus_encoder_create(sampleRate, channels, OPUS_APPLICATION_AUDIO, &err); if (!encoder || err < 0) { fprintf(stderr, "failed to create an encoder: %s\n", opus_strerror(err)); if (!encoder) { opus_encoder_destroy(encoder); } return false; } const uint16_t *data = (uint16_t *) (input->Data() + sizeof(in_info)); size_t size = (input->Size() - sizeof(in_info)) / 2; opus_int16 pcm_bytes[FRAME_SIZE * MAX_CHANNELS]; size_t index = 0; size_t step = static_cast (FRAME_SIZE * channels); FileStream encodedData; unsigned char cbits[MAX_PACKET_SIZE]; size_t frameCount = 0; size_t readCount = 0; while (index < size) { memset(&pcm_bytes, 0, sizeof(pcm_bytes)); if (index + step <= size) { memcpy(pcm_bytes, data + index, step * sizeof(uint16_t)); index += step; } else { readCount = size - index; memcpy(pcm_bytes, data + index, (readCount) * sizeof(uint16_t)); index += readCount; } int nbBytes = opus_encode(encoder, pcm_bytes, channels * FRAME_SIZE, cbits, MAX_PACKET_SIZE); if (nbBytes < 0) { fprintf(stderr, "encode failed: %s\n", opus_strerror(nbBytes)); break; } ++frameCount; encodedData.AppendU32(static_cast (nbBytes)); encodedData.Append((char *) cbits, static_cast (nbBytes)); } WavInfo info = {}; info.bitsPerSample = bitsPerSample; info.sampleRate = sampleRate; info.channels = channels; output->SeekBeg(); output->Append((char *) &info, sizeof(info)); output->Append(encodedData.Data(), encodedData.Size()); opus_encoder_destroy(encoder); return true; } bool Opus2Wav(FileStream *input, FileStream *output) { WavInfo info = {}; input->SeekBeg(); size_t read = input->Read(&info, sizeof(info), 1); if (read != 1) { return false; } int channels = info.channels; if (channels > MAX_CHANNELS) { return false; } output->SeekBeg(); output->Append((char *) &info, sizeof(info)); int err = 0; OpusDecoder *decoder = opus_decoder_create(info.sampleRate, channels, &err); if (!decoder || err < 0) { fprintf(stderr, "failed to create decoder: %s\n", opus_strerror(err)); if (!decoder) { opus_decoder_destroy(decoder); } return false; } unsigned char cbits[MAX_PACKET_SIZE]; opus_int16 out[MAX_FRAME_SIZE * MAX_CHANNELS]; int frameCount = 0; while (true) { uint32_t nbBytes; size_t readed = input->Read(&nbBytes, sizeof(uint32_t), 1); if (readed == 0) { break; } if (nbBytes > sizeof(cbits)) { fprintf(stderr, "nbBytes > sizeof(cbits)\n"); break; } readed = input->Read(cbits, sizeof(char), nbBytes); if (readed != nbBytes) { fprintf(stderr, "readed != nbBytes\n"); break; } int frame_size = opus_decode(decoder, cbits, nbBytes, out, MAX_FRAME_SIZE, 0); if (frame_size < 0) { fprintf(stderr, "decoder failed: %s\n", opus_strerror(frame_size)); break; } ++frameCount; output->Append((char *) out, channels * frame_size * sizeof(out[0])); } opus_decoder_destroy(decoder); return true; } void splitpath(const char *path, char *drv, char *dir, char *name, char *ext) { const char *end; const char *p; const char *s; if (path[0] && path[1] == ':') { if (drv) { *drv++ = *path++; *drv++ = *path++; *drv = '\0'; } } else if (drv) *drv = '\0'; for (end = path; *end && *end != ':';) end++; for (p = end; p > path && *--p != '\\' && *p != '/';) if (*p == '.') { end = p; break; } if (ext) for (s = end; (*ext = *s++);) ext++; for (p = end; p > path;) if (*--p == '\\' || *p == '/') { p++; break; } if (name) { for (s = p; s < end;) *name++ = *s++; *name = '\0'; } if (dir) { for (s = path; s < p;) *dir++ = *s++; *dir = '\0'; } } void opus2wav(const char *in_file, char *out_file) { FileStream input; FileStream output; input.ReadFromFile(in_file); Opus2Wav(&input, &output); stream2wav(&output, out_file); } void wav2opus(char *in_file, char *out_file) { FileStream input; FileStream output; wav2stream(in_file, &input); Wav2Opus(&input, &output); output.WriteToFile(out_file); } int main(int argc, char *argv[]) { printf("Opus Demo\n"); printf("blog:http://cpuimage.cnblogs.com/\n"); printf("e-mail:[email protected]\n"); if (argc < 2) return -1; char *in_file = argv[1]; char drive[3]; char dir[256]; char fname[256]; char ext[256]; char out_file[1024]; splitpath(in_file, drive, dir, fname, ext); if (memcmp(".wav", ext, strlen(ext)) == 0) { sprintf(out_file, "%s%s%s.out", drive, dir, fname); wav2opus(in_file, out_file); } else if (memcmp(".out", ext, strlen(ext)) == 0) { sprintf(out_file, "%s%s%s_out.wav", drive, dir, fname); opus2wav(in_file, out_file); } printf("done.\n"); printf("press any key to exit.\n"); getchar(); return 0; }
项目地址:
https://github.com/cpuimage/opus
示例具体流程为:
1.压缩
加载wav(拖放wav文件到可执行文件上)->压缩->保存为out
2.解压
加载out(拖放out文件到可执行文件上)->解压->保存为wav
示例比较简单,用cmake即可进行编译示例代码,详情见CMakeLists.txt。
若有其他相关问题或者需求也可以邮件联系俺探讨。
邮箱地址是:
[email protected]