FFmpeg - 为什么从pcm提取音频，编码为mp3以后，时长变多了，声音也颤抖
```
#include <memory>
extern "C" {
#include "libavcodec/avcodec.h"
#include "libavformat/avformat.h"
#include "libavutil/imgutils.h"
#include "libswresample/swresample.h"
#include "libavutil/opt.h"
#include "libavutil/avassert.h"
#include "libavutil/audio_fifo.h"
}
// 函数退出后执行
#define DEFER(T, P, Fn) std::shared_ptr<T> P##deletor(P, [&](T *) {Fn;})
// 打印异常信息，并退出main函数
#define FATAL(M, ...) printf(M, ##__VA_ARGS__); return -1
// 自定义变量
const char *src_media = "E:/2.pcm";
const char *dst_media = "E:/3.mp3";
// PCM的原始参数
#define PCM_IN_FORMAT AV_SAMPLE_FMT_S16
#define PCM_IN_CHANNEL_LAYOUT AV_CH_LAYOUT_MONO
#define PCM_IN_NB_SAMPLES 1024
#define PCM_IN_SAMPLE_RATE 44100
int main(int argc, char **argv) {
// 申请一个输出的上下文
AVFormatContext *fmt_ctx = nullptr;
avformat_alloc_output_context2(&fmt_ctx, nullptr, nullptr, dst_media);
if (fmt_ctx == nullptr) {
FATAL("alloc output format context failed.");
}
DEFER(AVFormatContext, fmt_ctx, avio_closep(&fmt_ctx->pb);
avformat_close_input(&fmt_ctx));
// 查询编码器
AVCodec *audio_enc;
if ((audio_enc = avcodec_find_encoder(fmt_ctx->oformat->audio_codec)) == nullptr) {
FATAL("find audio encoder failed.");
}
AVStream *audio_stream = avformat_new_stream(fmt_ctx, audio_enc);
audio_stream->id = fmt_ctx->nb_streams - 1;
// 为编码器申请上下文
AVCodecContext *audio_enc_ctx = nullptr;
if ((audio_enc_ctx = avcodec_alloc_context3(audio_enc)) == nullptr) {
FATAL("allocate audio enc context failed.");
}
DEFER(AVCodecContext, audio_enc_ctx, avcodec_free_context(&audio_enc_ctx));
// 为编码器配置编码参数
audio_enc_ctx->sample_fmt = audio_enc->sample_fmts ? audio_enc->sample_fmts[0] : AV_SAMPLE_FMT_FLTP;
audio_enc_ctx->bit_rate = 128000;
audio_enc_ctx->sample_rate = 44100;
audio_enc_ctx->channel_layout = AV_CH_LAYOUT_STEREO;
audio_enc_ctx->channels = av_get_channel_layout_nb_channels(audio_enc_ctx->channel_layout);
if (fmt_ctx->oformat->flags & AVFMT_GLOBALHEADER) {
audio_enc_ctx->flags |= AV_CODEC_FLAG_GLOBAL_HEADER;
}
audio_stream->time_base = AVRational{1, audio_enc_ctx->sample_rate};
// 打开编码器
if (avcodec_open2(audio_enc_ctx, audio_enc, nullptr) < 0) {
FATAL("codec open failed.");
}
if (avcodec_parameters_from_context(audio_stream->codecpar, audio_enc_ctx) < 0) {
FATAL("copy params failed.");
}
if (avio_open(&fmt_ctx->pb, dst_media, AVIO_FLAG_WRITE) < 0) {
FATAL("open dst file failed.");
}
// 写入头信息
if (avformat_write_header(fmt_ctx, nullptr) < 0) {
FATAL("write header failed.");
}
// 申请一个音频frame
AVFrame *audio_frame = nullptr;
if ((audio_frame = av_frame_alloc()) == nullptr) {
FATAL("allocate frame failed.");
}
DEFER(AVFrame, audio_frame, av_frame_free(&audio_frame));
audio_frame->format = audio_enc_ctx->sample_fmt;
audio_frame->channel_layout = audio_enc_ctx->channel_layout;
audio_frame->nb_samples = audio_enc_ctx->frame_size;
audio_frame->sample_rate = audio_enc_ctx->sample_rate;
// 给frame的data和size分配空间
if (av_frame_get_buffer(audio_frame, 0) < 0) {
FATAL("allocate frame data failed.");
}
// 创建swr的上下文
SwrContext *swr_cxt = swr_alloc();
DEFER(SwrContext, swr_cxt, swr_free(&swr_cxt));
// 获取输入与输出的采样格式
enum AVSampleFormat in_sample_fmt = PCM_IN_FORMAT;
enum AVSampleFormat out_sample_fmt = audio_enc_ctx->sample_fmt;
// 获取输入与输出的采样率
int in_sample_rate = PCM_IN_SAMPLE_RATE;
int out_sample_rate = audio_enc_ctx->sample_rate;
// 获取输入与输出的声道格式
uint64_t in_ch_layout = PCM_IN_CHANNEL_LAYOUT;
uint64_t out_ch_layout = audio_enc_ctx->channel_layout;
// 设置swr的配置项
swr_alloc_set_opts(swr_cxt, out_ch_layout, out_sample_fmt, out_sample_rate,
in_ch_layout, in_sample_fmt, in_sample_rate, 0, nullptr);
// 初始化swr
swr_init(swr_cxt);
// 申请一个packet，并初始化
AVPacket pkt;
av_init_packet(&pkt);
pkt.data = nullptr;
pkt.size = 0;
FILE *input = nullptr;
if ((input = fopen(src_media, "rb")) == nullptr) {
FATAL("no readable file.");
}
DEFER(FILE, input, fclose(input));
// 从文件中读取出来的数据，临时存储在该帧中
AVFrame *temp_frame = av_frame_alloc();
DEFER(AVFrame, temp_frame, av_frame_free(&temp_frame));
temp_frame->format = PCM_IN_FORMAT;
temp_frame->nb_samples = PCM_IN_NB_SAMPLES;
temp_frame->channel_layout = PCM_IN_CHANNEL_LAYOUT;
temp_frame->channels = av_get_channel_layout_nb_channels(PCM_IN_CHANNEL_LAYOUT);
// 给frame的data和size分配空间
if (av_frame_get_buffer(temp_frame, 0) < 0) {
FATAL("allocate frame data failed.");
}
// 循环读取frame数据
int frame_count = 0;
while (true) {
if (av_frame_make_writable(audio_frame) < 0) {
FATAL("frame is not writable");
}
if (fread(temp_frame->data[0], 1, (size_t) temp_frame->linesize[0], input) < 0) {
FATAL("read input file failed.");
} else if (feof(input)) {
break;
}
swr_convert(swr_cxt, audio_frame->data, audio_frame->nb_samples, (const uint8_t **) temp_frame->data, temp_frame->nb_samples);
audio_frame->pts = av_rescale_q(frame_count, (AVRational){1, audio_enc_ctx->sample_rate}, audio_enc_ctx->time_base);
frame_count += audio_frame->nb_samples;
// 发送一个frame
if (avcodec_send_frame(audio_enc_ctx, audio_frame) < 0) {
FATAL("send frame exception.");
}
// 接受编码完的内容
while (true) {
auto packet_ret = avcodec_receive_packet(audio_enc_ctx, &pkt);
// 判断是否完全接受了packet
if (packet_ret == AVERROR(EAGAIN) || packet_ret == AVERROR_EOF) {
break;
}
// 检查是否接受异常
if (packet_ret < 0) {
FATAL("receive packet exception.");
}
av_packet_rescale_ts(&pkt, audio_enc_ctx->time_base, audio_stream->time_base);
pkt.stream_index = audio_stream->index;
av_interleaved_write_frame(fmt_ctx, &pkt);
av_packet_unref(&pkt);
}
}
// 发送一个空的frame，告知编码器，已经编码完成，清空缓冲区
if (avcodec_send_frame(audio_enc_ctx, nullptr) < 0) {
FATAL("send frame exception.");
}
// 接受编码完的内容
while (true) {
auto packet_ret = avcodec_receive_packet(audio_enc_ctx, &pkt);
// 判断是否完全接受了packet
if (packet_ret == AVERROR(EAGAIN) || packet_ret == AVERROR_EOF) {
break;
}
// 检查是否接受异常
if (packet_ret < 0) {
FATAL("receive packet exception.");
}
av_packet_rescale_ts(&pkt, audio_enc_ctx->time_base, audio_stream->time_base);
pkt.stream_index = audio_stream->index;
av_interleaved_write_frame(fmt_ctx, &pkt);
av_packet_unref(&pkt);
}
av_write_trailer(fmt_ctx);
return 0;
}
```
### 问题已经解决，请看下面的代码
```cpp
#include <memory>
extern "C" {
#include "libavcodec/avcodec.h"
#include "libavformat/avformat.h"
#include "libswresample/swresample.h"
}
// 申请智能指针变量
#define NEW_PTR(T, P, V, Fn) T *P = V; std::shared_ptr<T> P##P(P, [&P](T *){if(P != nullptr){Fn;}})
// 打印异常信息，并退出main函数
#define FATAL(M, ...) printf(M, ##__VA_ARGS__); return -1
// 自定义变量
const char *src_media = "D:/2.pcm";
const char *dst_media = "D:/2.mp3";
// PCM的原始参数
#define PCM_IN_FORMAT AV_SAMPLE_FMT_S16
#define PCM_IN_CHANNELS 1
#define PCM_IN_SAMPLE_RATE 44100
int main(int argc, char **argv) {
// 申请一个输出的上下文
NEW_PTR(AVFormatContext, fmt_ctx, nullptr, avio_closep(&fmt_ctx->pb); avformat_close_input(&fmt_ctx));
avformat_alloc_output_context2(&fmt_ctx, nullptr, nullptr, dst_media);
if (fmt_ctx == nullptr) {
FATAL("alloc output format context failed.");
}
// 查询编码器
AVCodec *audio_enc;
if ((audio_enc = avcodec_find_encoder(fmt_ctx->oformat->audio_codec)) == nullptr) {
FATAL("find audio encoder failed.");
}
// 为编码器申请上下文
NEW_PTR(AVCodecContext, audio_enc_ctx, nullptr, avcodec_free_context(&audio_enc_ctx));
if ((audio_enc_ctx = avcodec_alloc_context3(audio_enc)) == nullptr) {
FATAL("allocate audio enc context failed.");
}
// 为编码器配置编码参数
audio_enc_ctx->sample_fmt = audio_enc->sample_fmts ? audio_enc->sample_fmts[0] : AV_SAMPLE_FMT_FLTP;
audio_enc_ctx->bit_rate = 64000;
audio_enc_ctx->sample_rate = 44100;
audio_enc_ctx->channel_layout = AV_CH_LAYOUT_STEREO;
audio_enc_ctx->channels = av_get_channel_layout_nb_channels(audio_enc_ctx->channel_layout);
if (fmt_ctx->oformat->flags & AVFMT_GLOBALHEADER) {
audio_enc_ctx->flags |= AV_CODEC_FLAG_GLOBAL_HEADER;
}
// 打开编码器
if (avcodec_open2(audio_enc_ctx, audio_enc, nullptr) < 0) {
FATAL("codec open failed.");
}
AVStream *audio_stream = avformat_new_stream(fmt_ctx, audio_enc);
audio_stream->id = fmt_ctx->nb_streams - 1;
audio_stream->time_base = AVRational{1, audio_enc_ctx->sample_rate};
if (avcodec_parameters_from_context(audio_stream->codecpar, audio_enc_ctx) < 0) {
FATAL("copy params failed.");
}
if (avio_open(&fmt_ctx->pb, dst_media, AVIO_FLAG_WRITE) < 0) {
FATAL("open dst file failed.");
}
// 写入头信息
if (avformat_write_header(fmt_ctx, nullptr) < 0) {
FATAL("write header failed.");
}
// 申请一个音频frame
NEW_PTR(AVFrame, audio_frame, nullptr, av_frame_free(&audio_frame));
if ((audio_frame = av_frame_alloc()) == nullptr) {
FATAL("allocate frame failed.");
}
audio_frame->format = audio_enc_ctx->sample_fmt;
audio_frame->channel_layout = audio_enc_ctx->channel_layout;
audio_frame->nb_samples = audio_enc_ctx->frame_size;
audio_frame->sample_rate = audio_enc_ctx->sample_rate;
if (av_frame_get_buffer(audio_frame, 0) < 0) {
FATAL("audio frame get buffer failed.");
}
// 创建一个frame，用来存储从pcm读取的数据
NEW_PTR(AVFrame, buf_frame, nullptr, av_frame_free(&buf_frame));
if ((buf_frame = av_frame_alloc()) == nullptr) {
FATAL("allocate buf frame failed.");
}
buf_frame->format = PCM_IN_FORMAT;
buf_frame->nb_samples = audio_frame->nb_samples;
buf_frame->channel_layout = (uint64_t) av_get_default_channel_layout(PCM_IN_CHANNELS);
buf_frame->sample_rate = PCM_IN_SAMPLE_RATE;
if (av_frame_get_buffer(buf_frame, 0) < 0) {
FATAL("create buf frame buffer failed.");
}
// 从pcm文件中读取适应音频帧的尺寸数据
auto readSize = av_samples_get_buffer_size(nullptr, buf_frame->channels, buf_frame->nb_samples,
(AVSampleFormat) buf_frame->format, 1);
NEW_PTR(uint8_t , buf, (uint8_t*)av_malloc((size_t)readSize), av_freep(&buf));
// 创建swr的上下文
NEW_PTR(SwrContext, swr_ctx, swr_alloc(), swr_free(&swr_ctx));
swr_alloc_set_opts(swr_ctx,
audio_frame->channel_layout,
(AVSampleFormat)audio_frame->format,
audio_frame->sample_rate,
av_get_default_channel_layout(PCM_IN_CHANNELS),
PCM_IN_FORMAT,
PCM_IN_SAMPLE_RATE,
0, nullptr);
swr_init(swr_ctx);
// 申请一个packet，并初始化
AVPacket pkt;
av_init_packet(&pkt);
pkt.data = nullptr;
pkt.size = 0;
NEW_PTR(FILE, input, nullptr, fclose(input));
if ((input = fopen(src_media, "rb")) == nullptr) {
FATAL("no readable file.");
}
// 循环读取frame数据
int audio_pts = 0;
while (true) {
// 用来编码的帧
AVFrame * encode_frame = nullptr;
if (fread(buf, 1, (size_t) readSize, input) < 0) {
FATAL("read input file failed.");
} else if (!feof(input)) {
// 文件没有到结尾，则获取编码帧
av_samples_fill_arrays(buf_frame->data, buf_frame->linesize, (const uint8_t*)buf, buf_frame->channels, buf_frame->nb_samples, (AVSampleFormat)buf_frame->format, 1);
swr_convert(swr_ctx, audio_frame->data, audio_frame->nb_samples, (const uint8_t**)buf_frame->data, buf_frame->nb_samples);
audio_frame->pts = audio_pts;
audio_pts += av_rescale_q(audio_frame->nb_samples, AVRational{1, audio_enc_ctx->frame_size},
audio_enc_ctx->time_base);
encode_frame = audio_frame;
} else {
// 文件结束了，则发送一个空指针的frame，用来清空缓冲区
encode_frame = nullptr;
}
// 发送一个frame
if (avcodec_send_frame(audio_enc_ctx, encode_frame) < 0) {
FATAL("send frame exception.");
}
// 接受编码完的内容
while (true) {
auto packet_ret = avcodec_receive_packet(audio_enc_ctx, &pkt);
// 判断是否完全接受了packet
if (packet_ret == AVERROR(EAGAIN) || packet_ret == AVERROR_EOF) {
break;
}
// 检查是否接受异常
if (packet_ret < 0) {
FATAL("receive packet exception.");
}
av_packet_rescale_ts(&pkt, audio_enc_ctx->time_base, audio_stream->time_base);
pkt.stream_index = audio_stream->index;
av_interleaved_write_frame(fmt_ctx, &pkt);
av_packet_unref(&pkt);
}
// 编码帧为空，则表示已经处理完所有的编码，退出该循环
if (encode_frame == nullptr) break;
}
av_write_trailer(fmt_ctx);
return 0;
}
```
1724阅读
1评论
0点赞
1