FFmpeg音频降噪功能的使用
网上关于FFmpeg的使用实在太少，这里自己摸索了一下
此处对pcm_s16le，单声道，256kbit/s的音频进行高低通降噪
除下高低通，还有lv2，afftdn等，可以参考官网
为了通用性，此处直接读取相关数据到unsigned char*数组中，并直接构造AVCodecContext。
创建音频滤镜图
在FFmpeg里，降噪也是一种滤镜，要想使用该滤镜，需要构造一个滤镜图。
图的起点和终点分别为abuffer和abuffersink。在构造后使用avfilter_graph_parse_ptr传入filter_desr完成图的构造。filter_desr为滤镜效果字符串。
此处使用"highpass=f=200,lowpass=f=3000"作为filter_desr，为高低通滤波器，200-3000hz为人声区域
:为同级分隔符，可以对配置传入多个参数，,可以另起滤镜配置
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
codec = avcodec_find_decoder(AV_CODEC_ID_PCM_S16LE);
dec_ctx = avcodec_alloc_context3(codec);
dec_ctx->sample_rate = 16000;
dec_ctx->sample_fmt = AV_SAMPLE_FMT_S16;
dec_ctx->channel_layout = AV_CH_LAYOUT_MONO;
dec_ctx->channels = 1;
dec_ctx->bit_rate = 256000;
int ret = avcodec_open2(dec_ctx, codec, NULL);
if (ret < 0) {
LOGE("no decoder!!!");
return;
}
char args[512];
snprintf(args, sizeof(args),
"sample_rate=%d:sample_fmt=%s:channel_layout=0x%" PRIx64,
dec_ctx->sample_rate,
av_get_sample_fmt_name(dec_ctx->sample_fmt), dec_ctx->channel_layout);
filter_graph = avfilter_graph_alloc();
if (!filter_graph) {
LOGE("alloc failed");
return;
}
// 获取对应的filter
src = avfilter_get_by_name("abuffer");
sink = avfilter_get_by_name("abuffersink");
if (!src || !sink) {
LOGE("no highpass_filter and lowpass_filter!!!");
}
ret = avfilter_graph_create_filter(&src_context, src, "in", args, NULL,
filter_graph);
if (ret < 0) {
LOGE("fail to create highpass filter.");
}
static const enum AVSampleFormat out_sample_fmts[] = {AV_SAMPLE_FMT_S16};
static const int64_t out_channel_layouts[] = {AV_CH_LAYOUT_MONO, -1};
static const int out_sample_rates[] = {16000, -1};
ret = avfilter_graph_create_filter(&sink_context, sink, "out", NULL,
NULL, filter_graph);
if (ret < 0) {
LOGE("fail to create sink");
}
ret = av_opt_set_int_list(sink_context, "sample_fmts", out_sample_fmts, -1,
AV_OPT_SEARCH_CHILDREN);
ret = av_opt_set_int_list(sink_context, "channel_layouts", out_channel_layouts, -1,
AV_OPT_SEARCH_CHILDREN);
ret = av_opt_set_int_list(sink_context, "sample_rates", out_sample_rates, -1,
AV_OPT_SEARCH_CHILDREN);
inputs->name = av_strdup("in");
inputs->filter_ctx = src_context;
inputs->pad_idx = 0;
inputs->next = NULL;
outputs->name = av_strdup("out");
outputs->filter_ctx = sink_context;
outputs->pad_idx = 0;
outputs->next = NULL;
ret = avfilter_graph_parse_ptr(filter_graph, filter_desr, &outputs, &inputs, NULL);
if (ret < 0) {
LOGE("fail to parse graph.");
}
ret = avfilter_graph_config(filter_graph, NULL);
if (ret < 0) {
LOGE("fail to config graph.");
}
逻辑代码
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
unsigned char *DenoiseManager::denoise_audio(unsigned char *data, int length) {
AVPacket* packet = av_packet_alloc();
AVFrame *frame = av_frame_alloc();
AVFrame *filt_frame = av_frame_alloc();
// 设置frame
frame->channel_layout = AV_CH_LAYOUT_MONO;
frame->sample_rate = 16000;
frame->format = AV_SAMPLE_FMT_S16;
auto * denoise_buf = (unsigned char*) malloc(sizeof(char)*length);
int denoise_index = 0;
packet->data = data;
packet->size = length;
packet->flags = 0;
packet->duration = 0;
packet->pos = 0;
avcodec_flush_buffers(dec_ctx);
int ret = avcodec_send_packet(dec_ctx,packet);
auto data_size = av_get_bytes_per_sample(AV_SAMPLE_FMT_S16);
if (ret < 0){
return data;
}
while (ret >= 0){
ret = avcodec_receive_frame(dec_ctx,frame);
if (ret == AVERROR(EAGAIN)){
continue;
}
if (ret == AVERROR_EOF)
return data;
ret = av_buffersrc_add_frame_flags(src_context,frame,AV_BUFFERSRC_FLAG_NO_CHECK_FORMAT); // 不检查格式，加快速度
if (ret < 0){
return data;
}
while(true){
ret = av_buffersink_get_frame(sink_context,filt_frame);
if (ret == AVERROR(EAGAIN)){
if(denoise_index == length){
break;
}
continue;
}
if (ret == AVERROR_EOF){
break;
}
if (ret < 0){
return data;
}
for (int i = 0; i < filt_frame->nb_samples; i++)
for (int ch = 0; ch < filt_frame->channels; ch++){
memcpy(denoise_buf + denoise_index,filt_frame->data[ch] + data_size * i,data_size);
denoise_index += data_size;
}
av_frame_unref(filt_frame);
}
av_frame_free(&frame);
av_frame_free(&filt_frame);
// 重新alloc avframe帧
filt_frame = av_frame_alloc();
frame = av_frame_alloc();
}
if (frame){
av_frame_free(&frame);
}
if (filt_frame){
av_frame_free(&filt_frame);
}
av_packet_unref(packet);
av_packet_free(&packet);
return denoise_buf;
}
