我有一个FFmpeg程序：
filtergraph本身相当复杂，可以直接从命令行运行：
ffmpeg -i demo.mp4 -filter_complex \
"[audio3]atrim=end=30:start=10[audio2];\
[video5]trim=end=30:start=10[video4];[audio2]anull[audio6];\
[video4]scale=width=1920:height=1080[video7];[audio6]anull[audio8];\
[video7]fps=fps=30[video9];[audio8]anull[audio10];\
[video9]format=pix_fmts=yuv420p[video11];\
[audio10]asetpts=expr=PTS-STARTPTS[audio12];\
[video11]setpts=expr=PTS-STARTPTS[video13];\
[audio15]concat=v=0:a=1:n=1[audio14];\
[video17]concat=v=1:a=0:n=1[video16];\
[audio12]afifo[audio15];[video13]fifo[video17];\
[audio14]afifo[audio18];[video16]fifo[video19];\
[audio18]anull[audio20];\
[video19]pad=width=1920:height=1080[video21];\
[audio20]anull[audio22];[video21]fps=fps=25[video23];\
[audio22]aformat=sample_fmts=fltp:sample_rates=44100:channel_layouts=stereo[fa];\
[video23]format=pix_fmts=yuv420p[fv];[0:a]afifo[audio3];\
[0:v]fifo[video5]" \
-map "[fv]" -map "[fa]" out.mp4
我意识到这是一个带有大量无操作过滤器的大规模过滤器图，它是自动生成的而不是手写的。 Here is a more cleaner version of the graph.（它是一个graphviz文件，您可以在命令行或here中运行它。）
无论如何，当我运行使用此过滤器图的程序时，我的内存使用量激增。我最终使用大约7 GB的RAM进行30秒的剪辑。但是，当我使用上面的ffmpeg命令运行程序时，它在大约600 MB的RAM中达到峰值。这使我相信问题不是过滤器图形的不合适的大小，而是我的程序如何使用它的问题。
程序设置filtergraph（使用av_filter_parse_ptr，给出上面显示的filtergraph字符串），编码器，复用器，解码器和解复用器，然后生成两个线程，一个将帧发送到filtergraph，另一个收到他们。发送它们的框架看起来像：
void decode () {
while(... more_frames ...) {
AVFrame *frame = av_frame_alloc();
... fill next frame of stream ...
av_buffersrc_write_frame(ctx, frame);
av_frame_free(&frame);
}
}
（我已经省略了av_send_packet/av_receive_frame函数，因为它们似乎没有泄漏内存。我也省略了刷新buffersrc的过程，因为直到最后才会发生这种情况，并且内存峰值很长在那之前。）
编码器线程看起来很相似：
void encode() {
while(... nodes_in_graph ...) {
AVFrame *frame = av_frame_alloc();
av_buffersink_get_frame(ctx, frame);
... ensure frame actually was filled ...
... send frame to encoder ...
av_frame_free(&frame);
}
}
与解码器一样，我已经省略了send_frame/receive_packet组合，因为它们似乎没有泄漏内存。此外，我已经省略了确保框架实际填充的细节。代码循环直到帧最终被填满。
我分配的每一帧我都很快解除分配。我另外处理了ffmpeg可以给出的所有错误情况（在示例中已经过了）。
我还尝试过只有一个帧用于编码器，一个用于解码器（并在循环的每次迭代中调用av_frame_unref）。
我忘了释放某些内容，或者我只是错误地使用了对libavfilter的调用，以至于它必须缓冲所有数据？我不认为泄漏是由内存图引起的，因为从命令行运行它似乎不会导致相同的内存爆炸。
FWIW，实际代码是here，尽管它是用Racket编写的。我有一个最小的例子似乎也复制了这种行为（从ffmpeg代码的doc/example/filtering_video.c文件修改：
#include <unistd.h>
#include <libavcodec/avcodec.h>
#include <libavformat/avformat.h>
#include <libavfilter/avfiltergraph.h>
#include <libavfilter/buffersink.h>
#include <libavfilter/buffersrc.h>
#include <libavutil/opt.h>
const char *filter_descr = "trim=start=10:end=30,scale=78:24,transpose=cclock";
static AVFormatContext *fmt_ctx;
static AVCodecContext *dec_ctx;
AVFilterContext *buffersink_ctx;
AVFilterContext *buffersrc_ctx;
AVFilterGraph *filter_graph;
static int video_stream_index = -1;
static int64_t last_pts = AV_NOPTS_VALUE;
static int open_input_file(const char *filename)
{
int ret;
AVCodec *dec;
if ((ret = avformat_open_input(&fmt_ctx, filename, NULL, NULL)) < 0) {
av_log(NULL, AV_LOG_ERROR, "Cannot open input file\n");
return ret;
}
if ((ret = avformat_find_stream_info(fmt_ctx, NULL)) < 0) {
av_log(NULL, AV_LOG_ERROR, "Cannot find stream information\n");
return ret;
}
/* select the video stream */
ret = av_find_best_stream(fmt_ctx, AVMEDIA_TYPE_VIDEO, -1, -1, &dec, 0);
if (ret < 0) {
av_log(NULL, AV_LOG_ERROR, "Cannot find a video stream in the input file\n");
return ret;
}
video_stream_index = ret;
/* create decoding context */
dec_ctx = avcodec_alloc_context3(dec);
if (!dec_ctx)
return AVERROR(ENOMEM);
avcodec_parameters_to_context(dec_ctx, fmt_ctx->streams[video_stream_index]->codecpar);
av_opt_set_int(dec_ctx, "refcounted_frames", 1, 0);
/* init the video decoder */
if ((ret = avcodec_open2(dec_ctx, dec, NULL)) < 0) {
av_log(NULL, AV_LOG_ERROR, "Cannot open video decoder\n");
return ret;
}
return 0;
}
static int init_filters(const char *filters_descr)
{
char args[512];
int ret = 0;
AVFilter *buffersrc = avfilter_get_by_name("buffer");
AVFilter *buffersink = avfilter_get_by_name("buffersink");
AVFilterInOut *outputs = avfilter_inout_alloc();
AVFilterInOut *inputs = avfilter_inout_alloc();
AVRational time_base = fmt_ctx->streams[video_stream_index]->time_base;
enum AVPixelFormat pix_fmts[] = { AV_PIX_FMT_GRAY8, AV_PIX_FMT_NONE };
filter_graph = avfilter_graph_alloc();
if (!outputs || !inputs || !filter_graph) {
ret = AVERROR(ENOMEM);
goto end;
}
/* buffer video source: the decoded frames from the decoder will be inserted here. */
snprintf(args, sizeof(args),
"video_size=%dx%d:pix_fmt=%d:time_base=%d/%d:pixel_aspect=%d/%d",
dec_ctx->width, dec_ctx->height, dec_ctx->pix_fmt,
time_base.num, time_base.den,
dec_ctx->sample_aspect_ratio.num, dec_ctx->sample_aspect_ratio.den);
ret = avfilter_graph_create_filter(&buffersrc_ctx, buffersrc, "in",
args, NULL, filter_graph);
if (ret < 0) {
av_log(NULL, AV_LOG_ERROR, "Cannot create buffer source\n");
goto end;
}
/* buffer video sink: to terminate the filter chain. */
ret = avfilter_graph_create_filter(&buffersink_ctx, buffersink, "out",
NULL, NULL, filter_graph);
if (ret < 0) {
av_log(NULL, AV_LOG_ERROR, "Cannot create buffer sink\n");
goto end;
}
ret = av_opt_set_int_list(buffersink_ctx, "pix_fmts", pix_fmts,
AV_PIX_FMT_NONE, AV_OPT_SEARCH_CHILDREN);
if (ret < 0) {
av_log(NULL, AV_LOG_ERROR, "Cannot set output pixel format\n");
goto end;
}
outputs->name = av_strdup("in");
outputs->filter_ctx = buffersrc_ctx;
outputs->pad_idx = 0;
outputs->next = NULL;
inputs->name = av_strdup("out");
inputs->filter_ctx = buffersink_ctx;
inputs->pad_idx = 0;
inputs->next = NULL;
if ((ret = avfilter_graph_parse_ptr(filter_graph, filters_descr,
&inputs, &outputs, NULL)) < 0)
goto end;
if ((ret = avfilter_graph_config(filter_graph, NULL)) < 0)
goto end;
end:
avfilter_inout_free(&inputs);
avfilter_inout_free(&outputs);
return ret;
}
int main(int argc, char **argv)
{
int ret;
AVPacket packet;
AVFrame *frame = av_frame_alloc();
AVFrame *filt_frame = av_frame_alloc();
if (!frame || !filt_frame) {
perror("Could not allocate frame");
exit(1);
}
if (argc != 2) {
fprintf(stderr, "Usage: %s file\n", argv[0]);
exit(1);
}
av_register_all();
avfilter_register_all();
if ((ret = open_input_file(argv[1])) < 0)
goto end;
if ((ret = init_filters(filter_descr)) < 0)
goto end;
/* read all packets */
while (1) {
if ((ret = av_read_frame(fmt_ctx, &packet)) < 0)
break;
if (packet.stream_index == video_stream_index) {
ret = avcodec_send_packet(dec_ctx, &packet);
if (ret < 0) {
av_log(NULL, AV_LOG_ERROR, "Error while sending a packet to the decoder\n");
break;
}
while (ret >= 0) {
ret = avcodec_receive_frame(dec_ctx, frame);
if (ret == AVERROR(EAGAIN) || ret == AVERROR_EOF) {
break;
} else if (ret < 0) {
av_log(NULL, AV_LOG_ERROR, "Error while receiving a frame from the decoder\n");
goto end;
}
if (ret >= 0) {
frame->pts = av_frame_get_best_effort_timestamp(frame);
/* push the decoded frame into the filtergraph */
if (av_buffersrc_add_frame_flags(buffersrc_ctx, frame, AV_BUFFERSRC_FLAG_KEEP_REF) < 0) {
av_log(NULL, AV_LOG_ERROR, "Error while feeding the filtergraph\n");
break;
}
/* pull filtered frames from the filtergraph */
while (1) {
ret = av_buffersink_get_frame(buffersink_ctx, filt_frame);
if (ret == AVERROR(EAGAIN) || ret == AVERROR_EOF)
break;
if (ret < 0)
goto end;
av_frame_unref(filt_frame);
}
av_frame_unref(frame);
}
}
}
av_packet_unref(&packet);
}
end:
avfilter_graph_free(&filter_graph);
avcodec_free_context(&dec_ctx);
avformat_close_input(&fmt_ctx);
av_frame_free(&frame);
av_frame_free(&filt_frame);
return ret;
}