CODEWITHSUNDEEP
ffmpeg
Jerikc XIONG
Jerikc XIONG

Reputation: 3607

How to understand the given ffplay C code snippet?

The following code snippet is from ffplay:

static int decoder_decode_frame(Decoder *d, AVFrame *frame, AVSubtitle *sub) {
    int got_frame = 0;

    do {
        int ret = -1;

        if (d->queue->abort_request)
            return -1;

        if (!d->packet_pending || d->queue->serial != d->pkt_serial) {
            AVPacket pkt;
            do {
                if (d->queue->nb_packets == 0)
                    SDL_CondSignal(d->empty_queue_cond);
                if (packet_queue_get(d->queue, &pkt, 1, &d->pkt_serial) < 0)
                    return -1;
                if (pkt.data == flush_pkt.data) {
                    avcodec_flush_buffers(d->avctx);
                    d->finished = 0;
                    d->next_pts = d->start_pts;
                    d->next_pts_tb = d->start_pts_tb;
                }
            } while (pkt.data == flush_pkt.data || d->queue->serial != d->pkt_serial);
            av_free_packet(&d->pkt);
            d->pkt_temp = d->pkt = pkt;
            d->packet_pending = 1;
        }

        switch (d->avctx->codec_type) {
            case AVMEDIA_TYPE_VIDEO:
                ret = avcodec_decode_video2(d->avctx, frame, &got_frame, &d->pkt_temp);
                if (got_frame) {
                    if (decoder_reorder_pts == -1) {
                        frame->pts = av_frame_get_best_effort_timestamp(frame);
                    } else if (decoder_reorder_pts) {
                        frame->pts = frame->pkt_pts;
                    } else {
                        frame->pts = frame->pkt_dts;
                    }
                }
                break;
            case AVMEDIA_TYPE_AUDIO:
                ret = avcodec_decode_audio4(d->avctx, frame, &got_frame, &d->pkt_temp);
                if (got_frame) {
                    AVRational tb = (AVRational){1, frame->sample_rate};
                    if (frame->pts != AV_NOPTS_VALUE)
                        frame->pts = av_rescale_q(frame->pts, d->avctx->time_base, tb);
                    else if (frame->pkt_pts != AV_NOPTS_VALUE)
                        frame->pts = av_rescale_q(frame->pkt_pts, av_codec_get_pkt_timebase(d->avctx), tb);
                    else if (d->next_pts != AV_NOPTS_VALUE)
                        frame->pts = av_rescale_q(d->next_pts, d->next_pts_tb, tb);
                    if (frame->pts != AV_NOPTS_VALUE) {
                        d->next_pts = frame->pts + frame->nb_samples;
                        d->next_pts_tb = tb;
                    }
                }
                break;
            case AVMEDIA_TYPE_SUBTITLE:
                ret = avcodec_decode_subtitle2(d->avctx, sub, &got_frame, &d->pkt_temp);
                break;
        }

        if (ret < 0) {
            d->packet_pending = 0;
        } else {
            d->pkt_temp.dts =
            d->pkt_temp.pts = AV_NOPTS_VALUE;
            if (d->pkt_temp.data) {
                if (d->avctx->codec_type != AVMEDIA_TYPE_AUDIO)
                    ret = d->pkt_temp.size;
                d->pkt_temp.data += ret;
                d->pkt_temp.size -= ret;
                if (d->pkt_temp.size <= 0)
                    d->packet_pending = 0;
            } else {
                if (!got_frame) {
                    d->packet_pending = 0;
                    d->finished = d->pkt_serial; // FLAG
                }
            }
        }
    } while (!got_frame && !d->finished);

    return got_frame;
}

It's difficult for me to understand the following code:

d->finished = d->pkt_serial; // FLAG

Can anyone help me ?

Thanks.

Upvotes: 2

Views: 1090

Answers (1)

Ronald S. Bultje
Ronald S. Bultje

Reputation: 11174

See this commit.

Serial has two purposes nowadays: the initial purpose in the commit was to be able to distinguish packets in the packet queue from before and after a seek. The demuxer (input for packet queue) runs in a separate thread. After a seek, we want to flush it, but we don't want to stop the producer thread because overhead. However, we also don't want to flush too few or too many packets. So, the serial field tells us which packets are pre- and post-flush and thus which packets to drop without having to stop the producer thread while we're dropping those packets.

The second purpose is your line of code: it tells us when EOF occurs. Finished is set to the last serial number of a packet from the packet queue used to decode a frame. If that serial number is also the tail of the packet queue (and no more packets are produced), it means we stopped producing packets and decoded the frame belonging to that packet. In other words: end-of-file. Elsewhere, you'll find a test along those lines, and then either playback stops or (if looping is enabled) we seek back to the beginning of the file (i.e. invoke the looping behaviour).

(This write-up was helpfully assisted by several FFmpeg developers on IRC.)

Upvotes: 2

Related Questions