avconv.c

/*
 * avconv main
 * Copyright (c) 2000-2011 The libav developers.
 *
 * This file is part of Libav.
 *
 * Libav is free software; you can redistribute it and/or
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
 * version 2.1 of the License, or (at your option) any later version.
 *
 * Libav is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
 * License along with Libav; if not, write to the Free Software
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 */

#include "config.h"
#include <ctype.h>
#include <string.h>
#include <math.h>
#include <stdlib.h>
#include <errno.h>
#include <signal.h>
#include <limits.h>
#include <unistd.h>
#include "libavformat/avformat.h"
#include "libavdevice/avdevice.h"
#include "libswscale/swscale.h"
#include "libavutil/opt.h"
#include "libavcodec/audioconvert.h"
#include "libavutil/audioconvert.h"
#include "libavutil/parseutils.h"
#include "libavutil/samplefmt.h"
#include "libavutil/colorspace.h"
#include "libavutil/fifo.h"
#include "libavutil/intreadwrite.h"
#include "libavutil/dict.h"
#include "libavutil/mathematics.h"
#include "libavutil/pixdesc.h"
#include "libavutil/avstring.h"
#include "libavutil/libm.h"
#include "libavutil/imgutils.h"
#include "libavformat/os_support.h"

#if CONFIG_AVFILTER
# include "libavfilter/avfilter.h"
# include "libavfilter/avfiltergraph.h"
# include "libavfilter/buffersrc.h"
# include "libavfilter/vsrc_buffer.h"
#endif

#if HAVE_SYS_RESOURCE_H
#include <sys/types.h>
#include <sys/time.h>
#include <sys/resource.h>
#elif HAVE_GETPROCESSTIMES
#include <windows.h>
#endif
#if HAVE_GETPROCESSMEMORYINFO
#include <windows.h>
#include <psapi.h>
#endif

#if HAVE_SYS_SELECT_H
#include <sys/select.h>
#endif

#include <time.h>

#include "cmdutils.h"

#include "libavutil/avassert.h"

#define VSYNC_AUTO       -1
#define VSYNC_PASSTHROUGH 0
#define VSYNC_CFR         1
#define VSYNC_VFR         2

const char program_name[] = "avconv";
const int program_birth_year = 2000;

/* select an input stream for an output stream */
typedef struct StreamMap {
    int disabled;           /** 1 is this mapping is disabled by a negative map */
    int file_index;
    int stream_index;
    int sync_file_index;
    int sync_stream_index;
} StreamMap;

/**
 * select an input file for an output file
 */
typedef struct MetadataMap {
    int  file;      ///< file index
    char type;      ///< type of metadata to copy -- (g)lobal, (s)tream, (c)hapter or (p)rogram
    int  index;     ///< stream/chapter/program number
} MetadataMap;

static const OptionDef options[];

static int video_discard = 0;
static int same_quant = 0;
static int do_deinterlace = 0;
static int intra_dc_precision = 8;
static int qp_hist = 0;

static int file_overwrite = 0;
static int do_benchmark = 0;
static int do_hex_dump = 0;
static int do_pkt_dump = 0;
static int do_pass = 0;
static char *pass_logfilename_prefix = NULL;
static int video_sync_method = VSYNC_AUTO;
static int audio_sync_method = 0;
static float audio_drift_threshold = 0.1;
static int copy_ts = 0;
static int copy_tb = 1;
static int opt_shortest = 0;
static char *vstats_filename;
static FILE *vstats_file;

static int audio_volume = 256;

static int exit_on_error = 0;
static int using_stdin = 0;
static int64_t video_size = 0;
static int64_t audio_size = 0;
static int64_t extra_size = 0;
static int nb_frames_dup = 0;
static int nb_frames_drop = 0;
static int input_sync;

static float dts_delta_threshold = 10;

static int print_stats = 1;

static uint8_t *audio_buf;
static unsigned int allocated_audio_buf_size;

#define DEFAULT_PASS_LOGFILENAME_PREFIX "av2pass"

typedef struct FrameBuffer {
    uint8_t *base[4];
    uint8_t *data[4];
    int  linesize[4];

    int h, w;
    enum PixelFormat pix_fmt;

    int refcount;
    struct InputStream *ist;
    struct FrameBuffer *next;
} FrameBuffer;

typedef struct InputStream {
    int file_index;
    AVStream *st;
    int discard;             /* true if stream data should be discarded */
    int decoding_needed;     /* true if the packets must be decoded in 'raw_fifo' */
    AVCodec *dec;
    AVFrame *decoded_frame;
    AVFrame *filtered_frame;

    int64_t       start;     /* time when read started */
    /* predicted dts of the next packet read for this stream or (when there are
     * several frames in a packet) of the next frame in current packet */
    int64_t       next_dts;
    /* dts of the last packet read for this stream */
    int64_t       last_dts;
    PtsCorrectionContext pts_ctx;
    double ts_scale;
    int is_start;            /* is 1 at the start and after a discontinuity */
    int showed_multi_packet_warning;
    AVDictionary *opts;

    /* a pool of free buffers for decoded data */
    FrameBuffer *buffer_pool;
} InputStream;

typedef struct InputFile {
    AVFormatContext *ctx;
    int eof_reached;      /* true if eof reached */
    int ist_index;        /* index of first stream in ist_table */
    int buffer_size;      /* current total buffer size */
    int64_t ts_offset;
    int nb_streams;       /* number of stream that avconv is aware of; may be different
                             from ctx.nb_streams if new streams appear during av_read_frame() */
    int rate_emu;
} InputFile;

typedef struct OutputStream {
    int file_index;          /* file index */
    int index;               /* stream index in the output file */
    int source_index;        /* InputStream index */
    AVStream *st;            /* stream in the output file */
    int encoding_needed;     /* true if encoding needed for this stream */
    int frame_number;
    /* input pts and corresponding output pts
       for A/V sync */
    // double sync_ipts;        /* dts from the AVPacket of the demuxer in second units */
    struct InputStream *sync_ist; /* input stream to sync against */
    int64_t sync_opts;       /* output frame counter, could be changed to some true timestamp */ // FIXME look at frame_number
    /* pts of the first frame encoded for this stream, used for limiting
     * recording time */
    int64_t first_pts;
    AVBitStreamFilterContext *bitstream_filters;
    AVCodec *enc;
    int64_t max_frames;
    AVFrame *output_frame;

    /* video only */
    int video_resample;
    AVFrame pict_tmp;      /* temporary image for resampling */
    struct SwsContext *img_resample_ctx; /* for image resampling */
    int resample_height;
    int resample_width;
    int resample_pix_fmt;
    AVRational frame_rate;
    int force_fps;
    int top_field_first;

    float frame_aspect_ratio;

    /* forced key frames */
    int64_t *forced_kf_pts;
    int forced_kf_count;
    int forced_kf_index;

    /* audio only */
    int audio_resample;
    ReSampleContext *resample; /* for audio resampling */
    int resample_sample_fmt;
    int resample_channels;
    int resample_sample_rate;
    int reformat_pair;
    AVAudioConvert *reformat_ctx;
    AVFifoBuffer *fifo;     /* for compression: one audio fifo per codec */
    FILE *logfile;

#if CONFIG_AVFILTER
    AVFilterContext *output_video_filter;
    AVFilterContext *input_video_filter;
    AVFilterBufferRef *picref;
    char *avfilter;
    AVFilterGraph *graph;
#endif

    int64_t sws_flags;
    AVDictionary *opts;
    int is_past_recording_time;
    int stream_copy;
    const char *attachment_filename;
    int copy_initial_nonkeyframes;
} OutputStream;


typedef struct OutputFile {
    AVFormatContext *ctx;
    AVDictionary *opts;
    int ost_index;       /* index of the first stream in output_streams */
    int64_t recording_time; /* desired length of the resulting file in microseconds */
    int64_t start_time;     /* start time in microseconds */
    uint64_t limit_filesize;
} OutputFile;

static InputStream *input_streams   = NULL;
static int         nb_input_streams = 0;
static InputFile   *input_files     = NULL;
static int         nb_input_files   = 0;

static OutputStream *output_streams = NULL;
static int        nb_output_streams = 0;
static OutputFile   *output_files   = NULL;
static int        nb_output_files   = 0;

typedef struct OptionsContext {
    /* input/output options */
    int64_t start_time;
    const char *format;

    SpecifierOpt *codec_names;
    int        nb_codec_names;
    SpecifierOpt *audio_channels;
    int        nb_audio_channels;
    SpecifierOpt *audio_sample_rate;
    int        nb_audio_sample_rate;
    SpecifierOpt *frame_rates;
    int        nb_frame_rates;
    SpecifierOpt *frame_sizes;
    int        nb_frame_sizes;
    SpecifierOpt *frame_pix_fmts;
    int        nb_frame_pix_fmts;

    /* input options */
    int64_t input_ts_offset;
    int rate_emu;

    SpecifierOpt *ts_scale;
    int        nb_ts_scale;
    SpecifierOpt *dump_attachment;
    int        nb_dump_attachment;

    /* output options */
    StreamMap *stream_maps;
    int     nb_stream_maps;
    /* first item specifies output metadata, second is input */
    MetadataMap (*meta_data_maps)[2];
    int nb_meta_data_maps;
    int metadata_global_manual;
    int metadata_streams_manual;
    int metadata_chapters_manual;
    const char **attachments;
    int       nb_attachments;

    int chapters_input_file;

    int64_t recording_time;
    uint64_t limit_filesize;
    float mux_preload;
    float mux_max_delay;

    int video_disable;
    int audio_disable;
    int subtitle_disable;
    int data_disable;

    /* indexed by output file stream index */
    int   *streamid_map;
    int nb_streamid_map;

    SpecifierOpt *metadata;
    int        nb_metadata;
    SpecifierOpt *max_frames;
    int        nb_max_frames;
    SpecifierOpt *bitstream_filters;
    int        nb_bitstream_filters;
    SpecifierOpt *codec_tags;
    int        nb_codec_tags;
    SpecifierOpt *sample_fmts;
    int        nb_sample_fmts;
    SpecifierOpt *qscale;
    int        nb_qscale;
    SpecifierOpt *forced_key_frames;
    int        nb_forced_key_frames;
    SpecifierOpt *force_fps;
    int        nb_force_fps;
    SpecifierOpt *frame_aspect_ratios;
    int        nb_frame_aspect_ratios;
    SpecifierOpt *rc_overrides;
    int        nb_rc_overrides;
    SpecifierOpt *intra_matrices;
    int        nb_intra_matrices;
    SpecifierOpt *inter_matrices;
    int        nb_inter_matrices;
    SpecifierOpt *top_field_first;
    int        nb_top_field_first;
    SpecifierOpt *metadata_map;
    int        nb_metadata_map;
    SpecifierOpt *presets;
    int        nb_presets;
    SpecifierOpt *copy_initial_nonkeyframes;
    int        nb_copy_initial_nonkeyframes;
#if CONFIG_AVFILTER
    SpecifierOpt *filters;
    int        nb_filters;
#endif
} OptionsContext;

#define MATCH_PER_STREAM_OPT(name, type, outvar, fmtctx, st)\
{\
    int i, ret;\
    for (i = 0; i < o->nb_ ## name; i++) {\
        char *spec = o->name[i].specifier;\
        if ((ret = check_stream_specifier(fmtctx, st, spec)) > 0)\
            outvar = o->name[i].u.type;\
        else if (ret < 0)\
            exit_program(1);\
    }\
}

static void reset_options(OptionsContext *o)
{
    const OptionDef *po = options;

    /* all OPT_SPEC and OPT_STRING can be freed in generic way */
    while (po->name) {
        void *dst = (uint8_t*)o + po->u.off;

        if (po->flags & OPT_SPEC) {
            SpecifierOpt **so = dst;
            int i, *count = (int*)(so + 1);
            for (i = 0; i < *count; i++) {
                av_freep(&(*so)[i].specifier);
                if (po->flags & OPT_STRING)
                    av_freep(&(*so)[i].u.str);
            }
            av_freep(so);
            *count = 0;
        } else if (po->flags & OPT_OFFSET && po->flags & OPT_STRING)
            av_freep(dst);
        po++;
    }

    av_freep(&o->stream_maps);
    av_freep(&o->meta_data_maps);
    av_freep(&o->streamid_map);

    memset(o, 0, sizeof(*o));

    o->mux_max_delay  = 0.7;
    o->recording_time = INT64_MAX;
    o->limit_filesize = UINT64_MAX;
    o->chapters_input_file = INT_MAX;

    uninit_opts();
    init_opts();
}

static int alloc_buffer(InputStream *ist, FrameBuffer **pbuf)
{
    AVCodecContext *s = ist->st->codec;
    FrameBuffer  *buf = av_mallocz(sizeof(*buf));
    int i, ret;
    const int pixel_size = av_pix_fmt_descriptors[s->pix_fmt].comp[0].step_minus1+1;
    int h_chroma_shift, v_chroma_shift;
    int edge = 32; // XXX should be avcodec_get_edge_width(), but that fails on svq1
    int w = s->width, h = s->height;

    if (!buf)
        return AVERROR(ENOMEM);

    if (!(s->flags & CODEC_FLAG_EMU_EDGE)) {
        w += 2*edge;
        h += 2*edge;
    }

    avcodec_align_dimensions(s, &w, &h);
    if ((ret = av_image_alloc(buf->base, buf->linesize, w, h,
                              s->pix_fmt, 32)) < 0) {
        av_freep(&buf);
        return ret;
    }
    /* XXX this shouldn't be needed, but some tests break without this line
     * those decoders are buggy and need to be fixed.
     * the following tests fail:
     * bethsoft-vid, cdgraphics, ansi, aasc, fraps-v1, qtrle-1bit
     */
    memset(buf->base[0], 128, ret);

    avcodec_get_chroma_sub_sample(s->pix_fmt, &h_chroma_shift, &v_chroma_shift);
    for (i = 0; i < FF_ARRAY_ELEMS(buf->data); i++) {
        const int h_shift = i==0 ? 0 : h_chroma_shift;
        const int v_shift = i==0 ? 0 : v_chroma_shift;
        if (s->flags & CODEC_FLAG_EMU_EDGE)
            buf->data[i] = buf->base[i];
        else
            buf->data[i] = buf->base[i] +
                           FFALIGN((buf->linesize[i]*edge >> v_shift) +
                                   (pixel_size*edge >> h_shift), 32);
    }
    buf->w       = s->width;
    buf->h       = s->height;
    buf->pix_fmt = s->pix_fmt;
    buf->ist     = ist;

    *pbuf = buf;
    return 0;
}

static void free_buffer_pool(InputStream *ist)
{
    FrameBuffer *buf = ist->buffer_pool;
    while (buf) {
        ist->buffer_pool = buf->next;
        av_freep(&buf->base[0]);
        av_free(buf);
        buf = ist->buffer_pool;
    }
}

static void unref_buffer(InputStream *ist, FrameBuffer *buf)
{
    av_assert0(buf->refcount);
    buf->refcount--;
    if (!buf->refcount) {
        buf->next = ist->buffer_pool;
        ist->buffer_pool = buf;
    }
}

static int codec_get_buffer(AVCodecContext *s, AVFrame *frame)
{
    InputStream *ist = s->opaque;
    FrameBuffer *buf;
    int ret, i;

    if (!ist->buffer_pool && (ret = alloc_buffer(ist, &ist->buffer_pool)) < 0)
        return ret;

    buf              = ist->buffer_pool;
    ist->buffer_pool = buf->next;
    buf->next        = NULL;
    if (buf->w != s->width || buf->h != s->height || buf->pix_fmt != s->pix_fmt) {
        av_freep(&buf->base[0]);
        av_free(buf);
        if ((ret = alloc_buffer(ist, &buf)) < 0)
            return ret;
    }
    buf->refcount++;

    frame->opaque        = buf;
    frame->type          = FF_BUFFER_TYPE_USER;
    frame->extended_data = frame->data;
    frame->pkt_pts       = s->pkt ? s->pkt->pts : AV_NOPTS_VALUE;

    for (i = 0; i < FF_ARRAY_ELEMS(buf->data); i++) {
        frame->base[i]     = buf->base[i];  // XXX h264.c uses base though it shouldn't
        frame->data[i]     = buf->data[i];
        frame->linesize[i] = buf->linesize[i];
    }

    return 0;
}

static void codec_release_buffer(AVCodecContext *s, AVFrame *frame)
{
    InputStream *ist = s->opaque;
    FrameBuffer *buf = frame->opaque;
    int i;

    for (i = 0; i < FF_ARRAY_ELEMS(frame->data); i++)
        frame->data[i] = NULL;

    unref_buffer(ist, buf);
}

static void filter_release_buffer(AVFilterBuffer *fb)
{
    FrameBuffer *buf = fb->priv;
    av_free(fb);
    unref_buffer(buf->ist, buf);
}

#if CONFIG_AVFILTER

static int configure_video_filters(InputStream *ist, OutputStream *ost)
{
    AVFilterContext *last_filter, *filter;
    /** filter graph containing all filters including input & output */
    AVCodecContext *codec = ost->st->codec;
    AVCodecContext *icodec = ist->st->codec;
    SinkContext sink_ctx = { .pix_fmt = codec->pix_fmt };
    AVRational sample_aspect_ratio;
    char args[255];
    int ret;

    ost->graph = avfilter_graph_alloc();

    if (ist->st->sample_aspect_ratio.num) {
        sample_aspect_ratio = ist->st->sample_aspect_ratio;
    } else
        sample_aspect_ratio = ist->st->codec->sample_aspect_ratio;

    snprintf(args, 255, "%d:%d:%d:%d:%d:%d:%d", ist->st->codec->width,
             ist->st->codec->height, ist->st->codec->pix_fmt, 1, AV_TIME_BASE,
             sample_aspect_ratio.num, sample_aspect_ratio.den);

    ret = avfilter_graph_create_filter(&ost->input_video_filter, avfilter_get_by_name("buffer"),
                                       "src", args, NULL, ost->graph);
    if (ret < 0)
        return ret;
    ret = avfilter_graph_create_filter(&ost->output_video_filter, &sink,
                                       "out", NULL, &sink_ctx, ost->graph);
    if (ret < 0)
        return ret;
    last_filter = ost->input_video_filter;

    if (codec->width != icodec->width || codec->height != icodec->height) {
        snprintf(args, 255, "%d:%d:flags=0x%X",
                 codec->width,
                 codec->height,
                 (unsigned)ost->sws_flags);
        if ((ret = avfilter_graph_create_filter(&filter, avfilter_get_by_name("scale"),
                                                NULL, args, NULL, ost->graph)) < 0)
            return ret;
        if ((ret = avfilter_link(last_filter, 0, filter, 0)) < 0)
            return ret;
        last_filter = filter;
    }

    snprintf(args, sizeof(args), "flags=0x%X", (unsigned)ost->sws_flags);
    ost->graph->scale_sws_opts = av_strdup(args);

    if (ost->avfilter) {
        AVFilterInOut *outputs = av_malloc(sizeof(AVFilterInOut));
        AVFilterInOut *inputs  = av_malloc(sizeof(AVFilterInOut));

        outputs->name    = av_strdup("in");
        outputs->filter_ctx = last_filter;
        outputs->pad_idx = 0;
        outputs->next    = NULL;

        inputs->name    = av_strdup("out");
        inputs->filter_ctx = ost->output_video_filter;
        inputs->pad_idx = 0;
        inputs->next    = NULL;

        if ((ret = avfilter_graph_parse(ost->graph, ost->avfilter, inputs, outputs, NULL)) < 0)
            return ret;
    } else {
        if ((ret = avfilter_link(last_filter, 0, ost->output_video_filter, 0)) < 0)
            return ret;
    }

    if ((ret = avfilter_graph_config(ost->graph, NULL)) < 0)
        return ret;

    codec->width  = ost->output_video_filter->inputs[0]->w;
    codec->height = ost->output_video_filter->inputs[0]->h;
    codec->sample_aspect_ratio = ost->st->sample_aspect_ratio =
        ost->frame_aspect_ratio ? // overridden by the -aspect cli option
        av_d2q(ost->frame_aspect_ratio * codec->height/codec->width, 255) :
        ost->output_video_filter->inputs[0]->sample_aspect_ratio;

    return 0;
}
#endif /* CONFIG_AVFILTER */

static void term_exit(void)
{
    av_log(NULL, AV_LOG_QUIET, "");
}

static volatile int received_sigterm = 0;
static volatile int received_nb_signals = 0;

static void
sigterm_handler(int sig)
{
    received_sigterm = sig;
    received_nb_signals++;
    term_exit();
}

static void term_init(void)
{
    signal(SIGINT , sigterm_handler); /* Interrupt (ANSI).    */
    signal(SIGTERM, sigterm_handler); /* Termination (ANSI).  */
#ifdef SIGXCPU
    signal(SIGXCPU, sigterm_handler);
#endif
}

static int decode_interrupt_cb(void *ctx)
{
    return received_nb_signals > 1;
}

static const AVIOInterruptCB int_cb = { decode_interrupt_cb, NULL };

void exit_program(int ret)
{
    int i;

    /* close files */
    for (i = 0; i < nb_output_files; i++) {
        AVFormatContext *s = output_files[i].ctx;
        if (!(s->oformat->flags & AVFMT_NOFILE) && s->pb)
            avio_close(s->pb);
        avformat_free_context(s);
        av_dict_free(&output_files[i].opts);
    }
    for (i = 0; i < nb_output_streams; i++) {
        AVBitStreamFilterContext *bsfc = output_streams[i].bitstream_filters;
        while (bsfc) {
            AVBitStreamFilterContext *next = bsfc->next;
            av_bitstream_filter_close(bsfc);
            bsfc = next;
        }
        output_streams[i].bitstream_filters = NULL;

        if (output_streams[i].output_frame) {
            AVFrame *frame = output_streams[i].output_frame;
            if (frame->extended_data != frame->data)
                av_freep(&frame->extended_data);
            av_freep(&frame);
        }

#if CONFIG_AVFILTER
        av_freep(&output_streams[i].avfilter);
#endif
    }
    for (i = 0; i < nb_input_files; i++) {
        avformat_close_input(&input_files[i].ctx);
    }
    for (i = 0; i < nb_input_streams; i++) {
        av_freep(&input_streams[i].decoded_frame);
        av_freep(&input_streams[i].filtered_frame);
        av_dict_free(&input_streams[i].opts);
        free_buffer_pool(&input_streams[i]);
    }

    if (vstats_file)
        fclose(vstats_file);
    av_free(vstats_filename);

    av_freep(&input_streams);
    av_freep(&input_files);
    av_freep(&output_streams);
    av_freep(&output_files);

    uninit_opts();
    av_free(audio_buf);
    allocated_audio_buf_size = 0;

#if CONFIG_AVFILTER
    avfilter_uninit();
#endif
    avformat_network_deinit();

    if (received_sigterm) {
        av_log(NULL, AV_LOG_INFO, "Received signal %d: terminating.\n",
               (int) received_sigterm);
        exit (255);
    }

    exit(ret);
}

static void assert_avoptions(AVDictionary *m)
{
    AVDictionaryEntry *t;
    if ((t = av_dict_get(m, "", NULL, AV_DICT_IGNORE_SUFFIX))) {
        av_log(NULL, AV_LOG_FATAL, "Option %s not found.\n", t->key);
        exit_program(1);
    }
}

static void assert_codec_experimental(AVCodecContext *c, int encoder)
{
    const char *codec_string = encoder ? "encoder" : "decoder";
    AVCodec *codec;
    if (c->codec->capabilities & CODEC_CAP_EXPERIMENTAL &&
        c->strict_std_compliance > FF_COMPLIANCE_EXPERIMENTAL) {
        av_log(NULL, AV_LOG_FATAL, "%s '%s' is experimental and might produce bad "
                "results.\nAdd '-strict experimental' if you want to use it.\n",
                codec_string, c->codec->name);
        codec = encoder ? avcodec_find_encoder(c->codec->id) : avcodec_find_decoder(c->codec->id);
        if (!(codec->capabilities & CODEC_CAP_EXPERIMENTAL))
            av_log(NULL, AV_LOG_FATAL, "Or use the non experimental %s '%s'.\n",
                   codec_string, codec->name);
        exit_program(1);
    }
}

static void choose_sample_fmt(AVStream *st, AVCodec *codec)
{
    if (codec && codec->sample_fmts) {
        const enum AVSampleFormat *p = codec->sample_fmts;
        for (; *p != -1; p++) {
            if (*p == st->codec->sample_fmt)
                break;
        }
        if (*p == -1) {
            av_log(NULL, AV_LOG_WARNING,
                   "Incompatible sample format '%s' for codec '%s', auto-selecting format '%s'\n",
                   av_get_sample_fmt_name(st->codec->sample_fmt),
                   codec->name,
                   av_get_sample_fmt_name(codec->sample_fmts[0]));
            st->codec->sample_fmt = codec->sample_fmts[0];
        }
    }
}

/**
 * Update the requested input sample format based on the output sample format.
 * This is currently only used to request float output from decoders which
 * support multiple sample formats, one of which is AV_SAMPLE_FMT_FLT.
 * Ideally this will be removed in the future when decoders do not do format
 * conversion and only output in their native format.
 */
static void update_sample_fmt(AVCodecContext *dec, AVCodec *dec_codec,
                              AVCodecContext *enc)
{
    /* if sample formats match or a decoder sample format has already been
       requested, just return */
    if (enc->sample_fmt == dec->sample_fmt ||
        dec->request_sample_fmt > AV_SAMPLE_FMT_NONE)
        return;

    /* if decoder supports more than one output format */
    if (dec_codec && dec_codec->sample_fmts &&
        dec_codec->sample_fmts[0] != AV_SAMPLE_FMT_NONE &&
        dec_codec->sample_fmts[1] != AV_SAMPLE_FMT_NONE) {
        const enum AVSampleFormat *p;
        int min_dec = -1, min_inc = -1;

        /* find a matching sample format in the encoder */
        for (p = dec_codec->sample_fmts; *p != AV_SAMPLE_FMT_NONE; p++) {
            if (*p == enc->sample_fmt) {
                dec->request_sample_fmt = *p;
                return;
            } else if (*p > enc->sample_fmt) {
                min_inc = FFMIN(min_inc, *p - enc->sample_fmt);
            } else
                min_dec = FFMIN(min_dec, enc->sample_fmt - *p);
        }

        /* if none match, provide the one that matches quality closest */
        dec->request_sample_fmt = min_inc > 0 ? enc->sample_fmt + min_inc :
                                  enc->sample_fmt - min_dec;
    }
}

static void choose_sample_rate(AVStream *st, AVCodec *codec)
{
    if (codec && codec->supported_samplerates) {
        const int *p  = codec->supported_samplerates;
        int best      = 0;
        int best_dist = INT_MAX;
        for (; *p; p++) {
            int dist = abs(st->codec->sample_rate - *p);
            if (dist < best_dist) {
                best_dist = dist;
                best      = *p;
            }
        }
        if (best_dist) {
            av_log(st->codec, AV_LOG_WARNING, "Requested sampling rate unsupported using closest supported (%d)\n", best);
        }
        st->codec->sample_rate = best;
    }
}

static void choose_pixel_fmt(AVStream *st, AVCodec *codec)
{
    if (codec && codec->pix_fmts) {
        const enum PixelFormat *p = codec->pix_fmts;
        if (st->codec->strict_std_compliance <= FF_COMPLIANCE_UNOFFICIAL) {
            if (st->codec->codec_id == CODEC_ID_MJPEG) {
                p = (const enum PixelFormat[]) { PIX_FMT_YUVJ420P, PIX_FMT_YUVJ422P, PIX_FMT_YUV420P, PIX_FMT_YUV422P, PIX_FMT_NONE };
            } else if (st->codec->codec_id == CODEC_ID_LJPEG) {
                p = (const enum PixelFormat[]) { PIX_FMT_YUVJ420P, PIX_FMT_YUVJ422P, PIX_FMT_YUVJ444P, PIX_FMT_YUV420P,
                                                 PIX_FMT_YUV422P, PIX_FMT_YUV444P, PIX_FMT_BGRA, PIX_FMT_NONE };
            }
        }
        for (; *p != PIX_FMT_NONE; p++) {
            if (*p == st->codec->pix_fmt)
                break;
        }
        if (*p == PIX_FMT_NONE) {
            if (st->codec->pix_fmt != PIX_FMT_NONE)
                av_log(NULL, AV_LOG_WARNING,
                       "Incompatible pixel format '%s' for codec '%s', auto-selecting format '%s'\n",
                       av_pix_fmt_descriptors[st->codec->pix_fmt].name,
                       codec->name,
                       av_pix_fmt_descriptors[codec->pix_fmts[0]].name);
            st->codec->pix_fmt = codec->pix_fmts[0];
        }
    }
}

static double
get_sync_ipts(const OutputStream *ost, int64_t pts)
{
    OutputFile *of = &output_files[ost->file_index];
    return (double)(pts - of->start_time) / AV_TIME_BASE;
}

static void write_frame(AVFormatContext *s, AVPacket *pkt, OutputStream *ost)
{
    AVBitStreamFilterContext *bsfc = ost->bitstream_filters;
    AVCodecContext          *avctx = ost->st->codec;
    int ret;

    /*
     * Audio encoders may split the packets --  #frames in != #packets out.
     * But there is no reordering, so we can limit the number of output packets
     * by simply dropping them here.
     * Counting encoded video frames needs to be done separately because of
     * reordering, see do_video_out()
     */
    if (!(avctx->codec_type == AVMEDIA_TYPE_VIDEO && avctx->codec)) {
        if (ost->frame_number >= ost->max_frames)
            return;
        ost->frame_number++;
    }

    while (bsfc) {
        AVPacket new_pkt = *pkt;
        int a = av_bitstream_filter_filter(bsfc, avctx, NULL,
                                           &new_pkt.data, &new_pkt.size,
                                           pkt->data, pkt->size,
                                           pkt->flags & AV_PKT_FLAG_KEY);
        if (a > 0) {
            av_free_packet(pkt);
            new_pkt.destruct = av_destruct_packet;
        } else if (a < 0) {
            av_log(NULL, AV_LOG_ERROR, "%s failed for stream %d, codec %s",
                   bsfc->filter->name, pkt->stream_index,
                   avctx->codec ? avctx->codec->name : "copy");
            print_error("", a);
            if (exit_on_error)
                exit_program(1);
        }
        *pkt = new_pkt;

        bsfc = bsfc->next;
    }

    pkt->stream_index = ost->index;
    ret = av_interleaved_write_frame(s, pkt);
    if (ret < 0) {
        print_error("av_interleaved_write_frame()", ret);
        exit_program(1);
    }
}

static int check_recording_time(OutputStream *ost)
{
    OutputFile *of = &output_files[ost->file_index];

    if (of->recording_time != INT64_MAX &&
        av_compare_ts(ost->sync_opts - ost->first_pts, ost->st->codec->time_base, of->recording_time,
                      AV_TIME_BASE_Q) >= 0) {
        ost->is_past_recording_time = 1;
        return 0;
    }
    return 1;
}

static void generate_silence(uint8_t* buf, enum AVSampleFormat sample_fmt, size_t size)
{
    int fill_char = 0x00;
    if (sample_fmt == AV_SAMPLE_FMT_U8)
        fill_char = 0x80;
    memset(buf, fill_char, size);
}

static int encode_audio_frame(AVFormatContext *s, OutputStream *ost,
                              const uint8_t *buf, int buf_size)
{
    AVCodecContext *enc = ost->st->codec;
    AVFrame *frame = NULL;
    AVPacket pkt;
    int ret, got_packet;

    av_init_packet(&pkt);
    pkt.data = NULL;
    pkt.size = 0;

    if (buf) {
        if (!ost->output_frame) {
            ost->output_frame = avcodec_alloc_frame();
            if (!ost->output_frame) {
                av_log(NULL, AV_LOG_FATAL, "out-of-memory in encode_audio_frame()\n");
                exit_program(1);
            }
        }
        frame = ost->output_frame;
        if (frame->extended_data != frame->data)
            av_freep(&frame->extended_data);
        avcodec_get_frame_defaults(frame);

        frame->nb_samples  = buf_size /
                             (enc->channels * av_get_bytes_per_sample(enc->sample_fmt));
        if ((ret = avcodec_fill_audio_frame(frame, enc->channels, enc->sample_fmt,
                                            buf, buf_size, 1)) < 0) {
            av_log(NULL, AV_LOG_FATAL, "Audio encoding failed\n");
            exit_program(1);
        }

        if (!check_recording_time(ost))
            return 0;

        ost->sync_opts += frame->nb_samples;
    }

    got_packet = 0;
    if (avcodec_encode_audio2(enc, &pkt, frame, &got_packet) < 0) {
        av_log(NULL, AV_LOG_FATAL, "Audio encoding failed\n");
        exit_program(1);
    }

    if (got_packet) {
        if (pkt.pts != AV_NOPTS_VALUE)
            pkt.pts      = av_rescale_q(pkt.pts,      enc->time_base, ost->st->time_base);
        if (pkt.duration > 0)
            pkt.duration = av_rescale_q(pkt.duration, enc->time_base, ost->st->time_base);

        write_frame(s, &pkt, ost);

        audio_size += pkt.size;
    }