ad_mpg123: add MP3 decoding through libmpg123

patch by Thomas Orgis, thomas-forum orgis org

git-svn-id: svn://svn.mplayerhq.hu/mplayer/trunk@31590 b3059339-0415-0410-9bf9-f77b7e298cf2

5 files changed, 621 insertions, 0 deletions

diff --git a/Makefile b/Makefile
index 5ee3f49f2a..1d951f40b9 100644
--- a/Makefile
+++ b/Makefile
@@ -175,6 +175,7 @@ SRCS_COMMON-$(LIVE555)               += libmpdemux/demux_rtp.cpp \
                                         stream/stream_live555.c
 SRCS_COMMON-$(MACOSX_FINDER)         += osdep/macosx_finder_args.c
 SRCS_COMMON-$(MNG)                   += libmpdemux/demux_mng.c
+SRCS_COMMON-$(MPG123)                += libmpcodecs/ad_mpg123.c
 
 SRCS_MP3LIB-X86-$(HAVE_AMD3DNOW)     += mp3lib/dct36_3dnow.c \
                                         mp3lib/dct64_3dnow.c
diff --git a/configure b/configure
index 3588cb7714..7853110b74 100755
--- a/configure
+++ b/configure
@@ -322,6 +322,7 @@ Codecs:
   --disable-ladspa          disable LADSPA plugin support [autodetect]
   --disable-libbs2b         disable libbs2b audio filter support [autodetect]
   --disable-libdv           disable libdv 0.9.5 en/decoding support [autodetect]
+  --disable-mpg123          disable libmpg123 MP3 decoding support [autodetect]
   --disable-mad             disable libmad (MPEG audio) support [autodetect]
   --disable-mp3lame         disable LAME MP3 encoding support [autodetect]
   --disable-toolame         disable Toolame (MPEG layer 2) encoding [autodetect]
@@ -586,6 +587,7 @@ _tremor_low=no
 _libvorbis=auto
 _speex=auto
 _theora=auto
+_mpg123=auto
 _mp3lib=auto
 _liba52=auto
 _libdca=auto
@@ -969,6 +971,8 @@ for ac_option do
   --disable-tremor-low) _tremor_low=no  ;;
   --enable-theora)      _theora=yes     ;;
   --disable-theora)     _theora=no      ;;
+  --enable-mpg123)      _mpg123=yes     ;;
+  --disable-mpg123)     _mpg123=no      ;;
   --enable-mp3lib)      _mp3lib=yes     ;;
   --disable-mp3lib)     _mp3lib=no      ;;
   --enable-liba52)      _liba52=yes     ;;
@@ -6405,6 +6409,25 @@ else
 fi
 echores "$_mp3lib"
 
+# Any version of libmpg123 shall be fine.
+echocheck "mpg123 support"
+def_mpg123='#undef CONFIG_MPG123'
+if test "$_mpg123" = auto; then
+  _mpg123=no
+  cat > $TMPC <<EOF
+#include <mpg123.h>
+int main(void){ mpg123_init(); return 0; }
+EOF
+  cc_check -lmpg123 && _mpg123=yes && extra_ldflags="$extra_ldflags -lmpg123"
+fi
+if test "$_mpg123" = yes ; then
+  def_mpg123='#define CONFIG_MPG123 1'
+  codecmodules="mpg123 $codecmodules"
+else
+  nocodecmodules="mpg123 $nocodecmodules"
+fi
+echores "$_mpg123"
+
 echocheck "liba52 support"
 def_liba52='#undef CONFIG_LIBA52'
 if test "$_liba52" = auto ; then
@@ -7831,6 +7854,7 @@ MGA = $_mga
 MNG = $_mng
 MP3LAME = $_mp3lame
 MP3LIB = $_mp3lib
+MPG123 = $_mpg123
 MUSEPACK = $_musepack
 NAS = $_nas
 NATIVE_RTSP = $_native_rtsp
@@ -8149,6 +8173,7 @@ $def_mp3lame
 $def_mp3lame_preset
 $def_mp3lame_preset_medium
 $def_mp3lib
+$def_mpg123
 $def_musepack
 $def_speex
 $def_theora
diff --git a/etc/codecs.conf b/etc/codecs.conf
index c723ebd2cf..7e912e3afd 100644
--- a/etc/codecs.conf
+++ b/etc/codecs.conf
@@ -4163,6 +4163,21 @@ audiocodec mp3
   fourcc "LAME"      ; used in mythtv .nuv files
   driver mp3lib
 
+audiocodec mpg123
+  ; this is preferred over ffmp2/ffmp3 since it is faster, generally
+  info "MPEG 1.0/2.0/2.5 layers I, II, III"
+  status working
+  comment "High-performance decoder using libmpg123."
+  format 0x50        ; layer-1 && layer-2
+  format 0x55        ; layer-3
+  format 0x5500736d  ; "ms\0\x55" older mp3 fcc (MOV files)
+  format 0x5000736d  ; "ms\0\x50" older mp2 fcc (MOV files)
+  format 0x55005354  ;  broken file
+  fourcc ".mp3"      ; CBR/VBR MP3 (MOV files)
+  fourcc "MP3 "      ; used in .nsv files
+  fourcc "LAME"      ; used in mythtv .nuv files
+  driver mpg123
+
 audiocodec ffmp3on4float
   info "FFmpeg Multi-channel MPEG layer-3 on MP4 audio"
   status working
diff --git a/libmpcodecs/ad.c b/libmpcodecs/ad.c
index f81f11ef0f..81cd2a5fca 100644
--- a/libmpcodecs/ad.c
+++ b/libmpcodecs/ad.c
@@ -32,6 +32,7 @@
 /* Missed vorbis, mad, dshow */
 
 //extern ad_functions_t mpcodecs_ad_null;
+extern const ad_functions_t mpcodecs_ad_mpg123;
 extern const ad_functions_t mpcodecs_ad_mp3lib;
 extern const ad_functions_t mpcodecs_ad_ffmpeg;
 extern const ad_functions_t mpcodecs_ad_liba52;
@@ -62,6 +63,9 @@ extern const ad_functions_t mpcodecs_ad_libdca;
 const ad_functions_t * const mpcodecs_ad_drivers[] =
 {
 //  &mpcodecs_ad_null,
+#ifdef CONFIG_MPG123
+  &mpcodecs_ad_mpg123,
+#endif
 #ifdef CONFIG_MP3LIB
   &mpcodecs_ad_mp3lib,
 #endif
diff --git a/libmpcodecs/ad_mpg123.c b/libmpcodecs/ad_mpg123.c
new file mode 100644
index 0000000000..cc2787f305
--- /dev/null
+++ b/libmpcodecs/ad_mpg123.c
@@ -0,0 +1,576 @@
+/*
+ * MPEG 1.0/2.0/2.5 audio layer I, II, III decoding with libmpg123
+ *
+ * Copyright (C) 2010 Thomas Orgis <thomas@orgis.org>
+ *
+ * MPlayer is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * MPlayer 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 General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with MPlayer; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <unistd.h>
+
+#include "config.h"
+
+#include "ad_internal.h"
+
+static const ad_info_t info = {
+    "MPEG 1.0/2.0/2.5 layers I, II, III",
+    "mpg123",
+    "Thomas Orgis",
+    "mpg123.org",
+    "High-performance decoder using libmpg123."
+};
+
+LIBAD_EXTERN(mpg123)
+
+#include "libvo/fastmemcpy.h"
+
+/* We avoid any usage of mpg123 API that is sensitive to the large file
+ * support setting. This ensures compatibility with a wide range of libmpg123
+ * installs. This code is intended to work with version 1.0.0 of libmpg123.
+ *
+ * Though the chosen API subset is not affected by the choice of large file
+ * support, the mpg123 header (old versions of which) might include a check
+ * for matching _FILE_OFFSET_BITS. Since MPlayer does always define this one
+ * for large file support, we are safe for any default mpg123 install that
+ * either doesn't have such checks or defaults to the large value of
+ * _FILE_OFFSET_BITS .
+ * So, in short: There's no worry unless you have a non-default libmpg123
+ * with intentionally disabled large file support. */
+/* You might need to #undef _FILE_OFFSET_BITS here on a 64 bit system
+   with released mpg123 1.12 when using callback API. SVN snapshots
+   should work fine. */
+#include <mpg123.h>
+
+/* Selection of mpg123 usage patterns:
+ * AD_MPG123_CALLBACK: Use callback API instead of feeding of memory buffers.
+ *   That needs mpg123>=1.12, on x86-64 SVN snapshot because of
+ *   _FILE_OFFSET_BITS being defined (see above).
+ * AD_MPG123_PACKET: Use packet-based input (including pts handling).
+ * AD_MPG123_SEEKBUFFER: Use internal mpg123 buffer to enhance stream parsing.
+ *   Makes sense with callback API only.
+ * Any of those might affect I/O performance, might be significant compared
+ * to the excessively optimized decoding.
+ */
+/* #define AD_MPG123_CALLBACK */
+#define AD_MPG123_PACKET
+/* #define AD_MPG123_SEEKBUFFER */
+
+/* Switch for updating bitrate info of VBR files. Not essential. */
+#define AD_MPG123_MEAN_BITRATE
+
+/* Funny thing, that. I assume I shall use it for selecting mpg123 channels.
+ * Please correct me if I guessed wrong. */
+extern int fakemono;
+
+struct ad_mpg123_context {
+    mpg123_handle *handle;
+#ifdef AD_MPG123_MEAN_BITRATE
+    /* Running mean for bit rate, stream length estimation. */
+    float mean_rate;
+    unsigned int mean_count;
+    /* Time delay for updates. */
+    short delay;
+#endif
+    /* If the stream is actually VBR. */
+    char vbr;
+#if (defined AD_MPG123_CALLBACK) && (defined AD_MPG123_PACKET)
+    unsigned char *packet;
+    int packleft;
+#endif
+};
+
+static void context_reset(struct ad_mpg123_context *con)
+{
+#ifdef AD_MPG123_MEAN_BITRATE
+    con->mean_rate  = 0.;
+    con->mean_count = 0;
+    con->delay      = 1;
+#endif
+#if (defined AD_MPG123_CALLBACK) && (defined AD_MPG123_PACKET)
+    con->packet   = NULL;
+    con->packleft = 0;
+#endif
+}
+
+
+#ifdef AD_MPG123_CALLBACK
+/* Mpg123 calls that for retrieving data.
+ * This wrapper is at least needed for the call frame (ssize_t vs. int). */
+static ssize_t read_callback(void *ash, void *buf, size_t count)
+{
+    sh_audio_t *sh = ash;
+#ifdef AD_MPG123_PACKET
+    struct ad_mpg123_context *con = sh->context;
+    unsigned char *target = buf;
+    int need = count;
+    ssize_t got = 0;
+    while (need > 0) {
+        if (con->packleft > 0) {
+            int get = need > con->packleft ? con->packleft : need;
+            /* Any difference to normal memcpy? */
+            fast_memcpy(target, con->packet, get);
+            /* OK, that does look redundant. */
+            con->packet   += get;
+            con->packleft -= get;
+            target += get;
+            need   -= get;
+            got    += get;
+        } else {
+            double pts;
+            /* Feed more input data. */
+            con->packleft = ds_get_packet_pts(sh->ds, &con->packet, &pts);
+            if (con->packleft <= 0)
+                break;          /* Apparently that's it. EOF. */
+
+            /* Next bytes from that presentation time. */
+            if (pts != MP_NOPTS_VALUE) {
+                sh->pts       = pts;
+                sh->pts_bytes = 0;
+            }
+        }
+    }
+    return got;
+#else
+    /* It returns int... with the meaning of byte count. */
+    return (ssize_t) demux_read_data(sh->ds, buf, count);
+#endif
+}
+
+/* Arbitrary input seeking is not supported with this MPlayer API(?).
+   That also means that we won't read any ID3v1 tags. */
+static off_t seek_callback(void *sh, off_t pos, int whence)
+{
+    return -1;
+}
+#endif
+
+/* This initializes libmpg123 and prepares the handle, including funky
+ * parameters. */
+static int preinit(sh_audio_t *sh)
+{
+    int err, flag;
+    struct ad_mpg123_context *con;
+    /* Assumption: You always call preinit + init + uninit, on every file.
+     * But you stop at preinit in case it fails.
+     * If that is not true, one must ensure not to call mpg123_init / exit
+     * twice in a row. */
+    if (mpg123_init() != MPG123_OK)
+        return 0;
+
+    sh->context = malloc(sizeof(struct ad_mpg123_context));
+    con = sh->context;
+    context_reset(con);
+
+    /* Auto-choice of optimized decoder (first argument NULL). */
+    con->handle = mpg123_new(NULL, &err);
+    if (!con->handle)
+        goto bad_end;
+
+#ifdef CONFIG_FAKE_MONO
+    /* Guessing here: Default value triggers forced upmix of mono to stereo. */
+    flag = fakemono == 0 ? MPG123_FORCE_STEREO :
+           fakemono == 1 ? MPG123_MONO_LEFT    :
+           fakemono == 2 ? MPG123_MONO_RIGHT   : 0;
+    if (mpg123_param(con->handle, MPG123_ADD_FLAGS, flag, 0.0) != MPG123_OK)
+        goto bad_end;
+#endif
+#ifdef AD_MPG123_CALLBACK
+    /* The I/O is handled via callbacks to MPlayer stream functions,
+     * actually only the reading, as general seeking does not seem to be available */
+    if (mpg123_replace_reader_handle(con->handle, read_callback,
+                                     seek_callback, NULL) != MPG123_OK) {
+        mp_msg(MSGT_DECAUDIO, MSGL_ERR, "mpg123 error: %s\n",
+               mpg123_strerror(con->handle));
+        mpg123_exit();
+        return 0;
+    }
+#endif
+
+    /* Basic settings.
+     * Don't spill messages, enable better resync with non-seekable streams.
+     * Give both flags individually without error checking to keep going with
+     * old libmpg123. Generally, it is not fatal if the flags are not
+     * honored */
+    mpg123_param(con->handle, MPG123_ADD_FLAGS, MPG123_QUIET, 0.0);
+    /* Old headers don't know MPG123_SEEKBUFFER yet, so use the plain 0x100. */
+#ifdef AD_MPG123_SEEKBUFFER
+    mpg123_param(con->handle, MPG123_ADD_FLAGS, 0x100, 0.0);
+#endif
+
+    /* Open decisions: Configure libmpg123 to force encoding (or stay open about
+     * library builds that support only float or int32 output), (de)configure
+     * gapless decoding (won't work with seeking in MPlayer, though).
+     * Don't forget to eventually enable ReplayGain/RVA support, too.
+     * Let's try to run with the default for now. */
+
+    /* Example for RVA choice (available since libmpg123 1.0.0):
+    mpg123_param(con->handle, MPG123_RVA, MPG123_RVA_MIX, 0.0) */
+
+    return 1;
+
+  bad_end:
+    if (!con->handle)
+        mp_msg(MSGT_DECAUDIO, MSGL_ERR, "mpg123 preinit error: %s\n",
+               mpg123_plain_strerror(err));
+    else
+        mp_msg(MSGT_DECAUDIO, MSGL_ERR, "mpg123 preinit error: %s\n",
+               mpg123_strerror(con->handle));
+
+    if (con->handle)
+        mpg123_delete(con->handle);
+    mpg123_exit();
+    free(sh->context);
+    sh->context = NULL;
+    return 0;
+}
+
+/* Compute bitrate from frame size. */
+static int compute_bitrate(struct mpg123_frameinfo *i)
+{
+    static const int samples_per_frame[4][4] = {
+        {-1, 384, 1152, 1152},  /* MPEG 1 */
+        {-1, 384, 1152,  576},  /* MPEG 2 */
+        {-1, 384, 1152,  576},  /* MPEG 2.5 */
+        {-1,  -1,   -1,   -1},  /* Unknown */
+    };
+    return (int) ((i->framesize + 4) * 8 * i->rate * 0.001 /
+                  samples_per_frame[i->version][i->layer] + 0.5);
+}
+
+/* Opted against the header printout from old mp3lib, too much
+ * irrelevant info. This is modelled after the mpg123 app's
+ * standard output line.
+ * If more verbosity is demanded, one can add more detail and
+ * also throw in ID3v2 info which libmpg123 collects anyway. */
+static void print_header_compact(struct mpg123_frameinfo *i)
+{
+    static const char *smodes[5] = {
+        "stereo", "joint-stereo", "dual-channel", "mono", "invalid"
+    };
+    static const char *layers[4] = {
+        "Unknown", "I", "II", "III"
+    };
+    static const char *versions[4] = {
+        "1.0", "2.0", "2.5", "x.x"
+    };
+
+    mp_msg(MSGT_DECAUDIO, MSGL_V, "MPEG %s layer %s, ",
+           versions[i->version], layers[i->layer]);
+    switch (i->vbr) {
+    case MPG123_CBR:
+        if (i->bitrate)
+            mp_msg(MSGT_DECAUDIO, MSGL_V, "%d kbit/s", i->bitrate);
+        else
+            mp_msg(MSGT_DECAUDIO, MSGL_V, "%d kbit/s (free format)",
+                   compute_bitrate(i));
+        break;
+    case MPG123_VBR:
+        mp_msg(MSGT_DECAUDIO, MSGL_V, "VBR");
+        break;
+    case MPG123_ABR:
+        mp_msg(MSGT_DECAUDIO, MSGL_V, "%d kbit/s ABR", i->abr_rate);
+        break;
+    default:
+        mp_msg(MSGT_DECAUDIO, MSGL_V, "???");
+    }
+    mp_msg(MSGT_DECAUDIO, MSGL_V, ", %ld Hz %s\n", i->rate,
+           smodes[i->mode]);
+}
+
+#ifndef AD_MPG123_CALLBACK
+/* This tries to extract a requested amount of decoded data.
+ * Even when you request 0 bytes, it will feed enough input so that
+ * the decoder _could_ have delivered something.
+ * Returns byte count >= 0, -1 on error.
+ *
+ * Thoughts on exact pts keeping:
+ * We have to assume that MPEG frames are cut in pieces by packet boundaries.
+ * Also, it might be possible that the first packet does not contain enough
+ * data to ensure initial stream sync... or re-sync on erroneous streams.
+ * So we need something robust to relate the decoded byte count to the correct
+ * time stamp. This is tricky, though. From the outside, you cannot tell if,
+ * after having fed two packets until the first output arrives, one should
+ * start counting from the first packet's pts or the second packet's.
+ * So, let's just count from the last fed package's pts. If the packets are
+ * exactly cut to MPEG frames, this will cause one frame mismatch in the
+ * beginning (when mpg123 peeks ahead for the following header), but will
+ * be corrected with the third frame already. One might add special code to
+ * not increment the base pts past the first packet's after a resync before
+ * the first decoded bytes arrived. */
+static int decode_a_bit(sh_audio_t *sh, unsigned char *buf, int count)
+{
+    int ret = MPG123_OK;
+    int got = 0;
+    struct ad_mpg123_context *con = sh->context;
+
+    /* There will be one MPG123_NEW_FORMAT message on first open.
+     * This will be implicitly handled in reopen_stream(). */
+    do {
+        size_t got_now = 0;
+        ret = mpg123_decode(con->handle, NULL, 0, buf + got, count - got,
+                            &got_now);
+        got += got_now;
+#ifdef AD_MPG123_PACKET
+        sh->pts_bytes += got_now;
+#endif
+
+        if (ret == MPG123_NEED_MORE) {
+            int incount;
+#ifdef AD_MPG123_PACKET
+            double pts;
+            unsigned char *inbuf;
+            /* Feed more input data. */
+            incount = ds_get_packet_pts(sh->ds, &inbuf, &pts);
+            if (incount <= 0)
+                break;          /* Apparently that's it. EOF. */
+
+            /* Next bytes from that presentation time. */
+            if (pts != MP_NOPTS_VALUE) {
+                sh->pts       = pts;
+                sh->pts_bytes = 0;
+            }
+#else
+            const int inbufsize = 4096;
+            unsigned char inbuf[inbufsize];
+            /* Feed more input data. */
+            incount = demux_read_data(((sh_audio_t *) sh)->ds,
+                                      inbuf, inbufsize);
+            if (incount == 0)
+                break;          /* Apparently that's it. EOF. */
+#endif
+
+            /* Do not use mpg123_feed(), added in later libmpg123 versions. */
+            ret = mpg123_decode(con->handle, inbuf, incount, NULL, 0, NULL);
+            /* Return value is checked in the loop condition.
+             * It could be MPG12_OK now, it could need more. */
+        }
+        /* Older mpg123 versions might indicate MPG123_DONE, so be prepared. */
+        else if (ret == MPG123_ERR || ret == MPG123_DONE)
+            break;
+
+    } while (ret == MPG123_NEED_MORE || got < count);
+
+    if (ret == MPG123_ERR) {
+        mp_msg(MSGT_DECAUDIO, MSGL_ERR, "mpg123 decoding failed: %s\n",
+               mpg123_strerror(con->handle));
+        mpg123_close(con->handle);
+        return -1;
+    }
+
+    return got;
+}
+#endif
+
+/* Close, reopen stream. Feed data until we know the format of the stream.
+ * 1 on success, 0 on error */
+static int reopen_stream(sh_audio_t *sh)
+{
+    long rate;
+    int chan, enc;
+    struct ad_mpg123_context *con = (struct ad_mpg123_context*) sh->context;
+
+    mpg123_close(con->handle);
+    context_reset(con);
+
+#ifdef AD_MPG123_CALLBACK
+    if (MPG123_OK == mpg123_open_handle(con->handle, sh) &&
+#else
+    if (/* Open and make sure we have fed enough data to get stream properties. */
+        MPG123_OK == mpg123_open_feed(con->handle) &&
+        /* Feed data until mpg123 is ready (has found stream beginning). */
+        !decode_a_bit(sh, NULL, 0) &&
+#endif
+        /* Not handing NULL pointers for compatibility with old libmpg123. */
+        MPG123_OK == mpg123_getformat(con->handle, &rate, &chan, &enc)) {
+        return 1;
+    } else {
+        mp_msg(MSGT_DECAUDIO, MSGL_ERR,
+               "mpg123 failed to reopen stream: %s\n",
+               mpg123_strerror(con->handle));
+        mpg123_close(con->handle);
+        return 0;
+    }
+}
+
+/* Now we really start accessing some data and determining file format.
+ * Paranoia note: The mpg123_close() on errors is not really necessary,
+ * But it ensures that we don't accidentally continue decoding with a
+ * bad state (possibly interpreting the format badly or whatnot). */
+static int init(sh_audio_t *sh)
+{
+    long rate    = 0;
+    int channels = 0;
+    int encoding = 0;
+    mpg123_id3v2 *v2;
+    struct mpg123_frameinfo finfo;
+    struct ad_mpg123_context *con = sh->context;
+
+    /* We're open about any output format that libmpg123 will suggest.
+     * Note that a standard build will always default to 16 bit signed and
+     * the native sample rate of the file. */
+    if (MPG123_OK == mpg123_format_all(con->handle) &&
+        reopen_stream(sh) &&
+        MPG123_OK == mpg123_getformat(con->handle, &rate, &channels, &encoding) &&
+        /* Forbid the format to change later on. */
+        MPG123_OK == mpg123_format_none(con->handle) &&
+        MPG123_OK == mpg123_format(con->handle, rate, channels, encoding) &&
+        /* Get MPEG header info. */
+        MPG123_OK == mpg123_info(con->handle, &finfo) &&
+        /* Since we queried format, mpg123 should have read past ID3v2 tags.
+         * We need to decide if printing of UTF-8 encoded text info is wanted. */
+        MPG123_OK == mpg123_id3(con->handle, NULL, &v2)) {
+        /* If we are here, we passed all hurdles. Yay! Extract the info. */
+        print_header_compact(&finfo);
+        /* Do we want to print out the UTF-8 Id3v2 info?
+        if (v2)
+            print_id3v2(v2); */
+
+        /* Have kb/s, want B/s
+         * For VBR, the first frame will be a bad estimate. */
+        sh->i_bps = (finfo.bitrate ? finfo.bitrate : compute_bitrate(&finfo))
+                    * 1000 / 8;
+        context_reset(con);
+        con->vbr = (finfo.vbr != MPG123_CBR);
+        sh->channels   = channels;
+        sh->samplerate = rate;
+        /* Without external force, mpg123 will always choose signed encoding,
+         * and non-16-bit only on builds that don't support it.
+         * Be reminded that it doesn't matter to the MPEG file what encoding
+         * is produced from it. */
+        switch (encoding) {
+        case MPG123_ENC_SIGNED_8:
+            sh->sample_format = AF_FORMAT_S8;
+            sh->samplesize    = 1;
+            break;
+        case MPG123_ENC_SIGNED_16:
+            sh->sample_format = AF_FORMAT_S16_NE;
+            sh->samplesize    = 2;
+            break;
+        /* To stay compatible with the oldest libmpg123 headers, do not rely
+         * on float and 32 bit encoding symbols being defined.
+         * Those formats came later */
+        case 0x1180: /* MPG123_ENC_SIGNED_32 */
+            sh->sample_format = AF_FORMAT_S32_NE;
+            sh->samplesize    = 4;
+            break;
+        case 0x200: /* MPG123_ENC_FLOAT_32 */
+            sh->sample_format = AF_FORMAT_FLOAT_NE;
+            sh->samplesize    = 4;
+            break;
+        default:
+            mp_msg(MSGT_DECAUDIO, MSGL_ERR,
+                   "Bad encoding from mpg123: %i.\n", encoding);
+            mpg123_close(con->handle);
+            return 0;
+        }
+
+        return 1;
+    } else {
+        mp_msg(MSGT_DECAUDIO, MSGL_ERR, "mpg123 init error: %s\n",
+               mpg123_strerror(con->handle));
+        mpg123_close(con->handle);
+        return 0;
+    }
+}
+
+static void uninit(sh_audio_t *sh)
+{
+    struct ad_mpg123_context *con = (struct ad_mpg123_context*) sh->context;
+
+    mpg123_close(con->handle);
+    mpg123_delete(con->handle);
+    free(sh->context);
+    sh->context = NULL;
+    mpg123_exit();
+}
+
+#ifdef AD_MPG123_MEAN_BITRATE
+/* Update mean bitrate. This could be dropped if accurate time display
+ * on audio file playback is not desired. */
+static void update_info(sh_audio_t *sh)
+{
+    struct ad_mpg123_context *con = sh->context;
+    if (con->vbr && --con->delay < 1) {
+        struct mpg123_frameinfo finfo;
+        if (MPG123_OK == mpg123_info(con->handle, &finfo)) {
+            if (++con->mean_count > ((unsigned int) -1) / 2)
+                con->mean_count = ((unsigned int) -1) / 4;
+
+            /* Might not be numerically optimal, but works fine enough. */
+            con->mean_rate = ((con->mean_count - 1) * con->mean_rate +
+                              finfo.bitrate) / con->mean_count;
+            sh->i_bps = (int) (con->mean_rate * 1000 / 8);
+
+            con->delay = 10;
+        }
+    }
+}
+#endif
+
+static int decode_audio(sh_audio_t *sh, unsigned char *buf, int minlen,
+                        int maxlen)
+{
+    int bytes;
+
+#ifdef AD_MPG123_CALLBACK
+    struct ad_mpg123_context *con = sh->context;
+    size_t got_bytes = 0;
+    if (MPG123_ERR == mpg123_read(con->handle, buf, minlen, &got_bytes)) {
+        mp_msg(MSGT_DECAUDIO, MSGL_ERR, "Decoding error in mpg123: %s\n",
+               mpg123_strerror(con->handle));
+        return -1;
+    }
+#ifdef AD_MPG123_PACKET
+    sh->pts_bytes += got_bytes;
+#endif
+    bytes = got_bytes;
+#else
+    bytes = decode_a_bit(sh, buf, minlen);
+#endif
+
+    if (bytes == 0)
+        return -1;              /* EOF */
+
+#ifdef AD_MPG123_MEAN_BITRATE
+    update_info(sh);
+#endif
+
+    return bytes;
+}
+
+static int control(sh_audio_t *sh, int cmd, void *arg, ...)
+{
+    switch (cmd) {
+    case ADCTRL_RESYNC_STREAM:
+        /* Close/reopen the stream for mpg123 to make sure it doesn't
+         * think that it still knows the exact stream position.
+         * Otherwise, we would have funny effects from the gapless code.
+         * Oh, and it helps to minimize artifacts from jumping in the stream. */
+        if (reopen_stream(sh)) {
+#ifdef AD_MPG123_MEAN_BITRATE
+            update_info(sh);
+#endif
+            return CONTROL_TRUE;
+        } else {
+            mp_msg(MSGT_DECAUDIO, MSGL_ERR,
+                   "mpg123 cannot reopen stream for resync.\n");
+            return CONTROL_FALSE;
+        }
+        break;
+    }
+    return CONTROL_UNKNOWN;
+}