path: root/Documentation/DocBook/media/dvb/audio.xml

<title>DVB Audio Device</title>
<para>The DVB audio device controls the MPEG2 audio decoder of the DVB hardware. It
can be accessed through <emphasis role="tt">/dev/dvb/adapter0/audio0</emphasis>. Data types and and
ioctl definitions can be accessed by including <emphasis role="tt">linux/dvb/audio.h</emphasis> in your
application.
</para>
<para>Please note that some DVB cards don&#8217;t have their own MPEG decoder, which results in
the omission of the audio and video device.
</para>
<para>
These ioctls were also used by V4L2 to control MPEG decoders implemented in V4L2. The use
of these ioctls for that purpose has been made obsolete and proper V4L2 ioctls or controls
have been created to replace that functionality.</para>

<section id="audio_data_types">
<title>Audio Data Types</title>
<para>This section describes the structures, data types and defines used when talking to the
audio device.
</para>

<section id="audio-stream-source-t">
<title>audio_stream_source_t</title>
<para>The audio stream source is set through the AUDIO_SELECT_SOURCE call and can take
the following values, depending on whether we are replaying from an internal (demux) or
external (user write) source.
</para>
<programlisting>
typedef enum {
	AUDIO_SOURCE_DEMUX,
	AUDIO_SOURCE_MEMORY
} audio_stream_source_t;
</programlisting>
<para>AUDIO_SOURCE_DEMUX selects the demultiplexer (fed either by the frontend or the
DVR device) as the source of the video stream. If AUDIO_SOURCE_MEMORY
is selected the stream comes from the application through the <emphasis role="tt">write()</emphasis> system
call.
</para>

</section>
<section id="audio-play-state-t">
<title>audio_play_state_t</title>
<para>The following values can be returned by the AUDIO_GET_STATUS call representing the
state of audio playback.
</para>
<programlisting>
typedef enum {
	AUDIO_STOPPED,
	AUDIO_PLAYING,
	AUDIO_PAUSED
} audio_play_state_t;
</programlisting>

</section>
<section id="audio-channel-select-t">
<title>audio_channel_select_t</title>
<para>The audio channel selected via AUDIO_CHANNEL_SELECT is determined by the
following values.
</para>
<programlisting>
typedef enum {
	AUDIO_STEREO,
	AUDIO_MONO_LEFT,
	AUDIO_MONO_RIGHT,
	AUDIO_MONO,
	AUDIO_STEREO_SWAPPED
} audio_channel_select_t;
</programlisting>

</section>
<section id="audio-status">
<title>struct audio_status</title>
<para>The AUDIO_GET_STATUS call returns the following structure informing about various
states of the playback operation.
</para>
<programlisting>
typedef struct audio_status {
	boolean AV_sync_state;
	boolean mute_state;
	audio_play_state_t play_state;
	audio_stream_source_t stream_source;
	audio_channel_select_t channel_select;
	boolean bypass_mode;
	audio_mixer_t mixer_state;
} audio_status_t;
</programlisting>

</section>
<section id="audio-mixer">
<title>struct audio_mixer</title>
<para>The following structure is used by the AUDIO_SET_MIXER call to set the audio
volume.
</para>
<programlisting>
typedef struct audio_mixer {
	unsigned int volume_left;
	unsigned int volume_right;
} audio_mixer_t;
</programlisting>

</section>
<section id="audio_encodings">
<title>audio encodings</title>
<para>A call to AUDIO_GET_CAPABILITIES returns an unsigned integer with the following
bits set according to the hardwares capabilities.
</para>
<programlisting>
 #define AUDIO_CAP_DTS    1
 #define AUDIO_CAP_LPCM   2
 #define AUDIO_CAP_MP1    4
 #define AUDIO_CAP_MP2    8
 #define AUDIO_CAP_MP3   16
 #define AUDIO_CAP_AAC   32
 #define AUDIO_CAP_OGG   64
 #define AUDIO_CAP_SDDS 128
 #define AUDIO_CAP_AC3  256
</programlisting>

</section>
<section id="audio-karaoke">
<title>struct audio_karaoke</title>
<para>The ioctl AUDIO_SET_KARAOKE uses the following format:
</para>
<programlisting>
typedef
struct audio_karaoke {
	int vocal1;
	int vocal2;
	int melody;
} audio_karaoke_t;
</programlisting>
<para>If Vocal1 or Vocal2 are non-zero, they get mixed into left and right t at 70% each. If both,
Vocal1 and Vocal2 are non-zero, Vocal1 gets mixed into the left channel and Vocal2 into the
right channel at 100% each. Ff Melody is non-zero, the melody channel gets mixed into left
and right.
</para>

</section>
<section id="audio-attributes-t">
<title>audio attributes</title>
<para>The following attributes can be set by a call to AUDIO_SET_ATTRIBUTES:
</para>
<programlisting>
 typedef uint16_t audio_attributes_t;
 /&#x22C6;   bits: descr. &#x22C6;/
 /&#x22C6;   15-13 audio coding mode (0=ac3, 2=mpeg1, 3=mpeg2ext, 4=LPCM, 6=DTS, &#x22C6;/
 /&#x22C6;   12    multichannel extension &#x22C6;/
 /&#x22C6;   11-10 audio type (0=not spec, 1=language included) &#x22C6;/
 /&#x22C6;    9- 8 audio application mode (0=not spec, 1=karaoke,