Wiki Page Content

Differences between revisions 39 and 40
Revision 39 as of 2015-10-29 21:24:42
Size: 4173
Comment: Added comment about SDL_zero() in example (Feedback 2015-10-03).
Revision 40 as of 2019-05-15 14:14:25
Size: 4605
Editor: ColaEuphoria
Comment: Removed misleading "clarity" note and ancient sampling rates games haven't used for 20 years.
Deletions are marked like this. Additions are marked like this.
Line 36: Line 36:
'''freq''' specifies the number of sample frames sent to the sound device per second. Common values are 11025, 22050, 44100 and 48000. Larger values produce cleaner audio, in much the same way that larger resolutions produce cleaner graphics. '''freq''' specifies the number of sample frames sent to the sound device per second. The Nyquist Theorem states that the audio sampling frequency must be exactly twice the highest frequency represented in the audio. Humans can hear up to slightly under 20kHz, declining to 16kHz or lower as we age. Standard CD quality audio uses 44100. DVDs and the [[https://opus-codec.org/|Opus audio codec]] use 48000. Values higher than 48000 generally should not be used for playback purposes because they use more memory, use more CPU, and can cause other problems as explained in [[https://people.xiph.org/~xiphmont/demo/neil-young.html|this blog post by Chris Montgomery of Xiph]].

SDL_AudioSpec

A structure that contains the audio output format. It also contains a callback that is called when the audio device needs more data.

Data Fields

int

freq

DSP frequency (samples per second); see Remarks for details

SDL_AudioFormat

format

audio data format; see Remarks for details

Uint8

channels

number of separate sound channels: see Remarks for details

Uint8

silence

audio buffer silence value (calculated)

Uint16

samples

audio buffer size in samples (power of 2); see Remarks for details

Uint32

size

audio buffer size in bytes (calculated)

SDL_AudioCallback

callback

the function to call when the audio device needs more data; see Remarks for details

void*

userdata

a pointer that is passed to callback (otherwise ignored by SDL)

Code Examples

SDL_AudioSpec want, have;
SDL_AudioDeviceID dev;

SDL_memset(&want, 0, sizeof(want)); /* or SDL_zero(want) */
want.freq = 48000;
want.format = AUDIO_F32;
want.channels = 2;
want.samples = 4096;
want.callback = MyAudioCallback;  // you wrote this function elsewhere.
dev = SDL_OpenAudioDevice(NULL, 0, &want, &have, SDL_AUDIO_ALLOW_FORMAT_CHANGE);

Remarks

This structure is used by SDL_OpenAudioDevice() and SDL_LoadWAV(). While all fields are used by SDL_OpenAudioDevice(), only freq, format, channels, and samples are used by SDL_LoadWAV().

freq specifies the number of sample frames sent to the sound device per second. The Nyquist Theorem states that the audio sampling frequency must be exactly twice the highest frequency represented in the audio. Humans can hear up to slightly under 20kHz, declining to 16kHz or lower as we age. Standard CD quality audio uses 44100. DVDs and the Opus audio codec use 48000. Values higher than 48000 generally should not be used for playback purposes because they use more memory, use more CPU, and can cause other problems as explained in this blog post by Chris Montgomery of Xiph.

format specifies the size and type of each sample element and may be one of the following:

8-bit support

AUDIO_S8

signed 8-bit samples

AUDIO_U8

unsigned 8-bit samples

16-bit support

AUDIO_S16LSB

signed 16-bit samples in little-endian byte order

AUDIO_S16MSB

signed 16-bit samples in big-endian byte order

AUDIO_S16SYS

signed 16-bit samples in native byte order

AUDIO_S16

AUDIO_S16LSB

AUDIO_U16LSB

unsigned 16-bit samples in little-endian byte order

AUDIO_U16MSB

unsigned 16-bit samples in big-endian byte order

AUDIO_U16SYS

unsigned 16-bit samples in native byte order

AUDIO_U16

AUDIO_U16LSB

32-bit support (new to SDL 2.0)

AUDIO_S32LSB

32-bit integer samples in little-endian byte order

AUDIO_S32MSB

32-bit integer samples in big-endian byte order

AUDIO_S32SYS

32-bit integer samples in native byte order

AUDIO_S32

AUDIO_S32LSB

float support (new to SDL 2.0)

AUDIO_F32LSB

32-bit floating point samples in little-endian byte order

AUDIO_F32MSB

32-bit floating point samples in big-endian byte order

AUDIO_F32SYS

32-bit floating point samples in native byte order

AUDIO_F32

AUDIO_F32LSB

See SDL_AudioFormat for more info.

channels specifies the number of output channels. As of SDL 2.0, supported values are 1 (mono), 2 (stereo), 4 (quad), and 6 (5.1).

samples specifies a unit of audio data. When used with SDL_OpenAudioDevice() this refers to the size of the audio buffer in sample frames. A sample frame is a chunk of audio data of the size specified in format multiplied by the number of channels. When the SDL_AudioSpec is used with SDL_LoadWAV() samples is set to 4096. This field's value must be a power of two.

The values silence and size are calculated by SDL_OpenAudioDevice().

Channel data is interleaved. Stereo samples are stored in left/right ordering. Quad is stored in front-left/front-right/rear-left/rear-right order. 5.1 is stored in front-left/front-right/center/low-freq/rear-left/rear-right ordering ("low-freq" is the ".1" speaker).

The function prototype for callback is:

void SDL_AudioCallback(void*  userdata,
                       Uint8* stream,
                       int    len)
  • where its parameters are:

    userdata

    an application-specific parameter saved in the SDL_AudioSpec structure's userdata field

    stream

    a pointer to the audio data buffer filled in by SDL_AudioCallback()

    len

    the length of that buffer in bytes

Once the callback returns, the buffer will no longer be valid. Stereo samples are stored in a LRLRLR ordering.

The callback must completely initialize the buffer; as of SDL 2.0, this buffer is not initialized before the callback is called. If there is nothing to play, the callback should fill the buffer with silence.

With SDL >= 2.0.4 you can choose to avoid callbacks and use SDL_QueueAudio() instead, if you like. Just open your audio device with a NULL callback.


CategoryStruct, CategoryAudio

None: SDL_AudioSpec (last edited 2019-05-15 14:14:25 by ColaEuphoria)

Feedback
Please include your contact information if you'd like to receive a reply.
Submit