Keyboard shortcuts

Press ← or → to navigate between chapters

Press S or / to search in the book

Press ? to show this help

Press Esc to hide this help

Kernel drivers are highly unstable when using ALSA API manually and they may cause a kernel panic. Consider interfacing with Audio HAL instead.

Audio

Modern kernel drivers use kernel streaming, multiple buffers are allocated and are routed dynamically using ALSA controls.

Terminology

  • ADDA - Analog to Digital & Digital to Analog Converter. Usually used for headphones due to ADDA's ability to handle both headset and it's microphone.
  • I2S - Inter-Integrated Circuit Sound. Usually used for speakers.
  • DL - Downlink/sink.
  • UL - Uplink/source.

Buffers

  • Playback_1/DL1 - Default playback buffer.
  • Playback_2/DL2 - Isolated deep buffer.
  • Playback_3/DL3 - Isolated deep buffer fallback, used when Playback_2 isn't available.
  • Playback_12/DL12 - VoIP playback buffer.

Even though these buffers have specific roles in Audio HAL, they are multipurpose and can be used for anything (e.g. Playback_12 can be used for music playback).

ALSA Controls

Scenario

  • deep_buffer_scenario - Controls if route is using deep buffer or not.

Headphones

  • HPL Mux - Headphones left channel muxer. Default is Open, should switch to Audio Playback when playing audio through headphones.
  • HPR Mux - Headphones right channel muxer. Default is Open, should switch to Audio Playback when playing audio through headphones.
  • ADDA_DL_GAIN - ADDA total gain.

Routing

Note: Usually CH1 is left channel and CH2 is right channel. It's not possible to route CH1 to CH2 and vice versa.

  • ADDA_DL_CHn DLx_CHn - Route channel n of ADDA to channel n of downlink x.
  • I2Sx_CHn DLy_CHn - Route channel n of I2S device x to channel n of downlink y.

Hi-Res Audio

Kernel drivers provide sample formats S24_LE and S32_LE as well as sample rates up to 192000.

If you get distorted audio when using S24_LE or S32_LE sample formats, it means your volume is too high, you can try lowering it by 48dB:

# ffmpeg -i [path to audio file] -af volume=-48dB -ac 2 -ar 48000 -f s32le - | aplay -D hw:0,2 -r 48000 -c 2 -f S32_LE

Note: Some devices might have problems with sample rates above 48000.

Examples

Routing deep buffer to headphones and playing audio using aplay

# amixer sset "HPL Mux" "Audio Playback"
# amixer sset "HPR Mux" "Audio Playback"
# amixer sset "ADDA_DL_CH1 DL2_CH1" on
# amixer sset "ADDA_DL_CH2 DL2_CH2" on
# amixer sset "deep_buffer_scenario" on
# aplay -D hw:0,2 [path to .wav file]