Linux Audio Support¶
TODO: Mention J23 and J8
Audio support for linux is based on the Linux for Qualcomm Software support. Information included here can also be reviewed on Qualcomm's site:
Development Board Configuration¶
Ensure that Switch S7 is set properly.- Switch 3 should be set to enable Audio, which is a top level enable to send 1.8 V to the audio codec chip SGTL5000XNBA3 (U16).
- Switch 4 can be either set or not set
- If Switch 4 is not set the 1.8V comes from the DC supply on the Development Kit
- If Switch 4 is set the 1.8V comes from the SOM
Hardware Overview¶
A hardware block diagram of the audio subsystem is shown below. The following components are involved on the MitySOM-QC6490 devkit:
- Application processor – This is the ARM CPU complex and handles audio processing tasks, including:
- Managing audio record and playback
- Decoding audio formats
- Using LPAI for postprocessing tasks
- Low Power AI (LPAI) – Subsystem that runs audio playback/record and voice-activation (VA) algorithms. It integrates with a dedicated Qualcomm® Hexagon™ Processor (QDSP6) and a low power island (LPI).
- Audio codec – For the devkit this is a SGTL5000 and it is connected to the LPAI via I2S0 connections. It provides:
- Analog-to-Digital Converter (ADC) for capturing microphone / line-in signals and converting them to I2S
- Digital-to-Analog Converter (DAC) for converting stereo I2S data to analog outptus for either the Headphones or line-out signals
- Headphone / Mic Jack
Software Overview (linux)¶
Note: Audio support for the MitySOM-QC6490 Devkit is not available in the 1.0 release of the BSP.
Examples Using the Headphone Jack (J8 on the Devkit)¶
This section covers some basic examples to checkout the audio functions on the Devkit J8 (CITA standard headphone+mic) audio jack. A simple headset with a mic was used to test these examples.
Capture from microphone¶
Using gstreamer¶
The development kit includes a pulsesink gstreamer plugin to support capture from pulse audio capture channels.
To capture a WAV file using streamer commands from the microphone:
gst-launch-1.0 pulsesrc ! audioconvert ! audioresample ! wavenc ! filesink location=output.wav
The gstreamer tools in the reference filesystem image also include support for recording an mp3 file or a flac compressed file with the following example commands:
gst-launch-1.0 pulsesrc ! audioconvert ! audioresample ! lamemp3enc bitrate=192 ! id3v2mux ! filesink location=output.mp3
gst-launch-1.0 pulsesrc ! flacenc ! flacparse ! filesink location=capture.flac
Using pulseaudio tools¶
Use the following command to record a WAV file from the microphone. Hit CTRL-C to stop:
parec -v --rate=48000 --format=s16le --channels=1 --file-format=wav test.wav
Playback to headphones¶
Using gstreamer¶
The development kit includes a pulsesink plugin that will configure the back end output of a PCM stream to route to the pulseaudio framework.
To play a simple 400 Hz test tone to the headphones (in stereo), run the following command. Note: all gstreamer commands can be cancelled / terminated using CTRL-C.
gst-launch-1.0 audiotestsrc ! audioconvert ! audio/x-raw,channels=2 ! pulsesink
To play an mp3 file, the following command could be used:
gst-play-1.0 02-Jump.mp3
Using pulseaudio tools¶
You can play back a WAV file, such as the one attached on this page, with the following command:
paplay LRMonoPhase4.wav -v
Examples Using the Audio Header (J23 on the Devkit)¶
This section covers some basic examples to checkout the audio functions on the Devkit J23 audio header. To test these examples, an adapter was used to connect a 3.5mm jack input/output to the 5 pin header. Input was ran from a PC but could be ran from any powered source. Output was ran to a simple headset and a set of speakers.
Capture from input (LINE_IN)¶
Record input¶
To capture input from a PC or other powered source, connect the input lines of J23 to whatever audio source you'd like to capture. Before trying to capture audio, ensure your tinymix settings are set similar to the ones below (this can be seen by running tinymix contents, as seen below):
root@qcs6490-mitysom-devkit:~# tinymix contents Number of controls: 28 ctl type num name device value 0 INT 2 PCM Playback Volume 0144, 144 (range 0->192) 1 INT 2 Capture Volume 015, 15 (range 0->15) 2 BOOL 1 Capture Attenuate Switch (-6dB) 0Off 3 BOOL 1 Capture ZC Switch 0On 4 BOOL 1 Capture Switch 0On 5 INT 2 Headphone Playback Volume 0127, 127 (range 0->127) 6 BOOL 1 Headphone Playback Switch 0On 7 BOOL 1 Headphone Playback ZC Switch 0On 8 INT 1 Mic Volume 03 (range 0->3) 9 INT 2 Lineout Playback Volume 031, 31 (range 0->31) 10 BOOL 1 Lineout Playback Switch 0On 11 INT 1 DAP Main channel 032768 (range 0->65535) 12 INT 1 DAP Mix channel 00 (range 0->65535) 13 BOOL 1 AVC Switch 0Off 14 BOOL 1 AVC Hard Limiter Switch 0Off 15 INT 1 AVC Max Gain Volume 01 (range 0->2) 16 INT 1 AVC Integrator Response 01 (range 0->3) 17 INT 1 AVC Threshold Volume 012 (range 0->96) 18 INT 1 BASS 0 047 (range 0->95) 19 INT 1 BASS 1 047 (range 0->95) 20 INT 1 BASS 2 047 (range 0->95) 21 INT 1 BASS 3 047 (range 0->95) 22 INT 1 BASS 4 047 (range 0->95) 23 ENUM 1 Capture Mux 0MIC_IN, > LINE_IN, 24 ENUM 1 Headphone Mux 0DAC, > LINE_IN, 25 ENUM 1 Digital Input Mux 0> ADC, I2S, Rsvrd, DAP, 26 ENUM 1 DAP Mux 0> ADC, I2S, 27 ENUM 1 DAP MIX Mux 0> ADC, I2S,
The most important settings to pay attention to are
Capture Mux, which must be set to LINE_IN, Capture Switch, which must be On, and Capture Volume, which should not be 0. If you have these settings and are encountering issues capturing audio, the configuration above is from a successful capture. Checking that your other settings match the above ones could be beneficial in alleviating issues. After configuring the CODEC with tinymix, running gst-launch-1.0 pulsesrc ! audioconvert ! audioresample ! wavenc ! filesink location=output.wav and then providing audio input from whatever source is used will allow audio to be captured and stored as output.wav. To playback the recorded audio, you can follow the above instructions on playback.
Live playback¶
Headphones (J8)
To get live playback over the headphones, set the following tinymix settings:
root@qcs6490-mitysom-devkit:~# tinymix set 'Headphone Mux' 1 root@qcs6490-mitysom-devkit:~# tinymix set 'Headphone Playback Volume' 127 127 root@qcs6490-mitysom-devkit:~# tinymix set 'Headphone Playback Switch' 1
Then, play audio from your source plugged into LINE_IN (J23), and it should be heard over the headphones plugged into the audio jack (J8).
Playback to audio output (LINEOUT)¶
To play audio over LINEOUT (J23), ensure the following tinymix settings are set and that something is connected to J23 to recieve the audio output (such as speakers or headphones).
root@qcs6490-mitysom-devkit:~# tinymix set 'Lineout Playback Switch' 1 root@qcs6490-mitysom-devkit:~# tinymix set 'Lineout Playback Volume' 31 31
These should ensure Lineout Playback is enabled and that the Playback Volume is not 0 (could set to any value in range of 0 to 31, but max volume is recommended to ensure output is heard). LINEOUT should always receive output as it's not MUXed with anything (per the diagram here). Now, running
gst-launch-1.0 audiotestsrc ! audioconvert ! audio/x-raw,channels=2 ! pulsesink will play a test tone to verify LINEOUT functionality. The commands used to playback to headphones in the above instructions can also be used to play audio files over LINEOUT.
Additional Notes and Troubleshooting¶
If any issues are encountered, referencing the CODEC and controls section, tracing through the system block diagram, and then checking your tinymix settings with tinymix contents could help spot any issues in your configuration. Also ensuring that settings from other configurations (such as going from trying to record input from MIC_IN to LINE_IN as those signals are MUXed) aren't interfering with the current configuration could alleviate issues.
SGTL5000 CODEC and Controls¶
The linux SGTL5000 codec driver includes several controls for the routing and data within the SGTL5000 chip (between the jacks and the I2S0 bus). A top level block diagram of the SGTL5000 is shown below:
From the picture, the 3 signal audio jack is connected to the Headphone outputs (stereo) and microphone inputs (mono) for use with a standard analog headset with microphone. The line in and line out signals are routed to the expansion header J23. By default, playback commands will be routed to the HeadPhone interface and capture commands will capture from the microphone input.
The available controls of the output SGTL5000 codec can be listed (along with their settings) with the tinymix command as shown below.
root@qcs6490-mitysom-devkit:~# tinymix contents Number of controls: 28 ctl type num name device value 0 INT 2 PCM Playback Volume 0144, 144 (range 0->192) 1 INT 2 Capture Volume 012, 12 (range 0->15) 2 BOOL 1 Capture Attenuate Switch (-6dB) 0Off 3 BOOL 1 Capture ZC Switch 0On 4 BOOL 1 Capture Switch 0On 5 INT 2 Headphone Playback Volume 0110, 110 (range 0->127) 6 BOOL 1 Headphone Playback Switch 0On 7 BOOL 1 Headphone Playback ZC Switch 0On 8 INT 1 Mic Volume 00 (range 0->3) 9 INT 2 Lineout Playback Volume 016, 16 (range 0->31) 10 BOOL 1 Lineout Playback Switch 0Off 11 INT 1 DAP Main channel 032768 (range 0->65535) 12 INT 1 DAP Mix channel 00 (range 0->65535) 13 BOOL 1 AVC Switch 0Off 14 BOOL 1 AVC Hard Limiter Switch 0Off 15 INT 1 AVC Max Gain Volume 01 (range 0->2) 16 INT 1 AVC Integrator Response 01 (range 0->3) 17 INT 1 AVC Threshold Volume 012 (range 0->96) 18 INT 1 BASS 0 047 (range 0->95) 19 INT 1 BASS 1 047 (range 0->95) 20 INT 1 BASS 2 047 (range 0->95) 21 INT 1 BASS 3 047 (range 0->95) 22 INT 1 BASS 4 047 (range 0->95) 23 ENUM 1 Capture Mux 0> MIC_IN, LINE_IN, 24 ENUM 1 Headphone Mux 0> DAC, LINE_IN, 25 ENUM 1 Digital Input Mux 0ADC, > I2S, Rsvrd, DAP, 26 ENUM 1 DAP Mux 0> ADC, I2S, 27 ENUM 1 DAP MIX Mux 0> ADC, I2S,
Note: the tinymix command does not put a space between the device number (always "0" in this case) and the settings.
In this case, the 'PCM Playback Volume' represents the digital DAC Volume Control, the 'Headphone Playback Volume' represents the analog gain on the Headphone volume control, the 'Capture Volume' represents the Analog Gain block entering the ADC, and the 'Mic Volume' represents the MIC GAIN block.
The tinymix tool can be used to adjust these settings. For example, to increase the analog output volume of the headphones channel:
tinymix set 'Headphone Playback Volume' 120 120
This will set both the left and right channel gain setting to be "120".
There are other features of the codec not discussed here (including routing the audio to the line-in and line-out and leveraging the CODEC audio signal processing blocks) but can be leveraged in a similar manner.
Go to top
