Audio Processing Application

This chapter describes the GK7206 audio processing application example — sample_audio. It contains 9 independent audio programs covering the complete audio function chain: audio acquisition, playback, encoding/decoding, recording, and double talk.

The application source code is located in the SDK directory sample/audio/. It provides a comprehensive demonstration of audio functions and serves as a reference template for developing audio-related applications.

1 Application Overview

1.1 Features

This application contains 9 independent audio programs:

  1. sample_audio_ai - audio acquisition (AI, Audio Input)
  2. sample_audio_ao - audio output (AO, Audio Output)
  3. sample_audio_encoder - audio encoding (PCM → G711/AAC)
  4. sample_audio_decoder - audio decoding (G711/AAC → PCM)
  5. sample_audio_record_v1 - recording (version 1)
  6. sample_audio_record_v2 - recording (version 2)
  7. sample_audio_record_v3 - recording (version 3)
  8. sample_audio_record_detect - recording + detection
  9. sample_audio_doubletalk_v1/v2 - double talk (AEC echo cancellation)

1.2 Technical Specifications

ParameterValue
Sample rates8000/16000/32000/44100/48000 Hz
Channels1 (mono) / 2 (stereo)
Sample bit depth16 bit
Encoding formatsG711A/U, AAC
Audio quality enhancementVQE (Voice Quality Enhancement)
Echo cancellationAEC (Acoustic Echo Cancellation)
Input/output filesPCM files (under /sd_card)

1.3 Directory Structure

sample/audio/
├── Makefile                          # Build script
├── sample_audio_ai                    # Audio acquisition program
├── sample_audio_ai.c                  # Audio acquisition source
├── sample_audio_ao                    # Audio output program
├── sample_audio_ao.c                  # Audio output source
├── sample_audio_encoder               # Audio encoding program
├── sample_audio_encoder.c            # Audio encoding source
├── sample_audio_decoder              # Audio decoding program
├── sample_audio_decoder.c            # Audio decoding source
├── sample_audio_record_v1            # Recording program v1
├── sample_audio_record_v2            # Recording program v2
├── sample_audio_record_v3            # Recording program v3
├── sample_audio_record_detect        # Recording + detection program
├── sample_audio_doubletalk_v1        # Double talk program v1
├── sample_audio_doubletalk_v2        # Double talk program v2
└── record_board/                     # Board-level recording
    ├── sample_record_board_detect
    └── sample_record_board_record

2 Build and Deployment

2.1 Prerequisites

Before building this application, make sure the following preparations have been completed:

  1. SDK environment is set up: follow SDK Compilation to set up the cross-compilation toolchain and SDK configuration
  2. Audio hardware preparation:
    • Acquisition tests: a microphone must be connected
    • Playback tests: a speaker must be connected

2.2 Build the Application

# Enter the sample directory
cd <SDK_PATH>/sample

# Build the audio sample
make -C audio clean && make -C audio

Or build from the top level:

make -C sample all

2.3 Deploy to the Board

# Transfer to the development board via SCP
scp sample/audio/sample_audio_* root@<board IP>:/sd_card/

# Or download via TFTP
tftp -g -r sample_audio_ai <board IP>

Info

Audio files are usually large, so it is recommended to run them from the SD card.

3 Audio Acquisition (AI)

3.1 Function Description

sample_audio_ai demonstrates audio acquisition, saving the audio data captured from the microphone into a PCM file.

3.2 Command Format

./sample_audio_ai [dev] [send_object] [binder_mode] [samplerate] [usr_samplerate] [channels] [workmode] [outfile] [vqe_type]

Parameters:

ParameterDescriptionValues
devAudio device number0/1
send_objectSend target0=AO, 1=encoder, 2=both
binder_modeBinder mode0/1
samplerateSample rate8000/16000/32000/44100/48000
usr_samplerateUser sample rate8000/16000/32000/44100/48000
channelsNumber of channels1/2
workmodeWorking mode0/1
outfileOutput file path/sd_card/ai.pcm
vqe_typeAudio quality enhancement type0/1

3.3 Typical Usage

# 8 kHz mono acquisition, saved to /sd_card/ai.pcm
./sample_audio_ai 0 0 0 8000 8000 1 0 /sd_card/ai.pcm 0

# 16 kHz mono acquisition with VQE enabled
./sample_audio_ai 0 0 0 16000 16000 1 0 /sd_card/ai.pcm 1

3.4 Expected Behavior

The program prints the following information when running:

audio ai sample running
ai dev: 0
samplerate: 8000
channels: 1
...
recording to /sd_card/ai.pcm

During acquisition it keeps printing frame counts; press Ctrl+C to stop.

4 Audio Output (AO)

4.1 Function Description

sample_audio_ao demonstrates audio playback, reading audio data from a PCM file and playing it through the speaker.

4.2 Command Format

./sample_audio_ao [infile] [dev] [dev_rate] [dev_chn] [frm_rate] [frm_chn] [mode] [pcm_sample] [vqe]

Parameters:

ParameterDescriptionValues
infileInput PCM file path/sd_card/in.pcm
devAudio device number0/1
dev_rateDevice sample rate8000/16000/32000/44100/48000
dev_chnDevice channel count1/2
frm_rateFrame sample rate8000/16000/32000/44100/48000
frm_chnFrame channel count1/2
modePlayback mode0/1
pcm_samplePCM sample bit depth16
vqeAudio quality enhancement0/1

4.3 Typical Usage

# 8 kHz mono playback
./sample_audio_ao /sd_card/in.pcm 0 8000 1 30 0 0 16 0

# 16 kHz mono playback with VQE enabled
./sample_audio_ao /sd_card/in.pcm 0 16000 1 30 0 0 16 1

4.4 Expected Behavior

The program prints the following information when running:

audio ao sample running
ao dev: 0
samplerate: 8000
channels: 1
...
playing /sd_card/in.pcm

During playback it keeps printing frame counts, and exits automatically when playback finishes.

5 Audio Encoding/Decoding

5.1 Audio Encoding (sample_audio_encoder)

Function: encodes a PCM file into G711/AAC format.

Command format:

./sample_audio_encoder [infile_samplerate] [infile_channels] [infile] [outfile]

Typical usage:

# Encode to G711
./sample_audio_encoder 8000 1 /sd_card/in.pcm /sd_card/out.g711

# Encode to AAC
./sample_audio_encoder 8000 1 /sd_card/in.pcm /sd_card/out.aac

5.2 Audio Decoding (sample_audio_decoder)

Function: decodes a G711/AAC file into PCM format.

Command format:

./sample_audio_decoder [mode] [infile_samplerate] [infile]

Typical usage:

# Decode G711
./sample_audio_decoder 0 8000 /sd_card/in.g711

# Decode AAC
./sample_audio_decoder 1 8000 /sd_card/in.aac

6 Recording Programs

6.1 Version Differences

There are three versions of the recording program; the main differences are in parameters and features:

VersionCharacteristics
sample_audio_record_v1Basic recording
sample_audio_record_v2Supports more parameters
sample_audio_record_v3Supports more audio quality enhancement options

6.2 Command Format

# v1
./sample_audio_record_v1 [samplerate] [usr_samplerate] [binder_mode] [codec_input_vol] [outfile] [vqe_type]

# v2
./sample_audio_record_v2 [samplerate] [usr_samplerate] [binder_mode] [codec_input_vol] [outfile] [vqe_type] [aec_mode]

# v3
./sample_audio_record_v3 [samplerate] [usr_samplerate] [binder_mode] [codec_input_vol] [outfile] [vqe_type] [aec_mode] [agc_mode]

6.3 Typical Usage

# Basic recording
./sample_audio_record_v1 8000 8000 0 0 /sd_card/rec.pcm 0

# Recording with VQE and AEC enabled
./sample_audio_record_v2 16000 16000 0 0 /sd_card/rec.pcm 1 1

7 Double Talk Program

7.1 Function Description

sample_audio_doubletalk_v1/v2 demonstrates the double talk feature — performing audio acquisition and playback simultaneously with AEC (echo cancellation) enabled.

7.2 Command Format

./sample_audio_doubletalk_v1 [samplerate] [usr_samplerate] [channels] [binder_mode] [outfile] [infile] [infile_samplerate]

7.3 Typical Usage

# Double talk test
./sample_audio_doubletalk_v1 8000 8000 1 0 /sd_card/out.pcm /sd_card/in.pcm 8000

7.4 Expected Behavior

The program performs acquisition and playback simultaneously; the AEC algorithm removes the echo to ensure bidirectional call quality.

8 Audio Quality Enhancement (VQE)

8.1 VQE Functions

VQE (Voice Quality Enhancement) provides the following functions:

FunctionDescription
ANRAdaptive noise reduction
AGCAutomatic gain control
AECAcoustic echo cancellation
EQEqualizer

8.2 Enabling VQE

Set the vqe_type argument to 1 in the command to enable VQE:

./sample_audio_ai 0 0 0 8000 8000 1 0 /sd_card/ai.pcm 1

9 Key Programming Points

9.1 AI Acquisition Flow

// Step 1: Initialize the AI module
xmedia_ai_set_dev_attr(dev, &attr);
xmedia_ai_enable(dev);

// Step 2: Get an audio frame
xmedia_ai_get_frame(dev, &frame, timeout);

// Step 3: Process the audio frame
// ... write to file or send to the encoder ...

// Step 4: Release the audio frame
xmedia_ai_release_frame(dev, &frame);

// Step 5: Disable the AI module
xmedia_ai_disable(dev);

9.2 AO Output Flow

// Step 1: Initialize the AO module
xmedia_ao_set_dev_attr(dev, &attr);
xmedia_ao_enable(dev);

// Step 2: Send an audio frame
xmedia_ao_send_frame(dev, &frame, timeout);

// Step 3: Disable the AO module
xmedia_ao_disable(dev);

9.3 Audio Encoding/Decoding Flow

// Encoding
xmedia_acodec_create_encoder(&chn, &type);
xmedia_acodec_send_frame(chn, &pcm_frame);
xmedia_acodec_get_stream(chn, &encoded_stream);

// Decoding
xmedia_acodec_create_decoder(&chn, &type);
xmedia_acodec_send_stream(chn, &encoded_stream);
xmedia_acodec_get_frame(chn, &pcm_frame);

10 Troubleshooting

ProblemPossible causeSolution
No sound outputSpeaker not connected or volume too lowCheck the hardware connection, adjust volume
No data acquiredMicrophone not connectedCheck the hardware connection
Distorted soundSample rate mismatchCheck the sample rate settings
Severe echoAEC not enabledEnable VQE and AEC
Encoding/decoding failureUnsupported encoding formatCheck that the encoding format is supported

11 References