STSoundDetect

Overview

ST Sound Detector module

Discussion

SoundDetector has the function of discriminating between stationary noise and sound. To do so, it has a noise estimator which tracks the slow changes of the noise. Two configurable thresholds (to allow hysteresis) are then used to decide the state flag value (active or inactive). A minimum noise power level can be adjusted in order to lower-limit the noise estimator. At last, an hangover period from active to inactive state is configurable, in order to avoid state toggling. This module intentionally uses simple estimator algorithms in order to provide a low MIPS solution. The current implementation internally uses 24 bits input format, mono signal only.

The minimum input frame is 96 and the maximum input frame is 480 stereo samples.

Type Definition

CODE

typedef struct _ModuleSTSoundDetect
{
    ModuleInstanceDescriptor instance;            // Common Audio Weaver module instance structure
    INT32 nAPIResult;                             // Return status for last call into the ST API
    INT32 LearningFrameNumber;                    // Learning Frame Number
    INT32 enable;                                 // Block enable
    INT32 ratio_thr1_dB;                          // Inactive to active threshold in dB
    INT32 ratio_thr2_dB;                          // Active to Inactive threshold in dB
    INT32 noise_pwr_min_dB;                       // Minimum power noise value dBFS
    INT32 hangover_nb_frame;                      // Number of frames of hangover
    INT32 output_state;                           //
    INT32 output_nrj;                             //
    INT32 output_noise_pwr;                       //
    INT32 output_thr1;                            //
    INT32 output_thr2;                            //
    void * pGlobals;                              // Points to the global variables structure
} ModuleSTSoundDetectClass;

Variables

Properties

Name	Type	Usage	isHidden	Default value	Range	Units
nAPIResult	int	state	0	1	Unrestricted
LearningFrameNumber	int	const	0	4	Unrestricted
enable	int	parameter	0	1	0:1:1
ratio_thr1_dB	int	parameter	0	9	1:1:40	dB
ratio_thr2_dB	int	parameter	0	7	1:1:40	dB
noise_pwr_min_dB	int	parameter	0	-60	-87:1:0	dB
hangover_nb_frame	int	parameter	0	20	1:1:40
output_state	int	state	0	0	Unrestricted
output_nrj	int	state	0	0	Unrestricted
output_noise_pwr	int	state	0	0	Unrestricted
output_thr1	int	state	0	0	Unrestricted
output_thr2	int	state	0	0	Unrestricted
pGlobals	void *	const	1		Unrestricted

Pins

Input Pins

Name: in

Description: audio input

Data type: fract32

Channel range: 1

Block size range: 96:480

Sample rate range: 8000, 16000, 44100, 48000

Complex support: Real

Output Pins

Name: out

Description: audio output

Data type: fract32

Scratch Pins

Channel count: 1

Block size: 96

Sample rate: 8000

MATLAB Usage

File Name: st_sound_detect_module.m

CODE

 M = st_sound_detect_module(NAME, LEARNING_FRAME_NUMBER)
 Audio Weaver wrapper for the ST Sound Detect module.  Arguments:
   NAME - name of the module.
   LEARNING_FRAME_NUMBER - 2 to 20 frames
 Copyright (c) 2017 DSP Concepts, Inc.  All Rights Reserved.
 Author:  Chris Perry