API Basics

Immersitech ClearVoice is a C/C++ library enabling advanced speech processing utilities. It operates on 10ms buffers of single-channel, floating-point audio data (values from -1 to 1), and has configurable input/output sample rate options. Usage requires a license file. (A trial license is included in your distribution of the library. If it has expired or is unavailable for any reason, please contact us and we will be happy to assist.) This page will outline the steps needed to process audio through ClearVoice.

Step-by-step setup in C

To begin using ClearVoice functions in your project, first include the appropriate header file.

#include immersitech_clearvoice.h

Next, create an imm_cv_config structure. The easiest way is to just use the default configuration, and edit any values relevent to your project. Here, we'll just set ANC Mix to 90%.

imm_cv_config config = imm_cv_get_default_config();
config.anc_mix = 90;

We can now use this configuration structure to initialize ClearVoice, and return a handle to our initialized library. We also require the path to a valid license file, and a pointer to an imm_error_code to return any errors.

const char* license_file = "./Immersitech_Engineering_sound_manager_license_key.dat";
imm_error_code error_code;
imm_cv_handle handle = imm_cv_init_from_file(license_file, config, &error_code);
 
if (error_code != IMM_ERROR_NONE) {
    printf("An error occured!\n");
}

With an initialized handle, all of the other functions are now available. For example, we can now change ANC Mix back to 100% using the imm_cv_set_config function.

error_code = imm_cv_set_config(handle, IMM_CV_MUT_CONF_ANC_MIX, 100);
if (error_code != IMM_ERROR_NONE) {
    printf("An error occured!\n");
}

As you can see above, many configuration settings from the structure used to initialize the library are mutable after initialization, and can even be changed in the audio thread. The full list of mutable configuration settings are in the enum imm_cv_mutable_config.

Most importantly, we can now process audio data.

float* input_audio[480] = {0.0f};
float* output_audio[480] = {0.0f};
imm_cv_output_metadata metadata;
error_code = imm_cv_process(handle, input_audio, output_audio, &metadata);

All together, a simple ClearVoice test executable would look like this:

#include <stdio.h>
#include "immersitech_clearvoice.h"
 
int main() {
 
    imm_cv_config config = imm_cv_get_default_config();
    config.anc_mix = 90;
 
    const char* license_file = "./Immersitech_Engineering_sound_manager_license_key.dat";
    imm_error_code error_code;
    imm_cv_handle handle = imm_cv_init_from_file(license_file, config, &error_code);
 
    if (error_code != IMM_ERROR_NONE) {
        printf("An error occured!\n");
    }
 
    error_code = imm_cv_set_config(handle, IMM_CV_MUT_CONF_ANC_MIX, 100);
    if (error_code != IMM_ERROR_NONE) {
        printf("An error occured!\n");
    }
 
    float input_audio[480] = {0.0f};
    float output_audio[480] = {0.0f};
    imm_cv_output_metadata metadata;
    error_code = imm_cv_process(handle, input_audio, output_audio, &metadata);
 
    imm_cv_destroy(handle);
 
    return 0;
}