aic.h File Reference

Header file for the ai-coustics speech enhancement SDK. More...

#include <cstddef>
#include <cstdint>

Go to the source code of this file.

Typedefs
using	aic::LogCallback = void(*)(const char *, LogLevel)

Enumerations
enum class	aic::LogLevel : uint32_t {   Error = 1 , Warn , Debug , Info ,   Trace }

Functions
uint32_t	aic::aic_log_init (LogCallback log_callback)
	Initializes the SDK logger.
AicModel *	aic::aic_new_model_s ()
	Creates a new AicModel, using the machine learning model ModelS.
AicModel *	aic::aic_new_model_m ()
	Creates a new AicModel, using the machine learning model ModelM.
AicModel *	aic::aic_new_model_l ()
	Creates a new AicModel, using the machine learning model ModelL.
AicModel *	aic::aic_new_model_z ()
	Creates a new AicModel, using the machine learning model ModelZ.
uint32_t	aic::aic_init (AicModel *model, size_t num_channels, size_t sample_rate, size_t num_frames)
	Initializes the audio processing settings for the AicModel.
uint32_t	aic::aic_process_interleaved (AicModel *model, float *buffer, size_t num_channels, size_t num_frames)
	Processes a buffer of interleaved audio data using the AicModel.
uint32_t	aic::aic_process_deinterleaved (AicModel *model, float *const *buffer, size_t num_channels, size_t num_frames)
	Processes a multi-channel buffer of audio data using the AicModel.
uint32_t	aic::aic_reset (AicModel *model)
	Processes a multi-channel buffer of audio data using the AicModel.
uint32_t	aic::aic_set_enhancement_strength (AicModel *model, float enhancement_strength)
	Sets the enhancement strength for the AicModel.
uint32_t	aic::aic_get_enhancement_strength (AicModel *model, float *enhancement_strength)
	Gets the current enhancement strength for the AicModel.
uint32_t	aic::aic_set_voice_gain (AicModel *model, float voice_gain)
	Sets the voice gain parameter of the AicModel. This is the gain that is added to the extracted voice before the mixback to the original signal is happening.
uint32_t	aic::aic_get_voice_gain (AicModel *model, float *voice_gain)
	Gets the current voice gain parameter of the AicModel. This is the gain that is added to the extracted voice before the mixback to the original signal is happening.
uint32_t	aic::aic_get_optimal_num_frames (AicModel *model, size_t *num_frames)
	Gets the optimal number of frames for the AicModel. This is the native number of frames of the model, that causes the lowest latency.
uint32_t	aic::aic_get_optimal_sample_rate (AicModel *model, size_t *sample_rate)
	Gets the optimal sample rate for the AicModel. This is the native sample rate of the model, that causes the lowest latency.
uint32_t	aic::aic_get_latency (AicModel *model, size_t *latency)
	Gets the current latency of the full process in samples.
void	aic::aic_free (AicModel *model)
	Frees the memory used by the AicModel.

Detailed Description

Header file for the ai-coustics speech enhancement SDK.

This file contains the definitions and declarations for the ai-coustics speech enhancement SDK, including initialization, processing, and configuration functions. The ai-coustics SDK provides advanced machine learning models for speech enhancement, that can be used in audio streaming contexts.

Copyright

Copyright (C) ai-coustics GmbH - All Rights Reserved

Unauthorized copying, distribution, or modification of this file, via any medium, is strictly prohibited.

For inquiries, please contact: info@.nosp@m.ai-c.nosp@m.ousti.nosp@m.cs.c.nosp@m.om

Typedef Documentation

LogCallback

using aic::LogCallback = typedef void(*)(const char*, LogLevel)

The SDK will call this function to allow the user to receive logs. The first parameter is a pointer to the log message, and the second parameter is the level of that log.

Enumeration Type Documentation

LogLevel

enum class aic::LogLevel : uint32_t

strong

Levels for the log callback.

Function Documentation

aic_free()

void aic::aic_free

(

AicModel *

model

)

Frees the memory used by the AicModel.

Parameters

model

A pointer to the AicModel.

aic_get_enhancement_strength()

uint32_t aic::aic_get_enhancement_strength	(	AicModel *	model,
		float *	enhancement_strength
	)

Gets the current enhancement strength for the AicModel.

Parameters

model	A pointer to the AicModel.
enhancement_strength	A pointer to the variable that will store the current strength.

Returns

AIC_PASS on success, or an error code on failure.

aic_get_latency()

uint32_t aic::aic_get_latency	(	AicModel *	model,
		size_t *	latency
	)

Gets the current latency of the full process in samples.

Parameters

model	A pointer to the AicModel.
latency	A pointer to the variable that will store the latency value.

Returns

AIC_PASS on success, or an error code on failure.

aic_get_optimal_num_frames()

uint32_t aic::aic_get_optimal_num_frames	(	AicModel *	model,
		size_t *	num_frames
	)

Gets the optimal number of frames for the AicModel. This is the native number of frames of the model, that causes the lowest latency.

Parameters

model	A pointer to the AicModel.
num_frames	A pointer to the variable that will store the optimal number of frames.

Returns

AIC_PASS on success, or an error code on failure.

aic_get_optimal_sample_rate()

uint32_t aic::aic_get_optimal_sample_rate	(	AicModel *	model,
		size_t *	sample_rate
	)

Gets the optimal sample rate for the AicModel. This is the native sample rate of the model, that causes the lowest latency.

Parameters

model	A pointer to the AicModel.
sample_rate	A pointer to the variable that will store the optimal sample rate.

Returns

AIC_PASS on success, or an error code on failure.

aic_get_voice_gain()

uint32_t aic::aic_get_voice_gain	(	AicModel *	model,
		float *	voice_gain
	)

Gets the current voice gain parameter of the AicModel. This is the gain that is added to the extracted voice before the mixback to the original signal is happening.

Parameters

model	A pointer to the AicModel.
voice_gain	A pointer to the variable that will store the voice gain parameter.

Returns

AIC_PASS on success, or an error code on failure.

aic_init()

uint32_t aic::aic_init	(	AicModel *	model,
		size_t	num_channels,
		size_t	sample_rate,
		size_t	num_frames
	)

Initializes the audio processing settings for the AicModel.

Parameters

model	A pointer to the AicModel.
num_channels	The number of audio channels being processed.
sample_rate	The sample rate of the audio being processed.
num_frames	The number of audio frames processed per callback.

Returns

AIC_PASS on success, or an error code on failure.

Note

This function has to be called before the process function.

aic_log_init()

uint32_t aic::aic_log_init

(

LogCallback

log_callback

)

Initializes the SDK logger.

Parameters

log_callback

A callback function that the SDK will use to send log messages

Returns

AIC_PASS on success, or AIC_FAIL if the SDK logging system was already initialized.

Note

This function must only be called once.

aic_new_model_l()

AicModel * aic::aic_new_model_l

(

)

Creates a new AicModel, using the machine learning model ModelL.

Returns

A pointer to the newly created AicModel.

aic_new_model_m()

AicModel * aic::aic_new_model_m

(

)

Creates a new AicModel, using the machine learning model ModelM.

Returns

A pointer to the newly created AicModel.

aic_new_model_s()

AicModel * aic::aic_new_model_s

(

)

Creates a new AicModel, using the machine learning model ModelS.

Returns

A pointer to the newly created AicModel.

aic_new_model_z()

AicModel * aic::aic_new_model_z

(

)

Creates a new AicModel, using the machine learning model ModelZ.

Returns

A pointer to the newly created AicModel.

aic_process_deinterleaved()

uint32_t aic::aic_process_deinterleaved	(	AicModel *	model,
		float *const *	buffer,
		size_t	num_channels,
		size_t	num_frames
	)

Processes a multi-channel buffer of audio data using the AicModel.

Parameters

model	A pointer to the AicModel.
buffer	A pointer to the audio data.
num_channels	The number of channels.
num_frames	The number of audio frames in the buffer.

Returns

AIC_PASS on success, or an error code on failure.

aic_process_interleaved()

uint32_t aic::aic_process_interleaved	(	AicModel *	model,
		float *	buffer,
		size_t	num_channels,
		size_t	num_frames
	)

Processes a buffer of interleaved audio data using the AicModel.

Parameters

model	A pointer to the AicModel.
buffer	A pointer to the audio data.
num_channels	The number of channels.
num_frames	The number of audio frames in the buffer.

Returns

AIC_PASS on success, or an error code on failure.

Note

The buffer has to be num_channels * num_frames long.

aic_reset()

uint32_t aic::aic_reset

(

AicModel *

model

)

Processes a multi-channel buffer of audio data using the AicModel.

Parameters

model	A pointer to the AicModel.
buffer	A pointer to the audio data.
num_channels	The number of channels.
num_frames	The number of audio frames in the buffer.

Returns

AIC_PASS on success, or an error code on failure.

aic_set_enhancement_strength()

uint32_t aic::aic_set_enhancement_strength	(	AicModel *	model,
		float	enhancement_strength
	)

Sets the enhancement strength for the AicModel.

Parameters

model	A pointer to the AicModel.
enhancement_strength	Value between 0.0 and 1.0, where 0.0 is equal to a bypass and 1.0 is the maximum enhancement.

Returns

AIC_PASS on success, or an error code on failure.

aic_set_voice_gain()

uint32_t aic::aic_set_voice_gain	(	AicModel *	model,
		float	voice_gain
	)

Sets the voice gain parameter of the AicModel. This is the gain that is added to the extracted voice before the mixback to the original signal is happening.

Parameters

model	A pointer to the AicModel.
voice_gain	The voice gain parameter to set.

Returns

AIC_PASS on success, or an error code on failure.