Partial Convolution

Collaboration diagram for Partial Convolution:

Functions
int32_t	tpt_conv_partial_f32 (f32_t *aOutData, f32_t *aInDataA, uint32_t aLen1, f32_t *aInDataB, uint32_t aLen2, uint32_t aStartindex, uint32_t aCount)
	Convolution Partial of the floating-point vectors. More...
int32_t	tpt_conv_partial_q7 (q7_t *aOutData, q7_t *aInDataA, uint32_t aLen1, q7_t *aInDataB, uint32_t aLen2, uint32_t aStartindex, uint32_t aCount)
	Convolution Partial of the q7 vectors. More...
int32_t	tpt_conv_partial_q15 (q15_t *aOutData, q15_t *aInDataA, uint32_t aLen1, q15_t *aInDataB, uint32_t aLen2, uint32_t aStartindex, uint32_t aCount)
	Convolution Partial of the q15 vectors. More...
int32_t	tpt_conv_partial_q31 (q31_t *aOutData, q31_t *aInDataA, uint32_t aLen1, q31_t *aInDataB, uint32_t aLen2, uint32_t aStartindex, uint32_t aCount)
	Convolution Partial of q31 vectors. More...

Detailed Description

With a starting index and a aCount specified, partial convolution functions do the convolution partially for two signals. The partial convolution result will be generated to the destination vector in the range [start_index, … , start_index + aCount - 1 ]. Thus, if the partial convolution result is not within the range [0, … , aLen1 + aLen2 - 2], a return value of -1 will be given for the argument error

TPT DSP library supports distinct partial convolution functions for the following data types: single-precision floating-point, Q31, Q15 and Q7. These functions are introduced in the subsections below.

Function Documentation

tpt_conv_partial_f32()

int32_t tpt_conv_partial_f32	(	f32_t *	aOutData,
		f32_t *	aInDataA,
		uint32_t	aLen1,
		f32_t *	aInDataB,
		uint32_t	aLen2,
		uint32_t	aStartindex,
		uint32_t	aCount
	)

Convolution Partial of the floating-point vectors.

Parameters

[out]	aOutData	points to the output vector where the length is aLen1 + aLen2 1.
[in]	aInDataA	points to the first input vector.
[in]	aLen1	length of the first input vector.
[in]	aInDataB	points to the second input vector.
[in]	aLen2	length of the second input vector.
[in]	aStartindex	is the first output sample to start with.
[in]	aCount	is the number of output points to be computed.

Returns

int32_t

tpt_conv_partial_q15()

int32_t tpt_conv_partial_q15	(	q15_t *	aOutData,
		q15_t *	aInDataA,
		uint32_t	aLen1,
		q15_t *	aInDataB,
		uint32_t	aLen2,
		uint32_t	aStartindex,
		uint32_t	aCount
	)

Convolution Partial of the q15 vectors.

Parameters

[out]	aOutData	points to the output vector where the length is aLen1 + aLen2 1.
[in]	aInDataA	points to the first input vector.
[in]	aLen1	length of the first input vector.
[in]	aInDataB	points to the second input vector.
[in]	aLen2	length of the second input vector.
[in]	aStartindex	is the first output sample to start with.
[in]	aCount	is the number of output points to be computed.

Returns

int32_t

tpt_conv_partial_q31()

int32_t tpt_conv_partial_q31	(	q31_t *	aOutData,
		q31_t *	aInDataA,
		uint32_t	aLen1,
		q31_t *	aInDataB,
		uint32_t	aLen2,
		uint32_t	aStartindex,
		uint32_t	aCount
	)

Convolution Partial of q31 vectors.

Parameters

[out]	aOutData	points to the output vector where the length is aLen1 + aLen2 1.
[in]	aInDataA	points to the first input vector.
[in]	aLen1	length of the first input vector.
[in]	aInDataB	points to the second input vector.
[in]	aLen2	length of the second input vector.
[in]	aStartindex	is the first output sample to start with.
[in]	aCount	is the number of output points to be computed.

Returns

int32_t

tpt_conv_partial_q7()

int32_t tpt_conv_partial_q7	(	q7_t *	aOutData,
		q7_t *	aInDataA,
		uint32_t	aLen1,
		q7_t *	aInDataB,
		uint32_t	aLen2,
		uint32_t	aStartindex,
		uint32_t	aCount
	)

Convolution Partial of the q7 vectors.

Parameters

[out]	aOutData	points to the output vector where the length is aLen1 + aLen2 1.
[in]	aInDataA	points to the first input vector.
[in]	aLen1	length of the first input vector.
[in]	aInDataB	points to the second input vector.
[in]	aLen2	length of the second input vector.
[in]	aStartindex	is the first output sample to start with.
[in]	aCount	is the number of output points to be computed.

Returns

int32_t