Class convolve_filter¶

`class` convolve_filter<T> dft ¶

template <typename T>
class convolve_filter : public filter<T> { … }

Streaming convolution filter using the overlap-add (block convolution) method.

Template parameters
T	Sample type: `float` , `double` , `complex<float>` or `complex<double>` .

`constructor` convolve_filter<T>(size_t, size_t) ¶

explicit convolve_filter(size_t size, size_t block_size = 1024)

Constructs a filter with a kernel of the given length (initially zero).

Parameters
size	Length (in samples) of the filter kernel.
block_size	Processing block size in samples; rounded up to the next power of two. Larger blocks reduce per-sample overhead at the cost of more memory and latency.

Fully qualified name: kfr::convolve_filter<T>::convolve_filter<T>(size_t, size_t)

Declared at dft/convolution.hpp:177

`constructor` convolve_filter<T>(const univector_ref<const T> &, size_t) ¶

explicit convolve_filter(const univector_ref<const T>& data, size_t block_size = 1024)

Constructs a filter and initializes its kernel from data .

Parameters
data	The filter kernel (impulse response).
block_size	Processing block size in samples; rounded up to the next power of two.

Fully qualified name: kfr::convolve_filter<T>::convolve_filter<T>(const univector_ref<const T> &, size_t)

Declared at dft/convolution.hpp:184

`function` set_data(const univector_ref<const T> &) ¶

void set_data(const univector_ref<const T>& data)

Sets (replaces) the filter kernel.

Re-transforms the kernel into frequency-domain blocks and resets the internal state. The number of history segments is resized to match the new kernel length.

Parameters
data	The new filter kernel (impulse response).

Fully qualified name: kfr::convolve_filter<T>::set_data(const univector_ref<const T> &)

Declared at dft/convolution.hpp:193

`function` reset() ¶

void reset() final

Resets the filter state (clears input history and overlap buffers).

Fully qualified name: kfr::convolve_filter<T>::reset()

Declared at dft/convolution.hpp:195

`function` input_block_size() ¶

size_t input_block_size() const

Returns the processing block size.

For optimal efficiency, drive the filter with input lengths that are multiples of this value.

Returns
	The block size in samples.

Fully qualified name: kfr::convolve_filter<T>::input_block_size()

Declared at dft/convolution.hpp:203

`function` process_expression(T *, const expression_handle<T> &, size_t) ¶

void process_expression(T* dest, const expression_handle<T>& src, size_t size) final

Fully qualified name: kfr::convolve_filter<T>::process_expression(T *, const expression_handle<T> &, size_t)

Declared at dft/convolution.hpp:206

`function` process_buffer(T , const T , size_t) ¶

void process_buffer(T* output, const T* input, size_t size) final

Fully qualified name: kfr::convolve_filter<T>::process_buffer(T *, const T *, size_t)

Declared at dft/convolution.hpp:211

`typedef` ST ¶

using ST                       = subtype<T>

Fully qualified name: kfr::convolve_filter<T>::ST

Declared at dft/convolution.hpp:213

`variable` real_fft ¶

constexpr static bool real_fft = !std::is_same_v<T, complex<ST>>

Fully qualified name: kfr::convolve_filter<T>::real_fft

Declared at dft/convolution.hpp:214

`typedef` plan_t ¶

using plan_t                   = internal_generic::dft_conv_plan<T>

Fully qualified name: kfr::convolve_filter<T>::plan_t

Declared at dft/convolution.hpp:215

`variable` data_size ¶

size_t data_size

Fully qualified name: kfr::convolve_filter<T>::data_size

Declared at dft/convolution.hpp:218

`variable` block_size ¶

const size_t block_size

Fully qualified name: kfr::convolve_filter<T>::block_size

Declared at dft/convolution.hpp:220

`variable` fft ¶

const plan_t fft

Fully qualified name: kfr::convolve_filter<T>::fft

Declared at dft/convolution.hpp:222

`variable` temp ¶

univector<u8> temp

Fully qualified name: kfr::convolve_filter<T>::temp

Declared at dft/convolution.hpp:224

`variable` segments ¶

std::vector<univector<complex<ST>>> segments

Fully qualified name: kfr::convolve_filter<T>::segments

Declared at dft/convolution.hpp:226

`variable` position ¶

size_t position

Fully qualified name: kfr::convolve_filter<T>::position

Declared at dft/convolution.hpp:228

`variable` ir_segments ¶

std::vector<univector<complex<ST>>> ir_segments

Fully qualified name: kfr::convolve_filter<T>::ir_segments

Declared at dft/convolution.hpp:230

`variable` saved_input ¶

univector<T> saved_input

Fully qualified name: kfr::convolve_filter<T>::saved_input

Declared at dft/convolution.hpp:232

`variable` input_position ¶

size_t input_position

Fully qualified name: kfr::convolve_filter<T>::input_position

Declared at dft/convolution.hpp:234

`variable` premul ¶

univector<complex<ST>> premul

Fully qualified name: kfr::convolve_filter<T>::premul

Declared at dft/convolution.hpp:236

`variable` cscratch ¶

univector<complex<ST>> cscratch

Fully qualified name: kfr::convolve_filter<T>::cscratch

Declared at dft/convolution.hpp:238

`variable` scratch1 ¶

univector<T> scratch1

Fully qualified name: kfr::convolve_filter<T>::scratch1

Declared at dft/convolution.hpp:240

`variable` scratch2 ¶

univector<T> scratch1, scratch2

Fully qualified name: kfr::convolve_filter<T>::scratch2

Declared at dft/convolution.hpp:240

`variable` overlap ¶

univector<T> overlap

Fully qualified name: kfr::convolve_filter<T>::overlap

Declared at dft/convolution.hpp:242

Fully qualified name: kfr::convolve_filter<T>

Declared at dft/convolution.hpp:167

This file was generated by cxxdox, a C++ documentation generator based on MkDocs Material and libclang.

Class convolve_filter¶