Structs for configuring audio processors used in the FDN. More...

Classes
struct	sfFDN::ScalarFeedbackMatrixOptions
	Options for configuring a scalar feedback matrix. More...

struct	sfFDN::CascadedFeedbackMatrixOptions
	Information structure for constructing a cascaded feedback matrix (also known as a filter feedback matrix). More...

struct	sfFDN::ModulationOptions
	Options for configuring signal modulation. More...

struct	sfFDN::ParallelGainsOptions
	Options for configuring parallel gain processing. More...

struct	sfFDN::DelayOptions
	Options for configuring delays. More...

struct	sfFDN::DelayBankOptions
	Options for configuring a delay bank. More...

struct	sfFDN::DelayBankTimeVaryingOptions
	Options for configuring a time-varying delay bank. More...

struct	sfFDN::FilterCoefficients
	Coefficients for a digital IIR filter. More...

struct	sfFDN::AllpassFilterOptions
	Options for configuring an allpass filter. More...

struct	sfFDN::SparseFirOptions
	Options for configuring a sparse FIR filter. More...

struct	sfFDN::CascadedBiquadsOptions
	Options for configuring cascaded biquad filters. More...

struct	sfFDN::FirOptions
	Options for configuring a FIR filter. More...

struct	sfFDN::SchroederAllpassSectionOptions
	Options for configuring a Schroeder allpass section consisting of `N` Schroeder allpass in series or in parallel. More...

struct	sfFDN::MultichannelSchroederAllpassSectionOptions
	Options for configuring a multichannel bank of Schroeder allpass sections. More...

struct	sfFDN::HomogenousFilterOptions
	Options for configuring a homogenous filter. More...

struct	sfFDN::TwoBandFilterOptions
	Options for configuring a two-band filter. More...

struct	sfFDN::ThreeBandFilterOptions
	Options for configuring a three-band filter. More...

struct	sfFDN::TenBandFilterOptions
	Options for configuring a ten-band filter. More...

struct	sfFDN::AttenuationFilterBankOptions
	Options for configuring an attenuation filter bank. More...

struct	sfFDN::GraphicEQOptions
	Options for configuring a graphic equalizer. More...

Typedefs
using	sfFDN::attenuation_filter_variant_t
	Variant type for holding different attenuation filter options.

using	sfFDN::feedback_matrix_variant_t = std::variant<CascadedFeedbackMatrixOptions, ScalarFeedbackMatrixOptions>
	Variant type for holding different feedback matrix options.

using	sfFDN::single_channel_processor_variant_t
	Variant type for holding different single-channel processor options.

using	sfFDN::multi_channel_processor_variant_t
	Variant type for holding different multi-channel processor options.

Enumerations
enum class	sfFDN::ScalarMatrixType : uint8_t { ScalarMatrixType::Identity = 0 , ScalarMatrixType::Random = 1 , ScalarMatrixType::Householder = 2 , ScalarMatrixType::RandomHouseholder = 3 , ScalarMatrixType::Hadamard = 4 , ScalarMatrixType::Circulant = 5 , ScalarMatrixType::Allpass = 6 , ScalarMatrixType::NestedAllpass = 7 , ScalarMatrixType::VariableDiffusion = 8 , Count = 9 }
	Represents the type of a scalar matrix. More...

enum class	sfFDN::DelayInterpolationType : uint8_t { DelayInterpolationType::None = 0 , DelayInterpolationType::Linear = 1 , DelayInterpolationType::Allpass = 2 , Lagrange = 3 , Count = 4 }
	Types of interpolation for fractional delay lengths. More...

enum class	sfFDN::DelayLengthType : uint8_t { DelayLengthType::Random = 0 , DelayLengthType::Gaussian = 1 , DelayLengthType::Primes = 2 , DelayLengthType::Uniform = 3 , DelayLengthType::PrimePower = 4 , DelayLengthType::SteamAudio = 5 , Count = 6 }
	Types of delay length distributions. More...

enum class	sfFDN::ParallelGainsMode : uint8_t { ParallelGainsMode::Split , ParallelGainsMode::Merge , ParallelGainsMode::Parallel }
	Enumeration for parallel gain processing modes. More...

Detailed Description

Structs for configuring audio processors used in the FDN.

Class Documentation

◆ sfFDN::ScalarFeedbackMatrixOptions

struct sfFDN::ScalarFeedbackMatrixOptions

Options for configuring a scalar feedback matrix.

Can be use to construct a ScalarFeedbackMatrix.

Class Members
optional< float >	arg {std::nullopt}	Optional argument for certain matrix types. For example, for the VariableDiffusion type, this could represent.
optional< vector< float > >	custom_matrix {std::nullopt}	Optional custom matrix values in col-major order. The size of the vector must be equal to matrix_size*matrix_size. If this is set, `type` is ignored.
uint32_t	matrix_size	Size of the feedback matrix.
uint32_t	rng_seed {0}	Optional. Seed for random number generation when type is Random or RandomHouseholder.
ScalarMatrixType	type {ScalarMatrixType::Random}	Type of the feedback matrix.

◆ sfFDN::CascadedFeedbackMatrixOptions

struct sfFDN::CascadedFeedbackMatrixOptions

Information structure for constructing a cascaded feedback matrix (also known as a filter feedback matrix).

A cascaded feedback matrix is composed of multiple stages, where each stage consists of a scalar feedback matrix followed by a bank of delay lines.

Can be used to construct a FilterFeedbackMatrix.

Class Members
float	gain_per_samples {1.f}	Gain per sample.
uint32_t	matrix_size	Size of the feedback matrix.
float	sparsity {1.f}	Sparsity level (>= 1). A value of 1 corresponds to a fully dense matrix, while higher values correspond to sparser matrices.
uint32_t	stage_count	Number of stages.
ScalarMatrixType	type	Type of the feedback matrix. The same type is used for all stages.

◆ sfFDN::ModulationOptions

struct sfFDN::ModulationOptions

Options for configuring signal modulation.

Class Members
float	amplitude {0.f}
float	frequency {0.f}
float	initial_phase {0.f}

◆ sfFDN::ParallelGainsOptions

struct sfFDN::ParallelGainsOptions

Options for configuring parallel gain processing.

Class Members
vector< float >	gains	Gain values for each channel. The size of the vector determines the number of channels.
ParallelGainsMode	mode {ParallelGainsMode::Split}	Mode of parallel gain processing.
vector< ModulationOptions >	time_varying_config {}	Optional time-varying modulation configuration for each channel. The size of the vector must match the size of `gains` if provided.

◆ sfFDN::DelayOptions

struct sfFDN::DelayOptions

Options for configuring delays.

Class Members
float	delay {256.f}
DelayInterpolationType	interp_type
optional< ModulationOptions >	lfo_config
uint32_t	max_delay {512}

◆ sfFDN::DelayBankOptions

struct sfFDN::DelayBankOptions

Options for configuring a delay bank.

Class Members
uint32_t	block_size {kDefaultBlockSize}
vector< float >	delays
DelayInterpolationType	interpolation_type

◆ sfFDN::DelayBankTimeVaryingOptions

struct sfFDN::DelayBankTimeVaryingOptions

Options for configuring a time-varying delay bank.

Class Members
vector< float >	delays
DelayInterpolationType	interpolation_type
uint32_t	max_delay
vector< ModulationOptions >	time_varying_config

◆ sfFDN::AllpassFilterOptions

struct sfFDN::AllpassFilterOptions

Options for configuring an allpass filter.

Class Members
float	coeff {0.f}	The coefficient for the allpass filter.

◆ sfFDN::SparseFirOptions

struct sfFDN::SparseFirOptions

Options for configuring a sparse FIR filter.

Class Members
vector< pair< uint32_t, float > >	coeffs

◆ sfFDN::CascadedBiquadsOptions

struct sfFDN::CascadedBiquadsOptions

Options for configuring cascaded biquad filters.

Class Members
vector< FilterCoefficients >	coeffs

◆ sfFDN::FirOptions

struct sfFDN::FirOptions

Options for configuring a FIR filter.

Class Members
vector< float >	coeffs {1.f}

◆ sfFDN::SchroederAllpassSectionOptions

struct sfFDN::SchroederAllpassSectionOptions

Options for configuring a Schroeder allpass section consisting of N Schroeder allpass in series or in parallel.

Class Members
vector< float >	delays
vector< float >	gains
bool	parallel {false}

◆ sfFDN::MultichannelSchroederAllpassSectionOptions

struct sfFDN::MultichannelSchroederAllpassSectionOptions

Options for configuring a multichannel bank of Schroeder allpass sections.

Each section processes one channel of audio.

Class Members
vector< SchroederAllpassSectionOptions >	sections

◆ sfFDN::HomogenousFilterOptions

struct sfFDN::HomogenousFilterOptions

Options for configuring a homogenous filter.

The homogenous filter has the same attenuation characteristics across all frequencies.

Class Members
float	delay
float	sample_rate = kDefaultSampleRate
float	t60 = 1.f

◆ sfFDN::TwoBandFilterOptions

struct sfFDN::TwoBandFilterOptions

Options for configuring a two-band filter.

The two-band filter allows for specifying a target t60 at DC and Nyquist. The filter is implemented as a one-pole filter based on [1].

[1] Jot, J. M., & Chaigne, A. (1991). Digital delay networks for designing artificial reverberators (pp. 1-12). Presented at the Proc. Audio Eng. Soc. Conv., Paris, France.

Class Members
float	delay
float	sample_rate = kDefaultSampleRate
array< float, 2 >	t60s {1.f, 0.5f}	Target T60 values for the low and high bands.

◆ sfFDN::ThreeBandFilterOptions

struct sfFDN::ThreeBandFilterOptions

Options for configuring a three-band filter.

The three-band filter is composed of a 2nd order low shelf and a 2nd order high shelf filter in series and allows control of the T60 over three bands with the first band being [0, freq[0]], the second band [freq[0], freq[1]], and the third band [freq[1], Nyquist].

Class Members
float	delay
array< float, 2 >	freqs {800.f, 8000.f}
float	q = 1.f / std::numbers::sqrt2_v<float>
float	sample_rate = kDefaultSampleRate
array< float, 3 >	t60s {1.f, 0.5f, 0.25f}

◆ sfFDN::TenBandFilterOptions

struct sfFDN::TenBandFilterOptions

Options for configuring a ten-band filter.

The ten-band filter allows control of the T60 over ten bands. The bands of the filter are set to {32, 64, 125, 250, 500, 1k, 2k, 4k, 8k, 16k} Hz.

The filter is implemented as a cascade of second-order biquad filters following the design method described in [1]. [1] V. Välimäki, K. Prawda, and S. J. Schlecht, "Two-Stage Attenuation Filter for Artificial Reverberation," IEEE Signal Processing Letters, vol. 31, pp. 391–395, 2024, doi: 10.1109/LSP.2024.3352510.

Class Members
float	delay	Delay in samples for the delay line preceding the filter. If set to <= 0, it will be updated automatically when accessed from `CreateFDNFromConfig()`
float	sample_rate = kDefaultSampleRate	Sample rate in Hz. This is used to calculate the filter coefficients based on the specified T60 values.
float	shelf_cutoff = 8000.f	Cutoff frequency for the shelf filters.
array< float, 10 >	t60s = {1.f, 0.9f, 0.8f, 0.7f, 0.6f, 0.5f, 0.4f, 0.3f, 0.2f, 0.1f}	Target T60 values for the ten bands.

◆ sfFDN::AttenuationFilterBankOptions

struct sfFDN::AttenuationFilterBankOptions

Options for configuring an attenuation filter bank.

Class Members
vector< attenuation_filter_variant_t >	filter_configs	Vector of attenuation filter configurations.

◆ sfFDN::GraphicEQOptions

struct sfFDN::GraphicEQOptions

Options for configuring a graphic equalizer.

Class Members
array< float, 10 >	freqs	Frequency values for the ten bands in Hz.
array< float, 10 >	gains_db	Target gains for the ten bands in dB.
float	sample_rate = kDefaultSampleRate	Sample rate in Hz.

Typedef Documentation

◆ attenuation_filter_variant_t

using sfFDN::attenuation_filter_variant_t

Initial value:

std::variant<HomogenousFilterOptions, TwoBandFilterOptions, ThreeBandFilterOptions, TenBandFilterOptions>

Variant type for holding different attenuation filter options.

◆ multi_channel_processor_variant_t

using sfFDN::multi_channel_processor_variant_t

Initial value:

 
    std::variant<ParallelGainsOptions, MultichannelSchroederAllpassSectionOptions, AttenuationFilterBankOptions,
                 DelayBankOptions, DelayBankTimeVaryingOptions, CascadedFeedbackMatrixOptions,
                 ScalarFeedbackMatrixOptions>

Variant type for holding different multi-channel processor options.

◆ single_channel_processor_variant_t

using sfFDN::single_channel_processor_variant_t

Initial value:

 
    std::variant<SchroederAllpassSectionOptions, AllpassFilterOptions, CascadedBiquadsOptions, FirOptions, DelayOptions,
                 GraphicEQOptions>

Variant type for holding different single-channel processor options.

Enumeration Type Documentation

◆ DelayInterpolationType

enum class sfFDN::DelayInterpolationType : uint8_t

strong

Types of interpolation for fractional delay lengths.

Enumerator
None	No interpolation. The delay length will be rounded to the nearest integer value.
Linear	Linear interpolation.
Allpass	Allpass interpolation.

◆ DelayLengthType

enum class sfFDN::DelayLengthType : uint8_t

strong

Types of delay length distributions.

Enumerator
Random	Delay lengths are generated randomly within the specified range based on a uniform distribution.
Gaussian	Delay lengths are generated based on a Gaussian distribution within the specified range.
Primes	Delay lengths are selected randomly from a list of prime numbers within the specified range.
Uniform	Delay lengths are uniformly distributed within the specified range.
PrimePower	Delay lengths are generated as powers of prime numbers within the specified range. Based on https://ccrma.stanford.edu/~jos/pasp/Prime_Power_Delay_Line_Lengths.html.
SteamAudio	Delay lengths are generated using the algorithm from the SteamAudio library.

◆ ParallelGainsMode

enum class sfFDN::ParallelGainsMode : uint8_t

strong

Enumeration for parallel gain processing modes.

Enumerator
Split	Process input as a single channel and output to multiple channels.
Merge	Process each input channel separately and output to one channel.
Parallel	Process each input channel separately and output to the same number of channels.

◆ ScalarMatrixType

enum class sfFDN::ScalarMatrixType : uint8_t

strong

Represents the type of a scalar matrix.

[1] D. Rocchesso and J. O. Smith, “Circulant and elliptic feedback delay networks for artificial reverberation,” IEEE Transactions on Speech and Audio Processing, vol. 5, no. 1, pp. 51–63, Jan. 1997, doi: 10.1109/89.554269.
[2] S. J. Schlecht, “FDNTB: the feedback delay network toolbox,” 23rd International Conference on Digital Audio Effects (DAFx2020), 2020.
[3] O. Das, E. K. Canfield-Dafilou, and J. S. Abel, “On The Behavior of Delay Network Reverberator Modes,” in 2019 IEEE Workshop on Applications of Signal Processing to Audio and Acoustics (WASPAA), Oct. 2019, pp. 50–54. doi: 10.1109/WASPAA.2019.8937260.

Enumerator
Identity	Identity matrix.
Random	Random orthogonal matrix.
Householder	Householder matrix.
RandomHouseholder	Random Householder matrix.
Hadamard	Hadamard matrix.
Circulant	Circulant matrix as described in [1].
Allpass	Allpass matrix. See [2]
NestedAllpass	Nested Allpass matrix. See [2]
VariableDiffusion	Variable diffusion matrix as described in [3].

Classes

Typedefs

Enumerations

Detailed Description

Class Documentation

◆ sfFDN::ScalarFeedbackMatrixOptions

◆ sfFDN::CascadedFeedbackMatrixOptions

◆ sfFDN::ModulationOptions

◆ sfFDN::ParallelGainsOptions

◆ sfFDN::DelayOptions

◆ sfFDN::DelayBankOptions

◆ sfFDN::DelayBankTimeVaryingOptions

◆ sfFDN::AllpassFilterOptions

◆ sfFDN::SparseFirOptions

◆ sfFDN::CascadedBiquadsOptions

◆ sfFDN::FirOptions

◆ sfFDN::SchroederAllpassSectionOptions

◆ sfFDN::MultichannelSchroederAllpassSectionOptions

◆ sfFDN::HomogenousFilterOptions

◆ sfFDN::TwoBandFilterOptions

◆ sfFDN::ThreeBandFilterOptions

◆ sfFDN::TenBandFilterOptions

◆ sfFDN::AttenuationFilterBankOptions

◆ sfFDN::GraphicEQOptions

Typedef Documentation

◆ attenuation_filter_variant_t

◆ multi_channel_processor_variant_t

◆ single_channel_processor_variant_t

Enumeration Type Documentation

◆ DelayInterpolationType

◆ DelayLengthType

◆ ParallelGainsMode

◆ ScalarMatrixType