pulp::host::sequencing::rungler Namespace Reference

Classes
struct	Instance

Typedefs
using	Rung = signal::Rungler

Functions
float	rungler_worst_case_gain ()
	Worst-case linear gain for the Forge registry (series law 8).
float	rungler_output_bound_v ()
	The largest absolute output the node can emit, in volts: the ceiling of the range_v param, because a baked param can be automated anywhere in range.
CustomNodeType	make_rungler_node (int reg_bits=Rung::kDefaultBits, std::uint32_t seed_pattern=Rung::kSeedPattern)
	reg_bits and seed_pattern are registration-time.
ForgeNodeDescriptor	descriptor ()

Variables
constexpr const char *	kTypeId = "sequencing.rungler"
constexpr state::ParamID	kRun = 1
constexpr state::ParamID	kDacBits = 2
constexpr state::ParamID	kFeedbackTap = 3
constexpr state::ParamID	kRangeV = 4
constexpr state::ParamID	kExternalData = 5
constexpr state::ParamID	kDataIn = 6

Typedef Documentation

Rung

using pulp::host::sequencing::rungler::Rung = signal::Rungler

Function Documentation

descriptor()

ForgeNodeDescriptor pulp::host::sequencing::rungler::descriptor ( )

inline

make_rungler_node()

CustomNodeType pulp::host::sequencing::rungler::make_rungler_node	(	int	reg_bits = Rung::kDefaultBits,
		std::uint32_t	seed_pattern = Rung::kSeedPattern )

reg_bits and seed_pattern are registration-time.

seed_pattern because a seed is which performance you are playing (law 2). reg_bits because the DSP reloads the seed when the register length changes — a register whose length moves mid-orbit has no meaningful "same state" — so as an injectable param it would re-pin the sequence on every sample and the node would emit one frozen level instead of a line.

rungler_output_bound_v()

float pulp::host::sequencing::rungler::rungler_output_bound_v

(

)

The largest absolute output the node can emit, in volts: the ceiling of the range_v param, because a baked param can be automated anywhere in range.

rungler_worst_case_gain()

float pulp::host::sequencing::rungler::rungler_worst_case_gain

(

)

Worst-case linear gain for the Forge registry (series law 8).

This is the one block in the module with a feedback path, so law 8 bites here — and the answer is a PROVEN bound rather than a measurement. The feedback is a single-bit XOR into a shift register: there is no gain-carrying analog nonlinearity, so no small-signal-gain compensation applies. The output is a D-bit DAC code mapped affinely onto [−range_v, +range_v], so |y| ≤ range_v BY CONSTRUCTION for any clock or data sequence whatsoever. The DSP suite asserts exactly that over 10⁶ clocks of adversarial data (all-ones, alternating, xorshift) at three configurations, and asserts the bound is TIGHT — both extremes are reached. Reported as 1.0 because there is no input-to-output amplitude path at all; the absolute output bound is range_v, whose param ceiling is below.

Variable Documentation

kDacBits

state::ParamID pulp::host::sequencing::rungler::kDacBits = 2

inlineconstexpr

kDataIn

state::ParamID pulp::host::sequencing::rungler::kDataIn = 6

inlineconstexpr

kExternalData

state::ParamID pulp::host::sequencing::rungler::kExternalData = 5

inlineconstexpr

kFeedbackTap

state::ParamID pulp::host::sequencing::rungler::kFeedbackTap = 3

inlineconstexpr

kRangeV

state::ParamID pulp::host::sequencing::rungler::kRangeV = 4

inlineconstexpr

kRun

state::ParamID pulp::host::sequencing::rungler::kRun = 1

inlineconstexpr

kTypeId

const char* pulp::host::sequencing::rungler::kTypeId = "sequencing.rungler"

inlineconstexpr