API Reference

This page documents the Python-facing API introduced by the Eradiate kernel. Only additions specific to this fork are covered here; the base Mitsuba 3 API is documented in the Mitsuba 3 documentation.

How to build this page

This page requires two steps before it renders correctly:

1. Build the sources.

2. Build the API subsite to generate the intermediate RST.

These steps are automated by the pixi run docs-api task. Then rebuild this site normally.

Constants

mitsuba.ERD_MI_VERSION: str = 0.4.3

---

mitsuba.ERD_MI_VERSION_MAJOR: int = 0

---

mitsuba.ERD_MI_VERSION_MINOR: int = 4

---

mitsuba.ERD_MI_VERSION_PATCH: int = 3

---

Medium

class mitsuba.Medium

Base class: mitsuba.Object

get_majorant(self, mi, active=True)

Returns the medium’s majorant used for delta tracking

Parameter mi (mitsuba.MediumInteraction3f):

no description available

Parameter active (bool):

Mask to specify active lanes.

Returns → mitsuba.Color3f:

no description available

get_scattering_coefficients(self, mi, active=True)

Returns the medium coefficients Sigma_s, Sigma_n and Sigma_t evaluated at a given MediumInteraction mi

Parameter mi (mitsuba.MediumInteraction3f):

no description available

Parameter active (bool):

Mask to specify active lanes.

Returns → tuple[mitsuba.Color3f, mitsuba.Color3f, mitsuba.Color3f]:

no description available

has_spectral_extinction()

Returns whether this medium has a spectrally varying extinction

Returns → bool:

no description available

intersect_aabb(self, ray)

Intersects a ray with the medium’s bounding box

Parameter ray (mitsuba.Ray3f):

no description available

Returns → tuple[bool, float, float]:

no description available

is_homogeneous()

Returns whether this medium is homogeneous

Returns → bool:

no description available

property m_has_spectral_extinction

(self) -> bool

property m_is_homogeneous

(self) -> bool

property m_sample_emitters

(self) -> bool

phase_function()

Return the phase function of this medium

Returns → mitsuba.PhaseFunction:

no description available

sample_interaction(self, ray, sample, channel, active)

Sample a free-flight distance in the medium.

This function samples a (tentative) free-flight distance according to an exponential transmittance. It is then up to the integrator to then decide whether the MediumInteraction corresponds to a real or null scattering event.

Parameter ray (mitsuba.Ray3f):

Ray, along which a distance should be sampled

Parameter sample (float):

A uniformly distributed random sample

Parameter channel (int):

The channel according to which we will sample the free-flight distance. This argument is only used when rendering in RGB modes.

Parameter active (bool):

Mask to specify active lanes.

Returns → mitsuba.MediumInteraction3f:

This method returns a MediumInteraction. The MediumInteraction will always be valid, except if the ray missed the Medium’s bounding box.

sample_interaction_analytical(self, ray, it, sample, channel, active)

Sample a free-flight distance in the medium analytically.

This function samples a (tentative) free-flight distance according to an exponential transmittance. It is then up to the integrator to then decide whether the MediumInteraction corresponds to a real or null scattering event.

Parameter ray (mitsuba.Ray3f):

Ray, along which a distance should be sampled

Parameter it (mitsuba.Interaction3f):

The boundary interaction that the sampled distance cannot exceed.

Parameter sample (float):

A uniformly distributed random sample

Parameter channel (int):

The channel according to which we will sample the free-flight distance. This argument is only used when rendering in RGB modes.

Parameter active (bool):

Mask to specify active lanes.

Returns → tuple[mitsuba.MediumInteraction3f, mitsuba.Color3f, mitsuba.Color3f]:

This method returns a MediumInteraction. The MediumInteraction will always be valid, except if the ray missed the Medium’s bounding box.

transmittance_eval_analytical(self, ray, it, active)

Compute the analytical transmittance along a ray to an interaction.

Parameter ray (mitsuba.Ray3f):

Ray, along which to compute the transmittance, use mint

Parameter si:

Interaction that marks the end of the segment along which to compute the transmittance.

Parameter it (mitsuba.Interaction3f):

no description available

Parameter active (bool):

Mask to specify active lanes.

Returns → mitsuba.Color3f:

The transmittance along a ray

transmittance_eval_pdf(self, mi, si, active)

Compute the transmittance and PDF

This function evaluates the transmittance and PDF of sampling a certain free-flight distance The returned PDF takes into account if a medium interaction occurred (mi.t <= si.t) or the ray left the medium (mi.t > si.t)

The evaluated PDF is spectrally varying. This allows to account for the fact that the free-flight distance sampling distribution can depend on the wavelength.

Parameter mi (mitsuba.MediumInteraction3f):

no description available

Parameter si (mitsuba.SurfaceInteraction3f):

no description available

Parameter active (bool):

Mask to specify active lanes.

Returns → tuple[mitsuba.Color3f, mitsuba.Color3f]:

This method returns a pair of (Transmittance, PDF).

use_emitter_sampling()

Returns whether this specific medium instance uses emitter sampling

Returns → bool:

no description available

---

Extremum

class mitsuba.ExtremumSegment

Stores the extremum (minorant/majorant) data for a ray segment.

Used as the output type of ExtremumStructure traversal. Tracks the segment’s entry/exit distances and the local extinction coefficient bounds within that interval.

__init__()

__init__(self, other)

Copy constructor

Parameter other (mitsuba.ExtremumSegment):

no description available

__init__(self, mint, maxt, minorant, majorant)

Construct from entry/exit distances and a combined extremum vector.

Parameter mint (float):

Segment entry distance

Parameter maxt (float):

Segment exit distance

Parameter value (mitsuba.Vector2f):

Extremum vector [minorant, majorant]

Parameter minorant (float):

no description available

Parameter majorant (float):

no description available

__init__(self, mint, maxt, value)

Construct from entry/exit distances and a combined extremum vector.

Parameter mint (float):

Segment entry distance

Parameter maxt (float):

Segment exit distance

Parameter value (mitsuba.Vector2f):

Extremum vector [minorant, majorant]

assign(self, arg)

Parameter arg (mitsuba.ExtremumSegment, /):

no description available

Returns → None:

no description available

majorant()

Majorant value over the segment. Accessor to the second element of value.

Returns → float:

no description available

property maxt

Segment exit distance along ray

minorant()

Minorant value over the segment. Accessor to the first element of value.

Returns → float:

no description available

property mint

Segment entry distance along ray

reset()

Mark the extremum segment as invalid.

This operation sets segment’s minimum and maximum distances to \(\infty\) and \(-\infty\), respectively.

Returns → None:

no description available

valid()

Check whether this is a valid segment

A segment is considered valid when

segment.mint < segment.maxt

Returns → bool:

no description available

property value

Extremum data stored as [minorant, majorant]

zero_(self, size=1)

Overloaded function.

1. zero_(self, size: int = 1) -> None

2. zero_(self, arg: int, /) -> None

This callback method is invoked by dr::zeros<>, and takes care of fields that deviate from the standard zero-initialization convention. In ExtremumSegment, the mint and maxt fields are set to + and - infinity respectively to to mark invalid intersection records.

Parameter size (int):

no description available

Returns → None:

no description available

---

class mitsuba.ExtremumStructure

Base class: mitsuba.Object

Abstract base class for extremum structures

ExtremumStructure provides an interface for spatial data structures that store local extrema (majorant/minorant) of volumetric extinction coefficients. This enables efficient delta tracking with locally- adaptive majorants.

To minimize virtual function overhead, the traverse_extremum() method encapsulates the entire traversal loop internally, requiring only a single virtual call per distance sample.

__init__(self, props)

Parameter props (mitsuba.Properties):

no description available

bbox()

Return the bounding box of the extremum structure

Returns → mitsuba.BoundingBox3f:

no description available

eval_1(self, it, active=True)

Evaluate the minorant and majorant at a medium interaction point.

This method performs point evaluation at interaction point specified in local space.

Parameter it (mitsuba.Interaction3f):

Interaction interaction point in local space

Parameter active (bool):

Mask for active lanes

Returns → tuple[float, float]:

The minorant and majorant values at the medium interaction point. Clamped values outside bounds.