Base class for computing sensor tuning parameters using sensor-specific constants. More...

Collaboration diagram for libcamera::ipa::CameraSensorHelper:

Classes
struct	AnalogueGainConstants
	Analogue gain model constants. More...

struct	AnalogueGainExpConstants
	Analogue gain constants for the exponential gain model. More...

struct	AnalogueGainLinearConstants
	Analogue gain constants for the linear gain model. More...

Public Member Functions
std::optional< int16_t >	blackLevel () const
	Construct a CameraSensorHelper instance. More...

virtual uint32_t	gainCode (double gain) const
	Compute gain code from the analogue gain absolute value. More...

virtual double	gain (uint32_t gainCode) const
	Compute the real gain from the V4L2 subdev control gain code. More...

Protected Types
enum	AnalogueGainType { AnalogueGainLinear, AnalogueGainExponential }
	The gain calculation modes as defined by the MIPI CCS. More...

Protected Attributes
std::optional< int16_t >	blackLevel_
	The black level of the sensor.

AnalogueGainType	gainType_
	The analogue gain model type.

AnalogueGainConstants	gainConstants_
	The analogue gain parameters used for calculation. More...

Detailed Description

Base class for computing sensor tuning parameters using sensor-specific constants.

Instances derived from CameraSensorHelper class are sensor-specific. Each supported sensor will have an associated base class defined.

Member Enumeration Documentation

◆ AnalogueGainType

enum libcamera::ipa::CameraSensorHelper::AnalogueGainType

protected

The gain calculation modes as defined by the MIPI CCS.

Describes the image sensor analogue gain capabilities. Two modes are possible, depending on the sensor: Linear and Exponential.

Enumerator

AnalogueGainLinear

Gain is computed using linear gain estimation.

The relationship between the integer gain parameter and the resulting gain multiplier is given by the following equation:

$gain=\frac{m0x+c0}{m1x+c1}$

Where 'x' is the gain control parameter, and m0, m1, c0 and c1 are image-sensor-specific constants of the sensor. These constants are static parameters, and for any given image sensor either m0 or m1 shall be zero.

The full Gain equation therefore reduces to either:

$gain=\frac{c0}{m1x+c1}$ or $\frac{m0x+c0}{c1}$

AnalogueGainExponential

Gain is expressed using an exponential model.

The relationship between the integer gain parameter and the resulting gain multiplier is given by the following equation:

$gain = a \cdot 2^{m \cdot x}$

Where 'x' is the gain control parameter, and 'a' and 'm' are image sensor-specific constants.

This is a subset of the MIPI CCS exponential gain model with the linear factor 'a' being a constant, but with the exponent being configurable through the 'm' coefficient.

When the gain is expressed in dB, 'a' is equal to 1 and 'm' to $log_{2}{10^{\frac{1}{20}}}$ .

Member Function Documentation

◆ blackLevel()

libcamera::ipa::CameraSensorHelper::blackLevel ( ) const

inline

Construct a CameraSensorHelper instance.

CameraSensorHelper derived class instances shall never be constructed manually but always through the CameraSensorHelperFactoryBase::create() function.

Fetch the black level of the sensor

This function returns the black level of the sensor scaled to a 16bit pixel width. If it is unknown an empty optional is returned.

Todo:: Fill the blanks and add pedestal values for all supported sensors. Once done, drop the std::optional<>.

Black levels are typically the result of the following phenomena:

Pedestal added by the sensor to pixel values. They are typically fixed, sometimes programmable and should be reported in datasheets (but documentation is not always available).
Dark currents and other physical effects that add charge to pixels in the absence of light. Those can depend on the integration time and the sensor die temperature, and their contribution to pixel values depend on the sensor gains.

The pedestal is usually the value with the biggest contribution to the overall black level. In most cases it is either known before or in rare cases (there is not a single driver with such a control in the linux kernel) can be queried from the sensor. This function provides that fixed, known value.

Returns: The black level of the sensor, or std::nullopt if not known

◆ gain()

double libcamera::ipa::CameraSensorHelper::gain ( uint32_t gainCode ) const

virtual

Compute the real gain from the V4L2 subdev control gain code.

Parameters

[in] gainCode The V4L2 subdev control gain

This function aims to abstract the calculation of the gain letting the IPA use the real gain for its estimations. It is the counterpart of the function CameraSensorHelper::gainCode.

Returns: The real gain

◆ gainCode()

uint32_t libcamera::ipa::CameraSensorHelper::gainCode ( double gain ) const

virtual

Compute gain code from the analogue gain absolute value.

Parameters

[in] gain The real gain to pass

This function aims to abstract the calculation of the gain letting the IPA use the real gain for its estimations.

Returns: The gain code to pass to V4L2

Member Data Documentation

◆ gainConstants_

libcamera::ipa::CameraSensorHelper::gainConstants_

protected

The analogue gain parameters used for calculation.

The analogue gain is calculated through a formula, and its parameters are sensor specific. Use this variable to store the values at init time.

The documentation for this class was generated from the following files:

src/ipa/libipa/camera_sensor_helper.h
src/ipa/libipa/camera_sensor_helper.cpp

Classes

Public Member Functions

Protected Types

Protected Attributes