[libcamera-devel] [PATCH 2/4] libipa: camera_sensor_helper: Implement exponential gain model

[Jacopo Mondi]

Hi Laurent,

On Mon, Mar 28, 2022 at 03:03:34PM +0300, Laurent Pinchart via libcamera-devel wrote:
> The CameraSensorHelper specifies two gain models, linear and an

s/an //

> exponential. They are modelled after the MIPI CCS specification. Only
> the linear model has been implemented, the exponential model was left
> for later.
>
> We now need to support sensors that configure their gain in a hardware
> register with a value expressed in dB. This has similarities with the
> MIPI CCS exponential gain model, but is only has an exponential factor,
> while CCS also allows sensors to support a configurable linear factor.
>
> The full CCS exponential model needs two values (for the linear and
> exponential factors) to express a gain, while IPAs use a single linear
> gain value internally. However, the exponential gain model example in
> the CCS specification has a fixed linear factor, which may indicate that
> it could be common for sensors that implement the exponential gain model
> to only use the exponential factor. For this reason, implement the
> exponential gain model with a fixed linear factor, but with a

Isn't the linear factor k.exp.a not fixed ?

        return k.exp.a * std::exp2(k.exp.m * gain);

Reading this I would have expected k.exp.a to be fixed to 1

> sensor-specific coefficient for the exponential factor that allows
> expressing the gain in dB (or other logarithmical units) instead of
> limiting it to powers of 2 as in the MIPI CCS specification.
>
> Signed-off-by: Laurent Pinchart <laurent.pinchart at ideasonboard.com>
> ---
>  src/ipa/libipa/camera_sensor_helper.cpp | 82 ++++++++++++++++++-------
>  src/ipa/libipa/camera_sensor_helper.h   |  6 ++
>  2 files changed, 67 insertions(+), 21 deletions(-)
>
> diff --git a/src/ipa/libipa/camera_sensor_helper.cpp b/src/ipa/libipa/camera_sensor_helper.cpp
> index 714cd86f039f..7bb999e19102 100644
> --- a/src/ipa/libipa/camera_sensor_helper.cpp
> +++ b/src/ipa/libipa/camera_sensor_helper.cpp
> @@ -7,6 +7,8 @@
>   */
>  #include "camera_sensor_helper.h"
>
> +#include <cmath>
> +
>  #include <libcamera/base/log.h>
>
>  /**
> @@ -51,20 +53,28 @@ namespace ipa {
>   * This function aims to abstract the calculation of the gain letting the IPA
>   * use the real gain for its estimations.
>   *
> - * The parameters come from the MIPI Alliance Camera Specification for
> - * Camera Command Set (CCS).
> - *
>   * \return The gain code to pass to V4L2
>   */
>  uint32_t CameraSensorHelper::gainCode(double gain) const
>  {
>  	const AnalogueGainConstants &k = gainConstants_;
>
> -	ASSERT(gainType_ == AnalogueGainLinear);
> -	ASSERT(k.linear.m0 == 0 || k.linear.m1 == 0);
> +	switch (gainType_) {
> +	case AnalogueGainLinear:
> +		ASSERT(k.linear.m0 == 0 || k.linear.m1 == 0);
>
> -	return (k.linear.c0 - k.linear.c1 * gain) /
> -	       (k.linear.m1 * gain - k.linear.m0);
> +		return (k.linear.c0 - k.linear.c1 * gain) /
> +		       (k.linear.m1 * gain - k.linear.m0);
> +
> +	case AnalogueGainExponential:
> +		ASSERT(k.exp.a != 0 && k.exp.m != 0);
> +
> +		return std::log2(gain / k.exp.a) / k.exp.m;
> +
> +	default:
> +		ASSERT(false);
> +		return 0;
> +	}
>  }
>
>  /**
> @@ -75,20 +85,29 @@ uint32_t CameraSensorHelper::gainCode(double gain) const
>   * use the real gain for its estimations. It is the counterpart of the function
>   * CameraSensorHelper::gainCode.
>   *
> - * The parameters come from the MIPI Alliance Camera Specification for
> - * Camera Command Set (CCS).
> - *
>   * \return The real gain
>   */
>  double CameraSensorHelper::gain(uint32_t gainCode) const
>  {
>  	const AnalogueGainConstants &k = gainConstants_;
> +	double gain = static_cast<double>(gainCode);
>
> -	ASSERT(gainType_ == AnalogueGainLinear);
> -	ASSERT(k.linear.m0 == 0 || k.linear.m1 == 0);
> +	switch (gainType_) {
> +	case AnalogueGainLinear:
> +		ASSERT(k.linear.m0 == 0 || k.linear.m1 == 0);
>
> -	return (k.linear.m0 * static_cast<double>(gainCode) + k.linear.c0) /
> -	       (k.linear.m1 * static_cast<double>(gainCode) + k.linear.c1);
> +		return (k.linear.m0 * gain + k.linear.c0) /
> +		       (k.linear.m1 * gain + k.linear.c1);
> +
> +	case AnalogueGainExponential:
> +		ASSERT(k.exp.a != 0 && k.exp.m != 0);
> +
> +		return k.exp.a * std::exp2(k.exp.m * gain);
> +
> +	default:
> +		ASSERT(false);
> +		return 0.0;
> +	}
>  }
>
>  /**
> @@ -120,15 +139,22 @@ double CameraSensorHelper::gain(uint32_t gainCode) const
>
>  /**
>   * \var CameraSensorHelper::AnalogueGainExponential
> - * \brief Gain is computed using exponential gain estimation
> - * (introduced in CCS v1.1)
> + * \brief Gain is expressed using an exponential model
>   *
> - * Starting with CCS v1.1, Alternate Global Analogue Gain is also available.
> - * If the image sensor supports it, then the global analogue gain can be
> - * controlled by linear and exponential gain formula:
> + * The relationship between the integer gain parameter and the resulting gain
> + * multiplier is given by the following equation:
>   *
> - * \f$gain = analogLinearGainGlobal * 2^{analogExponentialGainGlobal}\f$
> - * \todo not implemented in libipa
> + * \f$gain = a \cdot 2^{m \cdot x}\f$
> + *
> + * 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.

So I don't really get what 'constant' means, as this seems to match
the CCS spec (apart from the 'm' paramter)

> + *
> + * When the gain is expressed in dB, 'a' is equal to 1 and 'm' to
> + * \f$log_{2}{10^{frac{1}{20}}}\f$.
>   */
>
>  /**
> @@ -152,6 +178,17 @@ double CameraSensorHelper::gain(uint32_t gainCode) const
>   * \brief Constant used in the linear gain coding/decoding
>   */
>
> +/**
> + * \struct CameraSensorHelper::AnalogueGainExpConstants
> + * \brief Analogue gain constants for the exponential gain model
> + *
> + * \var CameraSensorHelper::AnalogueGainExpConstants::a
> + * \brief Constant used in the exponential gain coding/decoding
> + *
> + * \var CameraSensorHelper::AnalogueGainExpConstants::m
> + * \brief Constant used in the exponential gain coding/decoding
> + */
> +
>  /**
>   * \struct CameraSensorHelper::AnalogueGainConstants
>   * \brief Analogue gain model constants
> @@ -161,6 +198,9 @@ double CameraSensorHelper::gain(uint32_t gainCode) const
>   *
>   * \var CameraSensorHelper::AnalogueGainConstants::linear
>   * \brief Constants for the linear gain model
> + *
> + * \var CameraSensorHelper::AnalogueGainConstants::exp
> + * \brief Constants for the exponential gain model
>   */
>
>  /**
> diff --git a/src/ipa/libipa/camera_sensor_helper.h b/src/ipa/libipa/camera_sensor_helper.h
> index 6b96520ba601..7351fc7c2928 100644
> --- a/src/ipa/libipa/camera_sensor_helper.h
> +++ b/src/ipa/libipa/camera_sensor_helper.h
> @@ -41,8 +41,14 @@ protected:
>  		int16_t c1;
>  	};
>
> +	struct AnalogueGainExpConstants {
> +		double a;
> +		double m;
> +	};
> +
>  	union AnalogueGainConstants {
>  		AnalogueGainLinearConstants linear;
> +		AnalogueGainExpConstants exp;
>  	};

With my confusion about what 'constant' means clarified:

Reviewed-by: Jacopo Mondi <jacopo at jmondi.org>


>
>  	AnalogueGainType gainType_;
> --
> Regards,
>
> Laurent Pinchart
>