TEV definitions. More...

#include <stdint.h>
#include <stdbool.h>

Classes
struct	gx_tev_stage_t

Enumerations
enum	gx_tev_stage_id { GX_TEV_STAGE_0 , GX_TEV_STAGE_1 , GX_TEV_STAGE_2 , GX_TEV_STAGE_3 , GX_TEV_STAGE_4 , GX_TEV_STAGE_5 , GX_TEV_STAGE_6 , GX_TEV_STAGE_7 , GX_TEV_STAGE_8 , GX_TEV_STAGE_9 , GX_TEV_STAGE_10 , GX_TEV_STAGE_11 , GX_TEV_STAGE_12 , GX_TEV_STAGE_13 , GX_TEV_STAGE_14 , GX_TEV_STAGE_15 }

enum	gx_tev_io_t { GX_TEV_IO_PREVIOUS = 0x0 , GX_TEV_IO_REGISTER_0 = 0x2 , GX_TEV_IO_REGISTER_1 = 0x4 , GX_TEV_IO_REGISTER_2 = 0x6 , GX_TEV_IO_TEXTURE = 0x8 , GX_TEV_IO_RASTERIZER = 0xA , GX_TEV_IO_ONE = 0xC , GX_TEV_IO_HALF = 0xD , GX_TEV_IO_CONSTANT = 0xE , GX_TEV_IO_ZERO = 0xF , GX_TEV_IO_ALPHA = 0x1 }

enum	gx_tev_scale_t { GX_TEV_SCALE_1 , GX_TEV_SCALE_2 , GX_TEV_SCALE_4 , GX_TEV_SCALE_HALF }

enum	gx_tev_bias_t { GX_TEV_BIAS_0 , GX_TEV_BIAS_ADD_HALF , GX_TEV_BIAS_SUB_HALF }

enum	gx_tev_compare_t { GX_TEV_COMPARE_R8_GREATER = 0x0 , GX_TEV_COMPARE_R8_EQUAL = 0x1 , GX_TEV_COMPARE_GR16_GREATER = 0x2 , GX_TEV_COMPARE_GR16_EQUAL = 0x3 , GX_TEV_COMPARE_BGR24_GREATER = 0x4 , GX_TEV_COMPARE_BGR24_EQUAL = 0x5 , GX_TEV_COMPARE_RGB8_GT = 0x6 , GX_TEV_COMPARE_RGB8_EQ = 0x7 , GX_TEV_COMPARE_A8_GREATER = 0x6 , GX_TEV_COMPARE_A8_EQUAL = 0x7 }

Functions
void	gx_set_tev_stages (int count)
	Sets the number of tev stages.

void	gx_initialize_tev_stage (gx_tev_stage_t *tev)
	Generates a empty TEV stage.

void	gx_flash_tev_stage (gx_tev_stage_id id, const gx_tev_stage_t *tev)
	Flashes a TEV stage onto the BP.

void	gx_set_tev_stage_color_input (gx_tev_stage_t *tev, gx_tev_io_t a, gx_tev_io_t b, gx_tev_io_t c, gx_tev_io_t d)
	Sets the color inputs to a TEV stage.

void	gx_set_tev_stage_alpha_input (gx_tev_stage_t *tev, gx_tev_io_t a, gx_tev_io_t b, gx_tev_io_t c, gx_tev_io_t d)
	Sets the color inputs to a TEV stage.

void	gx_set_tev_stage_color_output (gx_tev_stage_t *tev, gx_tev_io_t out, bool clamp)
	Sets the color output to a TEV stage.

void	gx_set_tev_stage_alpha_output (gx_tev_stage_t *tev, gx_tev_io_t out, bool clamp)
	Sets the alpha output to a TEV stage.

void	gx_set_tev_stage_color_biasing (gx_tev_stage_t *tev, bool subtract, gx_tev_bias_t bias, gx_tev_scale_t scale)
	Sets the TEV color stage into biasing mode.

void	gx_set_tev_stage_color_comparison (gx_tev_stage_t *tev, gx_tev_compare_t comparison)
	Sets the TEV color stage into comparison mode.

void	gx_set_tev_stage_alpha_biasing (gx_tev_stage_t *tev, bool subtract, gx_tev_bias_t bias, gx_tev_scale_t scale)
	Sets the TEV alpha stage into biasing mode.

void	gx_set_tev_stage_alpha_comparison (gx_tev_stage_t *tev, gx_tev_compare_t comparison)
	Sets the TEV alpha stage into comparison mode.

void	gx_flash_tev_register_color (gx_tev_io_t reg, int r, int g, int b, int a)
	Set the value of one of the TEVs internal registers.

Detailed Description

TEV definitions.

Though the TEV is really part of the BP. The TEV registers have been made their own header. So that their functions are organized.

Diagram of a TEV Stage:

                                                                                    /|     +------------+
                                                                                   / |>----| PREVIOUS   |

+------—+ 8 +----------------------—+ / | +---------—+ | INPUT A |>-/–| | / | +------—+ | | | | +---------—+ | | | |>-—| REGISTER 1 | +------—+ 8 | COLOR MIXER | +------------—+ +----—+ 11 | | +---------—+ | INPUT B |>-/–| mix(a, b, c) |>-—| ADD & SCALE |>-—| CLAMP |>-/–| | +------—+ | (a * (1.0 - c) + b * c) | | ((c+d+b) * s) | +----—+ | | +---------—+ | | +------------—| | |>-—| REGISTER 2 | +------—+ 8 | | | | | +---------—+ | INPUT C |>-/–| | | \ | +------—+ +----------------------—+ | \ | +---------—+ | \ |>-—| REGISTER 3 | +------—+ 11 | | +---------—+ | INPUT D |—/-------------------------------------—+ +------—+

Notes:

Each of the inputs, A-D, are selectable.
Inputs are truncated to 8 bits, except for D that is a signed 11 bits, [-1024, 1023]
The addition is input D, additional bias, then scale is set to be 0.5, 1, 2, or 4
If clamping is enabled, color values are capped [0, 255], otherwise [-1024, 1023]
You can save the output to 1 of 4 TEV registers, or just use the tev previous register.
TEV previous becomes the pixels color at the end of the TEV pipeline.
TEVs default mode is biasing mode, it can be in compare mode too. No mixing, It just says, if operation on input A and B is true, pass C, otherwise, 0. It still adds D as a bias.

Author: Samuel Fitzsimons (rainbain)

Date: 2025 @license MIT (see LICENSE file)

Function Documentation

◆ gx_flash_tev_register_color()

void gx_flash_tev_register_color	(	gx_tev_io_t	reg,
		int	r,
		int	g,
		int	b,
		int	a )

extern

Set the value of one of the TEVs internal registers.

The TEV's internal registers can be used to store constant colors if wanted. Only GX_TEV_IO_REGISTER_0, GX_TEV_IO_REGISTER_1, and GX_TEV_IO_REGISTER_2 are valid.

Parameters

color	Register to put it into.
r	Red value [0, 255], or a signed 11 bit value [-1024, 1024]
g	Green value [0, 255], or a signed 11 bit value [-1024, 1024]
b	Blue value [0, 255], or a signed 11 bit value [-1024, 1024]
a	Alpha value [0, 255], or a signed 11 bit value [-1024, 1024]

◆ gx_flash_tev_stage()

void gx_flash_tev_stage	(	gx_tev_stage_id	id,
		const gx_tev_stage_t *	tev )

extern

Flashes a TEV stage onto the BP.

Copys TEV stages onto the BP.

Parameters

id	TEV stage id.
tev	Pointer to TEV stage.

◆ gx_initialize_tev_stage()

void gx_initialize_tev_stage ( gx_tev_stage_t * tev )

extern

Generates a empty TEV stage.

It will just pass rasterizer/color value.

Parameters

tev	TEV stage to set.

◆ gx_set_tev_stage_alpha_biasing()

void gx_set_tev_stage_alpha_biasing	(	gx_tev_stage_t *	tev,
		bool	subtract,
		gx_tev_bias_t	bias,
		gx_tev_scale_t	scale )

extern

Sets the TEV alpha stage into biasing mode.

In biasing mode, the formula is like (mix(a, b, c) +- d + bias) * scale This in contrary to compare mode where the formula is d + (a compare b ? c : 0)

Parameters

tev	TEV stage to modify.
subtract	If set, the d value will be subtracted
bias	Bias value, -0.5, 0, 0.5.
scale	Scale. 0.5, 1, 2, or 4

◆ gx_set_tev_stage_alpha_comparison()

void gx_set_tev_stage_alpha_comparison	(	gx_tev_stage_t *	tev,
		gx_tev_compare_t	comparison )

extern

Sets the TEV alpha stage into comparison mode.

In biasing mode, the formula is like (mix(a, b, c) +- d + bias) * scale This in contrary to compare mode where the formula is `d + (a compare b ? c : 0)

Parameters

tev	TEV stage to modify.
comparison	Comparison mode to use.

◆ gx_set_tev_stage_alpha_input()

void gx_set_tev_stage_alpha_input	(	gx_tev_stage_t *	tev,
		gx_tev_io_t	a,
		gx_tev_io_t	b,
		gx_tev_io_t	c,
		gx_tev_io_t	d )

extern

Sets the color inputs to a TEV stage.

Sets the alpha inputs to a TEV stage

In this you can not use GX_TEV_INPUT_ONE and GX_TEV_INPUT_HALF.

GX_TEV_INPUT_ALPHA will not do anything.

Parameters

tev	TEV stage to modify.
input_a	Input A's alpha value
input_b	Input B's alpha value
input_c	Input C's alpha value
input_d	Input D's 11 bit signed value

◆ gx_set_tev_stage_alpha_output()

void gx_set_tev_stage_alpha_output	(	gx_tev_stage_t *	tev,
		gx_tev_io_t	out,
		bool	clamp )

extern

Sets the alpha output to a TEV stage.

Sets the alpha output to a TEV stage

Parameters

tev	TEV stage to modify.
out	Output Destination
clamp	Clamp the output [0, 255] or else [-1024, 1023]

◆ gx_set_tev_stage_color_biasing()

void gx_set_tev_stage_color_biasing	(	gx_tev_stage_t *	tev,
		bool	subtract,
		gx_tev_bias_t	bias,
		gx_tev_scale_t	scale )

extern

Sets the TEV color stage into biasing mode.

In biasing mode, the formula is like (mix(a, b, c) +- d + bias) * scale This in contrary to compare mode where the formula is d + (a compare b ? c : 0)

Parameters

tev	TEV stage to modify.
subtract	If set, the d value will be subtracted
bias	Bias value, -0.5, 0, 0.5.
scale	Scale. 0.5, 1, 2, or 4

◆ gx_set_tev_stage_color_comparison()

void gx_set_tev_stage_color_comparison	(	gx_tev_stage_t *	tev,
		gx_tev_compare_t	comparison )

extern

Sets the TEV color stage into comparison mode.

In biasing mode, the formula is like (mix(a, b, c) +- d + bias) * scale This in contrary to compare mode where the formula is d + (a compare b ? c : 0)

Parameters

tev	TEV stage to modify.
comparison	Comparison mode to use.

◆ gx_set_tev_stage_color_input()

void gx_set_tev_stage_color_input	(	gx_tev_stage_t *	tev,
		gx_tev_io_t	a,
		gx_tev_io_t	b,
		gx_tev_io_t	c,
		gx_tev_io_t	d )

extern

Sets the color inputs to a TEV stage.

Sets the color inputs to a TEV stage

Setting the GX_TEV_INPUT_ALPHA bit on a input, will instead use the alpha channel for that input.

Parameters

tev	TEV stage to modify.
input_a	Input A's RGB8 value
input_b	Input B's RGB8 value
input_c	Input C's RGB8 value
input_d	Input D's 11 bit signed value

◆ gx_set_tev_stage_color_output()

void gx_set_tev_stage_color_output	(	gx_tev_stage_t *	tev,
		gx_tev_io_t	out,
		bool	clamp )

extern

Sets the color output to a TEV stage.

Sets the color output to a TEV stage

Parameters

tev	TEV stage to modify.
out	Output Destination
clamp	Clamp the output [0, 255] or else [-1024, 1023]

◆ gx_set_tev_stages()

void gx_set_tev_stages ( int count )

extern

Sets the number of tev stages.

Sets the number of TEV stages the hardware will process.

Parameters

count TEV Stage count 1 to 16

Classes

Enumerations

Functions

Detailed Description

Function Documentation

◆ gx_flash_tev_register_color()

◆ gx_flash_tev_stage()

◆ gx_initialize_tev_stage()

◆ gx_set_tev_stage_alpha_biasing()

◆ gx_set_tev_stage_alpha_comparison()

◆ gx_set_tev_stage_alpha_input()

◆ gx_set_tev_stage_alpha_output()

◆ gx_set_tev_stage_color_biasing()

◆ gx_set_tev_stage_color_comparison()

◆ gx_set_tev_stage_color_input()

◆ gx_set_tev_stage_color_output()

◆ gx_set_tev_stages()