Module fixed_point 

Q16.16 fixed-point math.

A saturating fixed-point number format (i32, 16 fractional bits) for pipelines that can’t afford float math (no FPU, or the accuracy/latency trade-off isn’t worth it): rasterizer transforms, angle math, scanline interpolation. Complements the q7/q15/q31/q63 aliases in crate::types, which are narrower fixed-point formats aimed at CMSIS-DSP-style signal processing rather than screen-space geometry.

• Type alias Q16, plus FP_ONE, Q16_MAX, Q16_MIN
• Conversion: f32 <-> Q16, i16 <-> Q16, q31 <-> Q16
• Arithmetic: mul_q16, mul_n_q16, mul_f_q16, div_q16, div_n_q16, div_f_q16
• Saturating: qadd_q16, qsub_q16, abs_q16
• Helpers: lerp_q16, angle_to_q16, recip_q16
• ScanlineInterp — accelerated per-scanline z + (u, v) interpolation

Structs

ScanlineInterp

Per-scanline interpolator for z-buffer depth and (u, v) texture coordinates.

Constants

FP_ONE

1.0 in Q16.16 representation.

FP_SHIFT

Fractional bit count for Q16.16.

Q16_MAX

Maximum positive value of a Q16.16 number (same as i32::MAX).

Q16_MIN

Minimum value of a Q16.16 number (same as i32::MIN).

Functions

abs_q16

Absolute value of a Q16.16 number.

angle_to_q16

Convert an angle in degrees to Q16.16 radians.

div_f_q16

Divide a Q16.16 value by an f32 scalar.

div_n_q16

Divide a Q16.16 value by a plain i32 integer. Returns 0 on division by zero.

div_q16

Divide two Q16.16 values. Returns 0 on division by zero.

from_i16_q16

Convert i16 integer -> Q16.16 (shift left 16).

from_q16

Convert Q16.16 -> f32.

lerp_q16

Linear interpolation in Q16.16: a + (b - a) * t / denom.

mul_f_q16

Multiply a Q16.16 value by an f32 scalar.

mul_n_q16

Multiply a Q16.16 value by a plain i32 integer (no fractional scaling).

mul_q16

Multiply two Q16.16 values. Uses i64 intermediate (SMULL on Cortex-M).

q16_to_q31

Reinterpret Q16.16 as Q31 by shifting left 16 bits (use i64 to avoid overflow).

q31_to_q16

Reinterpret a Q31 value as Q16.16 by shifting right 15 bits.

qadd_q16

Saturating add: clamps the result to [i32::MIN, i32::MAX].

qsub_q16

Saturating subtract: clamps the result to [i32::MIN, i32::MAX].

recip_q16

Fast reciprocal approximation: 1.0 / v in Q16.16.

to_i16_q16

Convert Q16.16 -> i16 integer (truncate fractional bits).

to_q16

Convert f32 -> Q16.16 with correct rounding.

Type Aliases

Q16

Type alias: Q16 is i32 stored in Q16.16 fixed-point format.