Utilizing full 3D, it is fairly straightforward. You are merely shifting & rotating the digicam relative to a textured quad or flat mesh.
Utilizing pixels solely, it is tougher. The mapping method is obtainable on Wikipedia:
Mode 7 graphics are generated for every pixel by mapping display screen coordinate $r$ to background coordinate $r’$ utilizing an affine transformation and sampling the corresponding background colour. The 2D affine transformation is specified for every scanline by 6 parameters; $a,b,c, d$ outline the matrix $M$, whereas $x_0$ and $y_0$ outline the vector $r_0$ locates the origin of the matrix transformation and is expounded to a translation vector. Particularly, display screen coordinate $r$is translated to the origin coordinate system, the matrix is utilized, and the result’s translated again to the unique coordinate system to acquire $r’$. In 2D matrix notation, that is written as:
$$r’=M(r-r_0)+r_0
start{bmatrix}
x’
y’
finish{bmatrix} =
start{bmatrix}
a & b
c & d
finish{bmatrix}
(
start{bmatrix}
x
y
finish{bmatrix} –
start{bmatrix}
x_0
y_0
finish{bmatrix}) +
start{bmatrix}
x_0
y_0
finish{bmatrix}
$$
All arithmetic is carried out on 16-bit signed mounted level numbers, whereas all offsets are restricted to 13 bits. The radix level is between bits 7 and eight.
In the event you’re trying on the historic, pixel manipulation technique, it is value mentioning that the authentic implementation had some rounding errors that resulted in visible artifacts – the output appeared considerably blocky & had visible artifacts:
The above was taken from this video clip.
As to which strategy you need to use – solely you possibly can reply that. That depends upon your time, curiosity, abilities, constraints and different issues not lined within the query & past the scope of GDSE suggestions. My recommendation could be to think about the professionals & cons of these components & select accordingly.
