Assume that the slit surfaces play a central role in the single-slit experiment. Light scatters from the two edges of the single-slit, causing the phase shift of the photons as they would in an optical filter (filters contains small holes/slits, basically). Two distinct scattered light rays in different phases from both edges reach the same single point on the screen. When doing so, light beams recombine back to phase or not-in-phase, which yields the on-off pattern with oscillating amplitude. Intensity (a measure of the amount of energy at each point on the screen per unit time) is not calculated in this example.
The diffraction pattern is calculated from the Path Length Difference between lines a and b. PLD is divided by the wavelength to get the phase n. Each whole number n depicts the on-phase. Each half-integer part is off-phase. With this information, we can derive a mathematical formula to calculate each point (x,y) in the circle and get its phase value. See ϕ_{on_x} and ϕ_{on_y} in the simulation.
This model gives a slightly different approach to the single-slit experiment since the pattern is projected to the circle. The flat projection band in this example illustrates the oscillation pattern. Light rays that actually recombine in the flat screen in a single point are demonstrated in the other GeoGebra document: https://www.geogebra.org/m/qqxg9exn
A double-slit pattern would be a trivial extension to this approach. Inner edges would cause an envelope pattern precisely the same way as in a single-slit pattern. The outer edges would case the fringe pattern inside the on-phases of the envelope.
Note: if the wavelength is bigger than the width of the slit, light also goes straight through in the middle, causing a big lump in the center of the projector screen, which covers the underlying diffraction pattern.
Suppose the pattern predicted by this model can be confirmed by accurate experiments. In that case, it could mean that the wave model, using path integrals and more complicated trigonometry in the N-slit cases, is not necessary. That would also raise new questions about the wave-particle duality dilemma, namely, if it is a duality problem at all, after all.
