Shocks3 damping calculation is wrong?

[danmackey]

I was looking into shocks3 and the logic behind it and I think there is a bug in the damping calculations.

In the original discussion for shocks3 this graph was posted:

I figured that making something to visualize different values would help me understand tuning a bit more.

So I looked at Actor::CalcShocks3 and reimplemented it in Python so I can quickly produce a plot of the same curve.

Here I noticed the same parameters produced a different plot:

I then plotted the original logic that Dohzer suggested:

Out of curiosity, I tried plotting the calculation that's in the game, but without the last d /= v

I assume the goal was to get damping curves like the one that Dohzer posted in the Discord, meaning this is a bug?

Here's the Python code I wrote:

# pip install numpy pandas plotly

from typing import Callable

import numpy as np
import pandas as pd
import plotly.express as px

def ror_shocks3(damp: float, slow: float, split: float, fast: float):
    def curve(v: float):
        v = np.clip(abs(v), 0.15, 20)
        d = damp * slow * min(v, split) + damp * fast * max(0, v - split)
        d /= v
        return d
    return curve

def ror_shocks3_fixed(damp: float, slow: float, split: float, fast: float):
    def curve(v: float):
        v = np.clip(abs(v), 0.15, 20)
        d = damp * slow * min(v, split) + damp * fast * max(0, v - split)
        return d
    return curve

def dohzer(damp: float, slow: float, split: float, fast: float):
    def curve(v: float):
        if v <= split:
            d = damp * slow * v
        else:
            d = damp * fast * (v - split) + (damp * slow * split)
        return d
    return curve

damp = 5000
slow = 0.5
split = 1.2
fast = 0.2

curves = {
    'Shocks3': ror_shocks3(damp, slow, split, fast),
    'Dohzer': dohzer(damp, slow, split, fast),
    'Shocks3 Fixed': ror_shocks3_fixed(damp, slow, split, fast),
}

def calc_curve(curve: Callable[[float], float], title: str) -> pd.DataFrame:
    v = np.linspace(0, 3.5, 100)
    d = [curve(x) for x in v]

    return pd.DataFrame({'velocity m/s': v, 'damping': d, 'curve': title})

curves_df = pd.concat([
    calc_curve(curve, title)
    for title, curve in curves.items()
]).reset_index(drop=True)

fig = px.line(
    curves_df,
    x='velocity m/s',
    y='damping',
    color='curve',
    line_dash='curve',
    range_y=[0, 6000],
    title=f'{damp=}, {slow=}, {split=}, {fast=}',
)

fig.show()

[ohlidalp]

Only a few select mods seem to use shocks3, so this should be applicable. I'm checking it out.