To set the stage, picture this. A dense smoke starting at full opacity (1) and fading over some time period to invisibility (opacity zero). At the halfway point opacity would be 0.5, and at other times the opacity would be calculated from a straight-line, linear function.
There are two problems with this.
1) A linear fading means that the RELATIVE rate of change per unit time gets larger as time progresses. For the first 10% of the live time opacity drops from 1 to 0.9. That's a relative change of (1.0 - 0.9) / 1.0 = 0.1, or 10%. For the same 10% of live time from the halfway point, opacity drops from 0.5 to 0.4, or the same decrease of 0.1 in opacity. But the relative decrease is now (0.5 - 0.4) / 0.5 = 0.2, or 20%. From the 80% point of the live time, opacity drops from 0.2 to 0.1, or a relative change of (0.2 - 0.1) / 0.2 = 0.5, or 50%.
We see the impact of this particularly well when watching something like a water ring effect (or my bomb blast water disturbance). As it gets more transparent, toward the end of its live time it appears to approach fade-out in a very accelerating way.
2) For smoke effects which have the individual effect particles getting larger with time, the amount of overlap due to this expansion tends to increase. This overlap makes for increased opacity. The effect of this is the trend for a smoke trail to either actually increase in opacity for some period early on, or at least not fade as rapidly as it should.
For smokes or other effects having particle overlap, these two aspects only exacerbate each other.
And so instead of a linear fall-off in opacity there should be a rate that's more logarithmic. For instance, such a curve might have the opacity at the halfway point to be NOT halfway between start and finish values (0.5 when start/finish are 1.0/0.0), but instead be perhaps 0.3. Having the opacity decrease initially most rapidly, and slowing continuously thereafter, would be so much more realistic.
For rising smokes from ground objects, the stock approach has been a scheme, intentionally or by accident, to overcome this detracting aspect. And that is the use of upward vertical acceleration to cause the effect particles to become increasingly separated. While the appearance is usually improved as regards density change, the unnatural result of a smoke continuing to accelerate until the whatever GasResist value in the .eff file imposes a terminal velocity is physics-defying. (The fastest vertical ascent is usually in the first seconds, while the buoyancy is greatest, slowing thereafter as turbulent intermixing of the heated gas with surrounding air cools it down.)
And this kind of 'faking' is present for stock aircraft smokes, too. The too-low GasResist value typically employed results in the smoke particles taking some time to slow down. Which in turn acts as an acceleration on the particles to cause their separation to increase while they fall farther behind the plane. The downside here is the visual oddity of the plane appearing to move through the air rather more slowly than it is. In reality smoke essentially instantaneously slows to the ambient air environment, because the gas containing the smoke has (or will have in a tiny fraction of a second after emission) the same density as the environs. An air parcel pushed into the slipstream has not the density to get very far. Throw a balloon as hard as you can to see this; and that's for a contained parcel. Uncontained parcels are rapidly turbulently disrupted, removing momentum even more quickly.
Just the expression of a wish that likely could never be realized, while I take a little break from Brain Pain
TM, otherwise known as Java
Author
Topic: If only fading smokes could have a non-linear decrease in opacity. (Read 167 times)
