It's really more of an implementation detail, but I'll explain.
Last season, with constant weather, we changed the way we computed which sky box to use, to make it based on the weather selections. The idea was to make the sky box selection deterministic based on the weather parameters. This season, for constant weather, the algorithm was updated to include relative humidity, and was also updated so that +1 in one parameter would not be offset by a -1 in another parameter.
What this means to you is probably not much!
The change makes the selection of the sky box still deterministic (so same weather -> same sky box), but the mapping from a set of weather parameters to a sky box should appear to be random.
Reproduced with permission from member Philippe L....
Let me
try to explain what it all means.
Weather 101
You will see different terms being used for weather:
dynamic, realistic, random, static, fixed, official and
default.
We like to make things complicated.
Starting this season, iRacing started using dynamic weather for all race and practice sessions.
Dynamic weather is also called "realistic" or "random".
It's actually more random than realistic, but that's another story.
For a while iRacing had to revert to fixed weather because of a bug in dynamic weather so we saw people discussing things like static and fixed weather, adding to the confusion.
Dynamic weather is now back (for good?).
To make it easier to judge lap times when all races and practice sessions are using dynamic (random) weather, all demo laps and lap time guides are based on default weather.
Weather terms glossary:
Default weather = 78ºF, 55% RH, Wind N @ 2 MPH, Partly Cloudy
Official weather = the weather set for the entire week by iRacing. This is irrelevant when dynamic weather is in effect
Fixed weather = used when
official weather is meant
Static weather = used when
official weather is meant
Dynamic weather = different weather for every session, weather even changes during a session
Realistic weather = same as
dynamic weather
Random weather = same as
dynamic weather
Effects of weather on performance:
Track temperature
Grip is mainly affected by track temperature.
Lower track temperatures mean more grip, faster cornering and quicker lap times.
The effect is quite large. A difference of 10ºF can make a difference of over 1 second on a 1:30 lap.
Track temperature only depends on air temperature and cloud cover in iRacing:
(TT = track temp)
(AT = air temp)
Clear: TT = AT + 27ºF
Partly cloudy: TT = AT + 26ºF
Mostly cloudy: TT = AT + 11ºF
Overcast: TT = AT + 5ºF
Foggy: TT = AT + 5ºF
Wind
In the Skip Barber car, wind mainly affects speed down the straights.
A headwind will make you slower, a tailwind will make you faster.
If you have a tailwind, you will probably adjust your braking point (brake earlier because you're going faster)
Air temperature / humidity / atmospheric pressure
(this is getting rather complicated so you can stop reading if you're not interested in the science behind it)
Normal (non-turbo) engine performance is affected by "dry air density" or the amount of available oxygen per volume.
The higher the density, the more power the engine can deliver.
Dry air density is a factor of atmospheric pressure, air temperature and relative humidity.
Higher pressure, lower temperature and lower humidity increase air density.
So engines perform best when atmospheric pressure is high, temperature is low and humidity is low.
They perform worst when pressure is low, temperature is high and humidity is high.
Simply put, cool and dry weather is quicker, hot and humid is slower.
By how much? The theoretical difference in performance between 65ºF and 90ºF is about 6%.
The difference between 0% humidity and 100% humidity is about 2% at 65ºF or 5% at 90ºF.
All of this at standard atmospheric pressure of 1013 hPa.
From one extreme (65ºF/0%) to the other (90ºF/100%), the difference is
10%!
It seems that the effect of temperature on engine performance is correctly modeled in iRacing.
I haven't done any tests to see if humidity is also simulated correctly. I would guess it is.
Temperature and humidity also affects drag and aerodynamic efficiency, which is less of a problem with the Skip Barber car but can make a big difference in high-downforce cars like the F1 car.
Aerodynamic performance is a factor of air density (not dry air density).
Air density is still affected by humidity, but not as much.
Still, humid air is less dense than dry air (this may seem odd, but it's true).
Air density decreases with higher temperatures and higher humidity and of course increases with lower temperatures and low humidity.
This means that cold and dry air will result in more downforce and more drag.
Last edited 4th Decemeber 2015. See Philippe's
thread for any recent updates.
iRacing staff (David T)
noted 2 points:
First, at night tracks the track temp equals the
air temp (no sun shine), and second air density is affected by track
altitude as well, so a sea level track will be faster than a mountain
track because of increased air (oxygen) and therefore increased engine
power. How that gets offset by increased drag, I don't know.
Also,
not to mix things up too much, but when we get Dan's new track heat
model out to you then track temp's will be affected by sun angle and
shadowing, along with the driving line, so the time of day will suddenly
matter as well.
Philippe
explained:
I didn't include night because we (Skippies) don't race at night
As for the altitude of the track, I did mention that engine performance and aerodynamics is affected by atmospheric pressure, and of course that changes with altitude. Since a track doesn't move higher or lower from one race to the next (just imagine), the original post was mainly intended to explain the effect of weather changes on performance.
More insight from David T (
posted in same thread 3rd Sept 2015):
We actually model the cooling of the track by radiating heat into the air, if the air is cooler than the track surface. Since this only happens if the cloud cover changes, or the sun sets, it has no real impact on the sim at this time. I think some of our 'dusk' races actually start off with a hotter track surface than the air temp, so maybe in that case you can experience it.
Once the tires start adding heat you will get heat radiating into the air, especially in shaded areas of the track where the tire heat has more impact.
And of course all of this work is a build up to day/night transitions. I have no idea how long that will take but we are actively working on it.
Here is a short list of things we have and need, keep in mind that this is a gross simplification.
Working:
- Dynamic shadows based on sun angle
- Track surface sensitive to shadows
- Dynamic weather (for better or worse) including wind
- Track banking creates 'wind shadows' so cooling effects are dynamic across the surface. Buildings may do this as well. Chances are there is more to do here, but I'm out of the loop. In the past the wind just blew through everything, but not any more.
- Day and night lighting (night only at some tracks) and the piles of shaders needed to make that work.
Not yet:
- Dynamic sky box, right now we have a small collection of static sky boxes and the sun angle is set to match the box
- Dynamic cloud cover with shadows, we have 'haze' but no cloud shadows.
- Day to night transition, things get complicated at sunset
- More reasonable weather that changes with time of day, winds are calm just after sunrise and get really wild around 2 in the afternoon when the ground starts to radiate heat (thermals) back into the sky. Right at sunset you get into an inversion where the ground is much hotter than the air, that affects everything including sounds.
Basically if we could generate the sky box on the fly we could simulate a full day from just after sunrise to just before sunset. We can already simulate a whole night for what that is worth Actually I don't think the moon cast a shadow at night, that may need to be on the to-do list.
