[RTP]

The parabole has an horizontal asintote, that limits the vertical motion.
Of course every particle must have an initial, random, velocity vector, dispersing the plume.
At this vector, can be added the wind vector, if any.
Also, every particle must have a life, composed of a constant, big for smoke, lower for steam, plus a random component.
That is for obtain a gradual dispersion of the plume.

Again:
Please excuse my poor english.

Regards.

[rdamurphy]

James Ross, on 20 January 2014 - 06:51 AM, said:

The particle emitter gets given two key values from the simulation: volume (m^3/s) and hole size (m^2). I found that emitting particles at the seemingly-logical speed of volume/hole size and decelerating over a couple of seconds resulted in them rising a km or more and other obviously wrong situations. So, I would like to find out whether these inputs are correct and it is the particle simulation that's off, or if the inputs are bogus.

FWIW, it currently looks like a bit of both, but I'll be constructing a new particle simulation with much faster deceleration which should help.



It looks like the smoke decelerates much faster than I was anticipating (although I still think we're getting a higher velocity than might actually be the case in reality), so hopefully when I create the new model of that we'll get lower smoke plumes. I'll be adding in the wind too, though that seems less useful than more randomness in particle direction (which I'll also be adding).

Lower smoke plumes result from simply changing the value of the coordinates in the .eng file for the locomotive. Same for smoke rise during movement.

Basically, the Offset is the location of the stack, or the beginning point. The other value is the "target" for the particles. If it's 20 meters up and the particle only lives for a couple of seconds, it will never make it, and will just go pretty much straight towards it. If it's lower, say 5 meters, the plume will be much shorter and the smoke will stream backwards much more quickly on movement.

To see this in action, and to examine the settings, compare the .eng files between the default D9, which has low fat smoke with a wide exhaust and the default SD40-2 with a high thin smoke with a narrow exhaust. Try running the two together in the sim, and you can see the difference in the effects real time, then compare the values.

James, I spent a couple of weeks in the particle emitter, and I fully understand what you're trying to do, and basically, what I couldn't figure out was how to get the particles to decelerate as they rose, and also, the distance should also be more random. However, much more than that won't work if hard-coded into the particle emitter. Try making some radical changes to the .eng files and you'll see what I mean.

I also have an idea that I tried out that randomizes the smoke quads slightly so it doesn't look so uniform. Would you like to try that?

There are some MSTS locos with remarkably realistic smoke in OR, others look terrible. Usually the same results in MSTS.

Robert

[James Ross]

rdamurphy, on 21 January 2014 - 12:17 AM, said:

James, I spent a couple of weeks in the particle emitter, and I fully understand what you're trying to do, and basically, what I couldn't figure out was how to get the particles to decelerate as they rose, and also, the distance should also be more random. However, much more than that won't work if hard-coded into the particle emitter. Try making some radical changes to the .eng files and you'll see what I mean.

I also have an idea that I tried out that randomizes the smoke quads slightly so it doesn't look so uniform. Would you like to try that?

Well, the deceleration bit has been solved. :) I need to tune the behaviour but it's there. I am thinking about ways to differentiate smoke/steam/diesel but for the moment it'll be just the colour, volume emitted and duration. Let's wait and see how it looks with a load more randomness.

[James Ross]

Here's a few WIP shots of added randomness to the smoke:

[copperpen]

I have been testing the look of the steam emission. Having tweaked the position and diameter of the plume to sit right on the locomotive, the actual plume looks quite good. Not too much vertical movement, a nice curve to the smoke/steam which could do with a gradual flattening as speed increases. Now we come to the difference between steam and diesel. A diesel exhaust is driven by the throttle setting and the load on the prime mover. A steam exhaust is driven by the cut-off percentage and the speed of the locomotive.

The current rate of steam exhaust gives a series of puffs at a slow rate, this should be linked to the train speed in order to look better. The decay rate of the particles is also too slow in the current iteration of the code. At rest the visual effect stops just above the emission point and when running is not visible long enough. I realise that this is in the early stages of development.

James. That last picture looks very good, but illustrates the short time decay of the particles.

[rdamurphy]

James, I had two ideas I tried, if you'd like, I'll add them in, one was a very slight randomness to the size of the smoke particles, and the other was lowering the height point of the particles in proportion to the speed of the locomotive.

If you'd like, I'll go ahead and throw those in, so you can see what you think?

Your work has definitely improved the appearance!

Robert

[gpz]

I feel the randomness of the vertical movement of particles at decay phase should not be symmetrical. There is the sucking effect of the turbulence created by the train below the particles, but nothing such is above. So the particles will move down with a bit higher chance, than up, I think. And I also think the lifetime should be longer, in real world the vapour is often still visible, when the train has already passed. But the current smoke looks a lot better than before!

[James Ross]

copperpen, on 21 January 2014 - 01:38 PM, said:

I have been testing the look of the steam emission. Having tweaked the position and diameter of the plume to sit right on the locomotive, the actual plume looks quite good. Not too much vertical movement, a nice curve to the smoke/steam which could do with a gradual flattening as speed increases. Now we come to the difference between steam and diesel. A diesel exhaust is driven by the throttle setting and the load on the prime mover. A steam exhaust is driven by the cut-off percentage and the speed of the locomotive.

The current rate of steam exhaust gives a series of puffs at a slow rate, this should be linked to the train speed in order to look better. The decay rate of the particles is also too slow in the current iteration of the code. At rest the visual effect stops just above the emission point and when running is not visible long enough. I realise that this is in the early stages of development.

James. That last picture looks very good, but illustrates the short time decay of the particles.

The particle code doesn't get any say in the lifetime of particles - that is dictated by the locomotive simulation code, which I've not been touching. It should be looked at by someone when I stop fiddling with the actual emissions. :ireful2:

rdamurphy, on 21 January 2014 - 01:44 PM, said:

James, I had two ideas I tried, if you'd like, I'll add them in, one was a very slight randomness to the size of the smoke particles, and the other was lowering the height point of the particles in proportion to the speed of the locomotive.

Mmm, the size could be randomised (yay, more vertex data!). Regarding height/speed, I'd see how the current code works before adding something like that. It isn't clear to me how much effect speed actually has (remember that because the particles are rising at a roughly constant speed after emission, a faster locomotive means the trail will be closer at a given point down the train without any extra work) over what I have just committed as X1953.

Anyway, new code in X1953 seems to work okay and even looks alright (though probably still not great) for diesels too.

[PA1930]

It looks pretty awesome so far, James! :D

http://i39.tinypic.com/308b41c.png

May I ask something, though? I'm really not sure if this is yet possible or not, but it really depends on different locomotives if when they're in full power make a lot of smoke or not, or just when increasing throttle they do a lot of smoke and then they look rather "clean" afterwards... Is this possible to configure on the engine file? If not, would it be possible to ask for some "ORTS only" config lines which determinate how much smoke the locomotive can make in its different throttle steps or just when increasing? :) Just a thought...

[Guille592]

Tested the new smoke on the Spanish ALCO class 321... You guys over did it should I say? :D

Attached thumbnail(s)

•