Static v AI.

ianmacmillan · Post by **ianmacmillan** » Mon Jan 09, 2006 3:47 pm

Just been reading an interesting tech details file included with Gavin Mclaughlan's Coal Road Route.

Among other interesting stuff was his findings on static v AI consists.

He has discovered that MSTS loads all static consists at startup, thus filling the ram space.
This can cause it to fail to load files at tile boundries as it hunts for space and can cause a crash or severe stutters.

In contrast, AI consists are only loaded just before you encounter them.

He suggests using AI wherever possible and avoiding large static items such as locos with lots of big ace files as much as possible.
These should be started just before they are encountered to hold them at signals until you have passed.

Running the route (well worth while) it is noticable that the frame rates do not dip as you pass occupied sidings.

Any comments?

jbilton · Post by **jbilton** » Mon Jan 09, 2006 4:06 pm

Hi Ian
It might explain a few problems that were found on activities for TM.
I certainly have never managed to use more than 7 statics.
However the loading of an AI, can cause MSTS to stutter.
'The Coals Road' was very advanced for its time, and Gavin Mclaughlan was/is very innovitive.
Cheers
Jon

PS The other thing is also the bind of getting the AI service to remain stationary.One bonus is steam engines will steam if they are just creeping...I think this has been discussed before though.

dforrest · Post by **dforrest** » Mon Jan 09, 2006 5:28 pm

Just a though on this. Is there anything the can be removed from the "dead" loco .eng file created by Route Riter which would give them no power and thus cause them to remain stationary when used as AI traffic?

jbilton · Post by **jbilton** » Mon Jan 09, 2006 5:36 pm

dforrest wrote:Just a though on this. Is there anything the can be removed from the "dead" loco .eng file created by Route Riter which would give them no power and thus cause them to remain stationary when used as AI traffic?

Hi David
I don't know....doesn't msts complain if it finds power as '0' in an eng file....or maybe I'm getting muddled with something else.
Cheers
Jon

jp4712 · Post by **jp4712** » Mon Jan 09, 2006 5:47 pm

If your AI has a path where it'll be stuck at a red, fine. But if not then set up a service for the AI with a performance level of 1% and if you play it right (ie get the timing right) it will do quite nicely, I would've thought. Just for a laugh I just set up a test activity and one of Simon Barnes' DMUs took about ten minutes to do just under half a mile set up in this way.

Paul

ianmacmillan · Post by **ianmacmillan** » Mon Jan 09, 2006 6:39 pm

An AI loco always stops with a gap between itself and the signal.

If you place a reversing point in that gap, the path will not reach the signal and it will stay red but the loco will wait forever to reach the reversing point.

jbilton · Post by **jbilton** » Mon Jan 09, 2006 7:25 pm

Hi
I'm with all that...but I thought the suggestion was to have yards full of AI trains?, instead of statics. Not all yards have signals.
I see David's idea ,and the AI even if it had a path, wouldn't move....but as I say I think msts doesn't like 0 power?
I remember a thread where someone was looking into having an AI moving very very slowly...so that it still steamed, I don't know if they were successful.
Cheers
Jon

BobLatimer · Post by **BobLatimer** » Mon Jan 09, 2006 9:36 pm

Stuart Williamson (scefhwil), in his Viewpoints activity for Skipton, created an invisible loco with continuous wheelslip so that is didn't move. Now a bunch of AI locos sitting in sidings with their wheels going round but not moving will look a bit silly. I thought perhaps an invisible loco with continuous wheelslip on the front of the "static AI" consist might work. I think it would work for a string of wagons but any AI loco in the string would probably want to try to move the consist. But if the loco was "dead" ($) then it wouldn't try to move (but wouldn't produce smoke either). (Forgive the rambling - just thinking it through as I type.)

So, how about this? Make up consists with an invisible loco in front with continuous wheelslip (a la scefhwil, with permission of course) and strings of wagons and $ locos behind. Put them in the traffic files as AI trains on short paths in sidings.

It still could cause an MSTS loading problem as a Player came in sight of a yard with lots of these "static AI" trains sitting around. However, it should be nowhere near the loading problem of trying to load all of the activity's loose consists if it had been done the "old" way.

BobLatimer · Post by **BobLatimer** » Tue Jan 10, 2006 2:19 am

Just had another thought.

What if the invisible loco produced smoke out of the location of the chimney of the $ loco? It would then smoke all of the time the invisible loco was running.

jbilton · Post by **jbilton** » Tue Jan 10, 2006 2:27 am

Hi Bob
All sounds very plausible to me....I'm sure someone will give it a go.
Cheers
Jon

mickrik · Post by **mickrik** » Tue Jan 10, 2006 7:42 pm

So, how about this? Make up consists with an invisible loco in front with continuous wheelslip

How do you make a loco have continueous wheelslip?
Then I will download the invisible loco and give it a try.

Cheers

Mickrik

scefhwil · Post by **scefhwil** » Tue Jan 10, 2006 9:11 pm

You can't make AI's wheelslip, because the wheel rotation is simply an animation associated with the travel speed of the AI loco. (The reason the default AI steam loco's appear to wheelslip is a bug in the sim, which RouteRiter can correct).

The concept of using non-moving AI's instead of static consists has been discussed a number of times over the years MSTS has been available. I recall it first being discussed back in the trainsimfiles days, pre uktrainsim! Problem is very few people seem to be interested in the value of it each time it's discussed so the "knowledge" soon gets forgotten.

I've never come across any way myself, nor can I recall any method proposed by someone else whereby a stationary AI has been made able to produce smoke. It seems no matter how weak or slow the AI steam loco is it has to be travelling with a visible movement before the sim decides to add the smoke. The smoke is just an animation effect.

However, I do use an invisible loco in a consist of coaches/wagons for filling up sidings. You need a path along the siding with it's end at the buffers end, and long enough to fit the full consist into. The invisble loco setup is not important because the trick is editing the consist file. You edit the old MaxVelocity line to have very low max speed and very low accelaration. I use ( 0.001 0.001 ) and this results in no appreciable movement at all.

Stuart

mickrik · Post by **mickrik** » Thu Jan 12, 2006 6:41 pm

HI Stuart,

A very interesting piece of info that on acceleration and velocity figures. I am now converting an activity that is giving frame rate problems to see what the effect will be. It takes a while to convert each consist and so far there does not seem to be any visible movement or sign that it is any different from an ordinary static. I was half expecting to see whisps of steam.

The biggest problem I have come across so far is on some sidings where there is a siding not much longer than the consist. When running the activity, MSTS will not place the AI.

There will also be a possible problem with a plethera of extra paths to be sorted logically and the consists to go with them. so far I have come up with these as examples

"BCD1#_Loose_Somewhere_siding3" for a traffic name
"BCD1#_INV_Loose_HTV_Coal_x9" for the consist.

The activity file size is getting smaller as I convert each static consist as I work my way along the activity.

I will let you know when I run with it.

Mickrik

ianmacmillan · Post by **ianmacmillan** » Thu Jan 12, 2006 7:38 pm

Remember that my Static wagons (search under my name) can be used like normal wagons in AI trains.

They are very low poly and the wheels are not animated to keep down the frame rates.

mickrik · Post by **mickrik** » Fri Jan 13, 2006 6:30 am

HI Ian,

I think that using lots of low polys is going to be the next logical part of the experiment. I was originally doing the experiment with just the stock files that came with the Burton route. Some of these are now going to have to change anyway so it will give an opportunity to try ideas out.

Mickrik