Smoother realistic_acceleration

sidew · **Posted:** Mon May 08, 2006 10:35 am

try:
svn://svn.openttd.org/tags/0.4.7

or Rrevision 4781

http://svn.openttd.org/cgi-bin/trac.cgi ... tags/0.4.7

SirkoZ · **Posted:** Mon May 08, 2006 12:53 pm

Thanks.

SirkoZ · **Posted:** Sat Jul 08, 2006 11:12 pm

Nightly and Mini-IN versions of (TT-)original vehicle names and loco-speeds patches are here together.

P.S.: I've added the same patch for MiniIN r5720 which has the physics patch added, which off-sets some lines in /tables/engines.h.
It'll be added soon into MiniIN, as richk67 said. Thank you, Rich.

Papegaaij · **Joined:** Wed Oct 27, 2004 10:49 pm **Posts:** 34

I'm trying to switch from a old nightly to the latest MiniIN without breaking my entire network. My network currently has about 800 trains running on a 512x512 map with a more-or-less realistic landscape. Minor changes in train handling can cause major problems on such a crowded network. The latest MiniIN contains the new realistic acceleration (V2.0) which causes my trains to run much slower. My network is designed for the old 'realistic_acceleration = true' which makes maglev trains accelerate quite fast. However there seems to be no way of reverting to this old behaviour.

I would really like to continue with this game, but also enjoy the new features of the MiniIN, however redesigning my entire network just to make it compatible with the new acceleration is just too much work (if it is even possible). Would it be possible to still include the old (not-so) realistic behaviour?

SirkoZ · **Posted:** Tue Oct 10, 2006 9:57 pm

Yes - the old real_acceleration is still in there, just go to Patches >> Vehicles and turn OFF the RA-T (Realistic acceleration for trains) - that ought to give you my old real_acceleration patch.

By the way - in future it is better to write the inquiries about MiniIN in MiniIN thread - General OpenTTD >> MiniIN, because I only made the "old" real_acceleration patch. Some other player has written the 2.0 realistic acceleration patch.

Papegaaij · **Joined:** Wed Oct 27, 2004 10:49 pm **Posts:** 34

Thanks. I've enabled the SRA-T option, which seems to be the closest to what it was in my old version. Trains still accelerate a little slower than they used to, but now it is something I can fix in my network

hertogjan · **Posted:** Wed Oct 11, 2006 8:13 pm

Papegaaij wrote:

(...) Trains still accelerate a little slower than they used to, but now it is something I can fix in my network

I know what causes the trains to accelerate slower, and who is responsible for it. :mrgreen:

SirkoZ · **Posted:** Wed Oct 11, 2006 8:34 pm

Oh you do now, don't you?

BTW - is my patch still in (I haven't checked for a while) - please restore it into the old behaviour if possible.

Papegaaij · **Joined:** Wed Oct 27, 2004 10:49 pm **Posts:** 34

Unfortunately the difference in behaviour is too big to update my network. I'm afraid I'm stuck with my nightly

Even the smallest difference in acceleration distroys the working of all my major junctions. Isn't there any way to revert to the old behaviour?

I've just tested the difference: a fully loaded goods train (Chimaera with 12 cars) accelerates from 0 to 640km/h over a length of 33 squares with the old version. With the new version it reaches only 540km/h in 33 squares and it wasn't even able to reach 640 on my quite long (straight) test track. Maybe you can point me to the source files that have changed, so I can restore the old acceleration for maglev trains?

Papegaaij · **Joined:** Wed Oct 27, 2004 10:49 pm **Posts:** 34

Is it correct that this might be the change I'm looking for:

Code:

} else {
  //"kickoff" acceleration
  force = (mass * 12) + resistance;
}

Which is in my version:

Code:

} else {
  //"kickoff" acceleration
  force = (mass * 8) + resistance;
}

To me that seems as if the influence of the weight on the force has been increased. But as many lines have changed (the diff for train_cmd.c is 2367 lines) I'm not sure if this is the only thing that changed.

glx · **Posted:** Thu Oct 12, 2006 1:13 pm

This part is only used to make the train start (at green signal or in stations).

Papegaaij · **Joined:** Wed Oct 27, 2004 10:49 pm **Posts:** 34

glx wrote:

This part is only used to make the train start (at green signal or in stations).

I see it (only when speed=0). So it should only help a really tiny bit

From what I can see the code for the train acceleration changed quite a bit between the version that was removed in 5757 and added back in 6397. I've just tested the nightlies and it still has the old behaviour, so it is something with the MiniIN.

hertogjan, as you wrote the code, can you tell me where the change might be?

hertogjan · **Posted:** Fri Oct 13, 2006 8:51 pm

There is a conversion from km/h to mph*, which I removed since the whole game works internally in km/h, not mph. But I'm not sure how it was meant to work, since the realistic acceleration function (which still exists in trunk) is poorly coded, and on top of it, poorly documented. So it is very hard to figure out what is happening, or what should be happening; let alone: why it is happening.
That was one of the main reasons I wrote a realistic acceleration function myself.

*There should be something in the code that says

Code:

int speed = v->cur_speed; //[mph]
(...)
speed *= 10;
speed /= 16;

(Note the comment on the first of those lines.)
I removed the last two lines of those. But if anyone can prove that this is a mistake, then those lines may be readded.

SirkoZ · **Posted:** Sat Oct 14, 2006 1:25 am

Papegaaij wrote:

Is it correct that this might be the change I'm looking for:

Code:

} else {
  //"kickoff" acceleration
  force = (mass * 12) + resistance;
}

Which is in my version:

Code:

} else {
  //"kickoff" acceleration
  force = (mass * 8) + resistance;
}

To me that seems as if the influence of the weight on the force has been increased. But as many lines have changed (the diff for train_cmd.c is 2367 lines) I'm not sure if this is the only thing that changed.

Yes, that is just what it says - a kick of - if that number * mass is low, the slower a train will "start to move" or "kick-off" - it'll be visible as a "jerk".

Papegaaij · **Joined:** Wed Oct 27, 2004 10:49 pm **Posts:** 34

hertogjan wrote:

There is a conversion from km/h to mph*, which I removed since the whole game works internally in km/h, not mph.
..cut..
I removed the last two lines of those. But if anyone can prove that this is a mistake, then those lines may be readded.

I did see those lines. The lines seem to convert a km/h speed to mph (divide by 1.6). So I guess if v->cur_speed is in km/h then they are correct, if speed is supposed to be mph. However it's strange that this part of the code works with mph while the rest works with km/h. I'll try adding those lines back in and see what happens.

Papegaaij · **Joined:** Wed Oct 27, 2004 10:49 pm **Posts:** 34

Ok, that did it. Trains are running as they should again. Now I only need to change some orders due to differences between YAPF and NPF. Thanks!

hertogjan · **Posted:** Mon Oct 16, 2006 10:47 am

Maybe I will post a patch this week with a new version of that function that "fixes" all the effects caused by code changes elsewhere, i.e. it should act as it was meant to act when it was written. But first I will have to investigate what changes may have affected it.

SirkoZ · **Posted:** Mon Oct 16, 2006 1:56 pm

Beautiful!

Thx.

Papegaaij · **Joined:** Wed Oct 27, 2004 10:49 pm **Posts:** 34

That would be really great! With the conversion back in I still notice a small difference. Trains now reach 643km/h in stead of the 640 they should, but I'm not complaining about that

SirkoZ · **Posted:** Tue Oct 17, 2006 8:29 pm

Oh - don't worry 'bout 640/43 it's a purely unit-issue.

Not to mention - I did set it to 643 with my vehicle_names_powers_and_speeds patch, so it's like in the TTD.

Transport Tycoon Forums

Smoother realistic_acceleration

Who is online