[Goku]

James Ross, on 06 October 2015 - 12:12 PM, said:

Open Rails uses only Tr_Watermark.

So it is wrong with many routes. I saw lots of world files with no trWatermatk but with different StaticDetailLevel. I don't know why if all routes are created using one RE. Maybe because many route builders indeed do world editing by hand - and it is one reason why hand editing in text editor is wrong.

[Goku]

Is here anyone good in math?

I am looking for ultimate solution for dynamic tracks flex and I think I found one:

http://i.imgur.com/ANS8gJ3.png

I have P1, P2, C and D.
I need to calculate A, R1, R2.
I have no idea how to do this ...

[Genma Saotome]

Goku, on 06 October 2015 - 12:18 PM, said:

So it is wrong with many routes. I saw lots of world files with no trWatermatk but with different StaticDetailLevel. I don't know why if all routes are created using one RE. Maybe because many route builders indeed do world editing by hand - and it is one reason why hand editing in text editor is wrong.

Just because you don't understand the reasoning does not make it wrong.

[Goku]

Genma Saotome, on 06 October 2015 - 05:26 PM, said:

Just because you don't understand the reasoning does not make it wrong.

??
If OR doesn't use StaticDetailLevel then lots of routes are broken in OR if someone uses lower 'world object density'.

[That Genset Foamer]

Goku, on 06 October 2015 - 01:17 PM, said:

Is here anyone good in math?

I am looking for ultimate solution for dynamic tracks flex and I think I found one:

http://i.imgur.com/ANS8gJ3.png

I have P1, P2, C and D.
I need to calculate A, R1, R2.
I have no idea how to do this ...

OOoh, this looks like a Horizontal curve problem in geometric design! I'll see if I can work through it later on tonight.

BTW, you'll want to insert a tangent between the horizontal curves by default. AREMA standards tend to place 30-meter minimums between curves.

[conductorchris]

AREMA standards did not exist when track was being built 150+ years ago (and it shows).
That doesn't mean placing a tangent by default is a bad idea, but I assume you mean optional!

Goku - if you are correct that your route editor will make text editing unnecessary, terrific, but your biggest help to making a truly awesome editor is listening to your users. That is not to say following their opinion without question, but keeping an open mind will help your result.

From what I can see so far, this DOES look like an awesome editor, or at least a great step forward. I am excited, admiring your work and cheering you on. But the devil is in the details and they can be humbling. Take all the help you can get!

Christopher

Oh, and by the way, I am expecting that your route editor is likely going to be the push to get me into route building.

[Genma Saotome]

Goku, on 06 October 2015 - 05:33 PM, said:

??
If OR doesn't use StaticDetailLevel then lots of routes are broken in OR if someone uses lower 'world object density'.

As seen in OR.

Display level 0
 dl01.jpg (62.35K)
Number of downloads: 28



Display level 10
 dl02.jpg (63.59K)
Number of downloads: 27

[Goku]

Genma Saotome, on 06 October 2015 - 07:57 PM, said:

As seen in OR.

You use TrWatermatks so everything is ok.
I have some routes with mixed StaticDetailLevel and TrWatermarks inside .w files.

Some progress with signals:

http://i.imgur.com/UWetPSo.png

[That Genset Foamer]

conductorchris, on 06 October 2015 - 07:01 PM, said:

AREMA standards did not exist when track was being built 150+ years ago (and it shows).
That doesn't mean placing a tangent by default is a bad idea, but I assume you mean optional!

I was thinking that, actually! Though what'd be cool is if the bearing and location of the endpoints of a flexible dynamic track section could be fixed, and the curve radius and/or added tangents affect all the parameters of the curve. Although that probably requires even more complex algorithms!!!

EDIT: Will get around to working all that mathematic stuff soon, one sec...

[That Genset Foamer]

Quick questions - what's the order of operations on the second curve? Also, could curve radius be constrained by the user (selecting a design speed or manual input), meaning that only arc and thus arc length need to be computed by the sim?

BTW I stumbled upon this. Wikibooks has a section on the math behind horizontal alignment.

With most

[Goku]

That Genset Foamer, on 06 October 2015 - 09:44 PM, said:

Also, could curve radius be constrained by the user (selecting a design speed or manual input), meaning that only arc and thus arc length need to be computed by the sim?

Here you can see example use: https://www.youtube....h?v=BH6pUjOMqdA
(I did only one curve solution that only sometimes works).

You know only position and rotation of two endpoints.

Instead of calculating R1, R2, A it will be fine to calculate P3 and something about vector V. Maybe that will be easier.
With P3 and vector V i can split two curve solution into two separate one curve solutions.

http://i.imgur.com/qKM5IQO.png

[That Genset Foamer]

Update--Found that the bearing for the tangent can be related to the bearing from the two angles. This new point where the opposing points of tangency intersect (Here, the PT of Curve 1 and the PC of Curve #2).


The first constraining equation is the one fixing the bearing of point "I", this intersection, to the known bearings.

I = C + A
I = 90 - A + D

C + A = 90 - A + D ---> 2A + C = 90 + D -----> 2A = 90 + D - C

Therefore A = (90 + D - C) / 2 is one constraining identity.

The second constraint is that the tangents must be identical between one bearing of the curve and

T₁ = R₁tan(A/2)
T₂ = R₂tan((90-A+D)/2)

The length between the PI of curve 1 and the PI of curve 2 are the sum of both tangents, plus any added tangent the user inserts.

Attached File(s)

•  CurveProblemWithTangents.jpg (19.16K)
  Number of downloads: 23

[Goku]

I think 90 - A + D on my image could be wrong.
Look here:

http://i.imgur.com/niDSfs8.png

[Goku]

Another question. Is any open source ACE compression software awailable? Something like aceit? Or I need to write all by myself?
Now I have only uncompressed rgb. No dxt and no mipmaps.

[James Ross]

Goku, on 07 October 2015 - 11:53 AM, said:

Another question. Is any open source ACE compression software awailable? Something like aceit? Or I need to write all by myself?
Now I have only uncompressed rgb. No dxt and no mipmaps.

My personal project has an ACE writer which includes compressing with DXT1 (lossy) and RGB with ZLIB/DEFLATE (lossless) in C#. License is BSD 2-clause so you should be able to do whatever you like with it (whether that's translating the code to C++ or just seeing how the formats are implemented). :)