Official v645 Thread

[nicober]

Good Afternoon!

Hello Mr Ross, I have tried your solution to add SignalLightText ( "" ) to all SignalType (2) without this line in the sgccfg.ad file.
Unfortunately when I launch the Quebec - Charlevoix route I have always this error message "Value cannot be null" and return to the main screen OR.

I don't know if I understand corretly your message. I add the last OR Log Text and the sigcfg.dat file corrected by me at your attention.

Thank for your attention

Bye!

[longiron]

Yes, it's a Toshiba Satellite A205, ha! Not too old, handles MSTS fairly well, averaging around 20 FPS on most routes.

Your Toshiba Laptop has Intel Core 2 Duo Processor T5500 1.66GHz and an Intel Graphics Media Accelerator 950 with 8MB-256MB dynamically allocated shared graphics memory.

What does that mean in plain English ... it will run MSTS fairly well as MSTS was designed to use the CPU for almost everything and the video card just rendering what the CPU told it to show. This technique requires very minimal video card horsepower as was the state of PCs when MSTS was coded up.

Open Rails, on the other hand, takes a more balanced approach. Open Rails uses the CPU the manage the route items, ENGs, WAGs to display and tells the video card -- "you figure it out". As a result, Open Rails needs much more video card power to fluidly display the game. The upside of this approach is with the right HW, studdering and other things that plagued MSTS are eliminated.

[muskokaandtahoe]

Dave, just so I’m sure we’re talking about the same shadow, in the first two screen shots on the previous page the shadow is above the B unit in the first screen (v645) and right above the A unit in the second screen (v360).

It didn’t look like a bug to me but an overall blurring like you mentioned in your response with all of the shadows in v645.

Yes, I see now I misunderstood what I was looking at. You are correct... it's blob shadows on both.

There is a bug out there somewhere that fails to cast a proper shadow from some trees. The shadow looks like it came off a pole instead of a tree. Not sure what is going on and AFAIK it has not been fixed yet.

[twpol]

Hello Mr Ross, I have tried your solution to add SignalLightText ( "" ) to all ...

Is that a typo in your post? It must be SignalLightTex - no "t" at the end.

[csantucci]

First of all, again my sincere thanks for the wonderful job and the underlying effort done by the ORTS dev team!
My first notes:
- I wish that the key assignment management will become something like that of MSTS, where the key assignment can be easily modified by the user; I would like to point out my case: I have an Italian notebook: well, only to make an example, in this ORTS version train brake increase is controlled by key " ' " , that I have on the second row of the keyboard, while train brake decrease is controlled by keys " Shift+, ", where the comma is on the 6th row of the keyboard (that where also the space key lies). Dynamic brake increase is controlled by key " , ", while dynamic brake decrease is controlled by key " . " on the numpad. It's really complicated to drive a train this way.
- I read someone complained about not working horns; I think I have found the problem; if at the beginning of the horn stream the Volume statement is set above 1.0, the stream is played at very low volume (practically muted). Volume parameter in a stream should be allowed above 1.0, provided that multiplying the Volume parameter of the Scalability level and the volume parameter of the stream you get a value not higher than 1.0. For example, Scalability level volume equal 0.7 and stream volume level equal 1.25 is correct (product equal to 0.875 )
- The trigger action SetStreamVolume ( volumelevel ) is still not implemented (for example
Variable_Trigger ( Variable2_Inc_Past 02 SetStreamVolume ( 0.30 ) )
)
I remind that such action overrides the Volume parameter on top of the stream, if available, and contributes to the whole volume level the same way as the volume parameter on top of the stream, that is
Resulting volume level = Scalability level volume * SetStreamVolume level * Volume level provided by the volume curve.
- The hidden signals linked to a signal common to more tracks are shown as the posts that you see in this picture

however nothing should be shown
- some cab controls and some signal parts provide for errors in the log file, see annex
OpenRailsLog_Bernina.zip
- this is a log file of a Bernina activity that does not load correctly. It loads correctly in version 360
OpenRailsLog_971.zip
- this is a log file of a Bernina Explore Route that does not load correctly, with a different error:
OpenRailsLog_Pontresina.zip .

I like seeing signal aspects and carspawners, although I understand this is only a small part of the job you have done!
I too miss the distant mountains!

Carlo

[GoGoran]

To confirm the high sound process values I get, I took two screenshots on my test route (GR Biggetal 2006 edition) with the same test train I used for the DynamicShadows test.

The first picture is with Sound slider at 5, plus BinPatch Sound - which results in 90 - 100% constant sound process load.

http://img199.imageshack.us/img199/9605/orwithsoundslider5plusb.jpg

The second is with sound slider at 0 and BinPatch box unchecked. 0% sound process there.

http://img130.imageshack.us/img130/8649/orwithoutsoundslider0.jpg

Comment: On the flat wagon the reason why I digg OR - high-poly models of tanks, trucks and other stuff, which just won't load into MSTS (not without advanced work on the shape file hierarchy, at least).

Apart from the high sound process load, the sounds don't play back as they should (e.g. very loud crossing sound) and many sounds are not loaded or at least played back by OR v0.645 at all.

[conrail1]

All of it can and probably will change as we gain futher understanding of how the various technique interact, what peoples' systems can actually support and where people want the rendering enegery focused.

With the simulation of train handling physics, A.I., rail traffic control, better sound (among other things) yet to be implemented, I would hope that the focus would be on putting as much rendering work on the GPU as possible.

There is a bug out there somewhere that fails to cast a proper shadow from some trees. The shadow looks like it came off a pole instead of a tree. Not sure what is going on and AFAIK it has not been fixed yet.

I missed that one; I’ll have to take a closer look.

Off the shadows/graphics subject I’ve also noticed that the occasional long consist stutter is still present with v645, but not to the extent that I’ve seen it in v360.

[abrahamx]

Greetings, here are some screenshots Open Rail v645. I have not had any problems, except it is very slow

[longiron]

Greetings, I have not had any problems, except it is very slow.

Please let us know your system specs, details on your video card are important
Thanks

[twpol]

Apart from the high sound process load, the sounds don't play back as they should (e.g. very loud crossing sound) and many sounds are not loaded or at least played back by OR v0.645 at all.

There's a new sound implementation in the works that hopefully will improve the volume issues; the missings sounds might be related to the high Sound Process you have but until we get to the bottom of that it's hard to say.

With the simulation of train handling physics, A.I., rail traffic control, better sound (among other things) yet to be implemented, I would hope that the focus would be on putting as much rendering work on the GPU as possible.

I think perhaps you misunderstood; I meant what visual features and effects we should put the GPU to, not where the actual work is carried out - that is always biased as much as we can manage to the GPU over the CPU. There's always more we can do though.

[GoGoran]

There's a new sound implementation in the works that hopefully will improve the volume issues; the missings sounds might be related to the high Sound Process you have but until we get to the bottom of that it's hard to say.

I don't know, but when sounds are missing i.e. not loaded or played back by OR, the sound process figure drops. E.g. on the original NEC route with the original Acela Express, several sounds are missing, and the sound process figure there is between 20 - 30% !

[conrail1]

I think perhaps you misunderstood; I meant what visual features and effects we should put the GPU to, not where the actual work is carried out - that is always biased as much as we can manage to the GPU over the CPU. There's always more we can do though.

Yes I did, got it now.

Do you know if the long consist stutter was addressed in v645 and what causes it?

[213 001-1]

OR logs for any and all crashes please.

Where are these logs saved?

[mnjrr]

I think on your desktop.

[kox]

Probably it was mentioned before, but i don't know:
-sometimes, the road level crossing instantly stops the train
-can't the render process work in more thread than one in dx9?

As you say that openrails is slower than v360 with dynamic shadows because it uses more gpu.
But now watching around on a station with dense scenery, i got 20 fps, but the gpu usage is only between 35-47%, and the cpu usage in same time is ~45% (dual core)