[dcarleton]

Running the latest Beta, in #4 view, it crashes to desktop after one or two run-bys. The log shows no errors, it just suddenly crashes like MSTS. The operating system, Windows 7, lists this in the Action Center: "OpenAL Installer for Windows has stopped working properly."

UPDATE

Following the pop-up from the Windows 7 Action Center, I went to OpenAL and downloaded an exe that installed two dll files under Windows\System32. I copied those two dll files into the folder which contained OR-X, overwriting other files with the same names. Then ran the same activities as before without any crashes (so far).

Still sometimes the sound cuts out and is replaced with a buzzsaw sound momentarily. Also, in #4 view all sound cuts out after about 30 cars go by.

UPDATE

Submitted a Bug Report.

Running a 'straight out of the box' installation of beta 1624, or the latest 1626 from the trunk, it is extremely unstable. We're talking a crash to desktop without any warning, just like MSTS. Replace OpenAL32.dll and add wrap_oal.dll by copying those files from windows\system32, and it becomes very stable but the sound cuts out after about 30 cars go by. (Although if you hit the #8 key the sounds come back. Very strange.)

What am I doing wrong here? I'm using a C-Media PCI card. Maybe it's incompatible with OpenAL? How would we know?

[jtr1962]

I've had a similar problem since X1617. It's been more a case of some sounds missing rather than crashes.

[disc]

I've had a crash just now. No error in console, no any openal related message, just "openrails have stopped working", close, send report etc.
But lookin in the event logs, i see this

Error 1000:
Faulting application name: runactivity.exe, version: 0.0.4901.36307, time stamp: 0x51ab9896
Faulting module name: OpenAL32.dll, version: 0.0.0.0, time stamp: 0x51a78d76
Exception code: 0xc0000005
Fault offset: 0x000236b4
Faulting process id: 0x%9

I'm also hearing weird sounds in game (some sounds get very loud for a brief moment, sometimes sounds like really fast wind, sometimes extremely high pitched sounds), since few weeks.

See/hear here

[engmod]

I too have just replaced openal32.dll and wrap_oal.dll, due to crashes, with the latest files from Creative Labs.

openAL32.dll is now version 6.14.357.25 (the supplied version did not have a version resource)
wrap_oal.dll is now version 2.2.0.7

cheers

[engmod]

Even after replacing the dll's its still unstable, I have gone back to build 1615.

I will record some crashes (1624) and post them here.

cheers

[BB25187]

Hi all,

I also noticed the weird sounds reported by disc, but not really the crashes.
Anyway: gpz has fixed the problem of missing streames which was subsequent to revision #1617. It may be worthwile to retest with revision #1628.

Regards

[dcarleton]

 BB25187, on 03 June 2013 - 10:30 AM, said:

It may be worthwile to retest with revision #1628.

Same problem as with 1624. Using the #4 view to watch very long freight trains, or just being in the cab when a very long freight train passes by on the other track, before too long OR crashes without warning. It appears that OR can handle only so many streams, and when it reaches its limit it chokes. Shouldn't it be relinquishing streams when it's done with them? It appears to be that the cumulative number of wags and engs that OR handles, and therefore the cumulative number of streams that it handles, is the issue; and it's exacerbated by handling very long consists.

Try running a 100 car freight and repeatedly watching it in the #4 view and see what happens. Even in explore where there's only one train running, if you watch the same train run by over and over, soon it crashes. It must be activating new streams for each and every car of each and every run-by. Shorter trains take longer to crash it and very long trains with complex sms files crash much sooner.

Went back to 1615 and it does not have these stability issues. The sound cuts out after about 30 cars go by, but it doesn't crash while running for miles and miles in the #4 view. It's too bad. The quality of the sounds in 1624 are superb when it's running correctly. There are some talented people on the OR team, and hopefully they can get this all straightened out.

[gpz]

Could I ask you to try to replace the OpenAL32.dll supplied with OR with the binary distributed version of OpenAL Soft? I wrote the procedure and download link to the bug's thread. It would be good to know whether that version also have problems with its default number of sources, and also with increased number of sources. (I also wrote the way to increase them to that thread.)

It would be valuable to know if the problem is with the own compiled version, or it lies elsewhere.

[dcarleton]

Ran some experiments, with the following results:

Downloaded openal-soft-1.15.1-bin.zip. Renamed Win32/soft_oal.dll from that package to OpenAL32.dll and put it into a folder with build 1628. Then ran a 100 car train of SLI equipment on Marias 5 starting from Whitefish in the #4 view. The sound cut out after the first few cars went by, no sounds for the rest of the train including the helpers on the rear. Stayed in #4 view and it repositioned for a second runby, same as the first. Then it repositioned for a third runby, and at about the point where the sound would have cut out instead it abruptly crashed.

Next test, the SLI BNSF Seligman route activity "Bad Day for Trains", straight out of the box, no modifications. It starts with observing two AI trains leaving Seligman from outside the cab. (It is necessary to change to Manual to get that to work on OR.) The AI trains did not appear to be making any sound, or were fully overwhelmed by the sound of the idling player train. After the AI trains were out of the picture, began moving the player train while still in #4 view. By the time the locomotives had finished going by, OR had already crashed. Once again Windows 7 says: "OpenAL Installer for Windows has stopped working properly."

Then something happened that may have improved the situation. A new build was put up: 0.0.4903.22327 (2013-06-04 12:24:14Z) The next test was done using that build and with the dll provided with it.

With this one, tried running SLI BNSF Seligman "Harvest Heavyweight" which is very intensive, and staying in the cab to see if the sounds of passing AI trains cause problems. Can't hear the passing trains. After running for perhaps half an hour, the sound cut out, it went to static, and the video began to studder too. It even halted briefly. It resolved back to normal spontaneously after a while. While waiting for the AI train at the siding, stuck head out the window and observed AI trains in both directions. Neither train appeared to be making sound. Then went to #4 view and observed the player train running by repeatedly for several miles with the sound cutting in and out but no crash. The saved the act. The only warning in the log is "Warning: Sound stream activation failed at number 1024"

Resumed running the act, in #4 view the sound cut out near the end of the train, then heard static. Camera repositioned, lead locomotives went by without proper sounds and only static. Near the end of the train the sounds resolved and the helper locomotive where heard. Camera repositioned, lead locomotives went by making their proper sounds and as they were going away OR crashed. Again, the only warning in the log is "Warning: Sound stream activation failed at number 1024"

UPDATE

Interesting news. There is a new build in the /trunk/Program this time it's revision 1632. Tried with SLI BNSF Seligman "Headin to Phoenix". The player train is four units and 60 wells. Made 31 consecutive whole train run-bys without any sounds cutting out, without any static, and obviously without any crashes. The Doppler effect is superb. If this small test is any indication, then Beta 1632 is very stable indeed!

[dcarleton]

Ran some more activities with 1632 and find it to be very stable. The sounds cut out once it activates it's maximum number of streams. That might be halfway through a very long train, or it might result in not being able to hear AI trains or other sounds outside the cab. Strange that it doesn't appear to relinquish or shed streams that are moving away from the camera to make room for streams moving toward the camera but there is a way to force it. Use the number 8 (free) view, and it restarts the sounds from that vantage point. For example, if you are in the cab and you go to the 3 camera to check on the helper on the back of your train but can't hear it, try going to 8 and then back to 3 again and you should have sound.

Suggest everyone try beta 1632. Hopefully all the improvements that are in 1632 will also be in the next weekly release.