740 replies to this topic

[fuzzel]

Unfortunely the VPX_4_beta_rev3262 does not start . It loads the table and crashes. Is there a log file to see what happens?

All files were updated and dll's allowed. MS .NET Framework 4.7 is installed also MS XNA Framework Redis... 3.0 and 3.1 and 4.0.

Previous Beta (VPX_4_beta_rev3259) doesen't crash. Any ideas?

If there is a crash dump you can find it in the same directory where you loaded the table.

[jimmihenry]

Found a old crash.dmp in my Tables Folder where my tables are (Duh). Deleted it and started the VPX_Beta_rev3262.

This time no new crash.dmp was created. Tired to start VPX4 with the table Nudge Test and Calibration.vpx and AFM_1.2.0.vpx.

Ended always in a crash but i can't find any crash.dmp. Tried a most recent table Xenon (Bally 1980) 1.2.5.vpx mod by hauntfreaks.

Table did pop up for a blinck of an eye, than crashed.

 

Oh almost forgot i am using PinMAME_Beta3.0.

Edited by jimmihenry, 14 December 2017 - 08:53 PM.

[hauntfreaks]

Found a old crash.dmp in my Tables Folder where my tables are (Duh). Deleted it and started the VPX_Beta_rev3262.

This time no new crash.dmp was created. Tired to start VPX4 with the table Nudge Test and Calibration.vpx and AFM_1.2.0.vpx.

Ended always in a crash but i can't find any crash.dmp. Tried a most recent table Xenon (Bally 1980) 1.2.5.vpx mod by hauntfreaks.

Table did pop up for a blinck of an eye, than crashed.

 

Oh almost forgot i am using PinMAME_Beta3.0.

just tested VPX_Beta_rev3262 with Xenon and is working on my end.... I know this doesnt fix your issue... 

Edited by hauntfreaks, 14 December 2017 - 09:02 PM.

[jimmihenry]

Thank you for you'r support!

 

It's not the table that causes thie crash it is my setup. But a good point

is that the previous versions ran swift like windhound, without any problems!

 

(It may be the PinMAME_Beta3.0 causes some mess up, may be.

It's just a gut feeling)

 

Thats my folderstructure, don't see anything odd there:

 

Edited by jimmihenry, 14 December 2017 - 10:43 PM.

[roccodimarco]

I have no idea if this is a pc issue or VPX beta issue.  But when I attempt to run SS tables in Forced fullscreen, the table loads, then just exits.  It will leave the backglass up and running however.  The only game that works is Blackhole and any EM games.  It's odd, because it was working fine, then randomly is just stopped working.  If I run it outside of forced full screen I have no problems.  I haven't been able to find a fix. 

[gtxjoe]

Exclusive full screen. Make sure the rom file registry has ddraw =0 and also d3ddraw=0

[fuzzel]

rev3264 is up:

 

- fix key handling in all manager dialogs (material, image, sound and so on)

- changelog updated

[jimmihenry]

Hello,

 

after a faceplant with the VPinball rev3262 i tried the rev3264 running into the same

odd crash. Inspired by GTXJoe had the idea to click the button "reset to defaults" in

Prferences -> Video Options. Done! This was the magic trick.

 

And it runs, it runs swift like windhound again .

 

I tried to repilcate the bug by setting my old settings, which caused the crash.

But i could't, VPin runs stable.

 

A big thanks to all involved helping me out! Also worth mention the quick tips are

awesome! Great work.

Edited by jimmihenry, 15 December 2017 - 07:04 PM.

[Thalamus]

http://vpinball.com/...e/4/#post-71434

 

Not tested this myself, but sounds like a valid report so, as a heads up ...

Edited by Thalamus, 17 December 2017 - 08:29 AM.

[fuzzel]

http://vpinball.com/...e/4/#post-71434
 
Not tested this myself, but sounds like a valid report so, as a heads up ...

Thanks I will change three files ending to pov in the next update

[fuzzel]

rev3270 is up:

 

- change POV settings file ending from .xml to .pov (to prevent issues with PinballX)

- add sam.vbs to the scripts

- update to Win32++ v8.5

- allow script to change ball spin (angular velocity and momentum) (see CommandReference.txt) (by DJRobX)

- experimental latency reduction code. A value of 600 works well on my cab. (by DJRobX)

 

Regarding the experimental latency reduction code:

Artificially lengthen the execution of the physics loop by X usecs, to give more opportunities to read changes from input(s) (try values in the multiple 100ms up to maximum 1000 range, in general: the more, the faster the CPU is)

Intended mainly to be used if vsync is enabled (e.g. most idle time is shifted from vsync-waiting to here)

If we've reached the end of the artificial delay cycle (probably about 40% of the way through), fire a "frame sync" timer event so VPM can react to input.   

This will effectively double the "-1" timer rate, but the goal, when this option is enabled, is to reduce latency and those "-1" timer calls should be roughly

halfway through the cycle, so it results in a pretty nice overall latency decrease.

 

There is also a new registry key under Player/MinPhysLoopTime (see RegistryKeys.txt) to define the minimal loop time

[Ben Logan]

Experimental latency reduction is big news. Looking forward to testing!

Can you tell us about the ballspin scripting option?

[fuzzel]

You can manipulate the ball spin values that are calculated by the physics engine. For manipulating the spin values you can change AngMomX/Y/Z (for the angular momentum) and AngVelX/Y/Z (for the angular speed). So in short you can influence the ball movement and physical behavior

[Thalamus]

10.4 is just miles above what I expected it to be coming from 10.3 !!

 

Thank you sooooo very much Toxie, fuzzel and DjRobX ... and, lets not forget the authors and artists that sees what is needed.

 

You are all awesome !

[DJRobX]

The idea behind the latency reduction stuff is to slow down the physics loop processing so it's closer to realtime.   When looking at my nudging problem I observed that the loop was processing 13-15ms worth of physics in 1ms, then rendering and waiting for VSync.   VP checks the input on each loop, but there is little benefit in checking for input 13 times back to back if you don't wait a bit until the next check.   The new option allows you to space the loops out more so you can get about 7-9 ms worth of input checking before you enter the render / vsync wait state (possibly more if your system is faster). 

 

I didn't make it a full blown option in the UI yet, I'd still like to figure out a way to make it require less manual tuning (I think I need to get the render and vsync states separated), but the result is pretty good with this simple approach on my cab so far.   Something you can play with.   If the number is too high you'll see slowdowns on heavy tables (you'll typically see it go from 60fps to 40fps due to missing vsyncs).  If it's too low the benefit reduces, of course.

[Ben Logan]

Nice loop processing detective work, DJRobX. Thanks for the explanation of the new routine. Interesting stuff.

[LeChucksBeard]

rev3270 = immediate quit to desktop with any table on my test computer. Reverting to 3235 works.

 

No crash dump. It behaves exactly as if the "q" key is pressed immediately after table load, displaying the usual quit text for the table then quitting to desktop. But VPX doesn't stay open as it normally does. Resetting the video settings to default does not fix it. 

 

Edit: I realized it actually works if I open VPX, cancel choosing a table, then open the table, then play it. Or if I double click on the table to start it that way. If I open VPX then choose a table when it prompts for one, it will quit.

 

Also no problem from PinballX.

Edited by LeChucksBeard, 17 December 2017 - 07:09 PM.

[toxie]

Nice loop processing detective work, DJRobX. Thanks for the explanation of the new routine. Interesting stuff.

 

please note that it's main use is only when vsync is active.. otherwise it can be counterproductive..

[Ben Logan]

Thanks, Toxie. I always use vsync with my cheap playfield TV in my cab, so this development is exciting stuff!

[coreduo0099]

Looks like 3270 resolves the issues I reported on the other forum...  Wow that was fast turn around!

 

Hi,

With the latest rev 3262 and 3264 when running old EM tables with the .xml table configuration, upon load I get “XML Parsing Error” in a modal.  reverting to rev3259 restores normal loading and playing without a modal for the xml pasing error.

e.g. TKO or super straight with Prospector_1977 as reference rom (same mfg/year as super straight).

Reading the changelog, it looks like the new POV feature breaks the tableconfig feature and the .xml files generated by DirectOutputConfigTester.exe table export function.  VPX is reading tableconfig files as a POV file, which really messes up the Pinball X launch experience.

I’d suggest a different file extension or something like POV as a suffix that VPX uniquely identifies on load.