4027 replies to this topic

[Killshadow66]

Hey all as a person that just got back into the hobby MAN HAS THINGS CHANGED lol.  I remembered I used to love Visual Pinball because I could play Black Knight.  So I installed a new version of 10.6 per Terry Red's awesome instruction and everything worked great.  I installed 10.7 last night I didn't get to play all of my tables but the ones I did play seemed to play better, faster and I thought the graphics looked a bit better.  All in all the change to 10.7 went smoothly and I noticed no problems in any tables old school EM or State of the art.  Just thought if you were on the fence this would help.  Always a fan and very grateful for all the help and the community.

 

CJ

 

aka Killshadow

[Wylte]

@wiesshund I did tests on your settings on UT, and wasn't seeing it anymore.  But then I reverted to "High" settings and it wasn't doing it anymore either...so I tested Goonies, not there...
Then I remembered that my wife had me install Webroot antivirus while trying out a new pirate source; I normally disable it, but it is on at startup and I'd restarted yesterday...

 

So that's the issue.  Seems that something 338 does that 310 doesn't is triggering active scanning by the antivirus.  It's not getting flagged, but it's doing something that the antivirus thinks is worth interrupting to check...over and over.  (Edit: Excluding VPX in the AV has no effect, it still kills performance on 338 and not 310.)  This is worse on tables that rely on the code in VPX more heavily, i.e. originals and EM's, since roms offload most things to VPM to run.

 

Just for reference, here's the comparison on usage between 338 running normally and Webroot thinking it's sus AF:

 

https://drive.google...iew?usp=sharing

https://drive.google...iew?usp=sharing

Edited by Wylte, 02 November 2021 - 06:43 PM.

[fuzzel]

I think it's vbs in general that your strange AV is checking all the time. Advanced originals like Goonies do sometimes heavy stuff in the script so that might be the reason why rom based games don't show any slowdowns.
The golden rule when playing VP is: disable any AV and you should be fine.

[sahesilva]

Tenho erro quando joga a bola na rampa do atack from Mars ,na rampa da vaca ,as vezes a bolinha vai ao contrário da rampa

[Wylte]

I think it's vbs in general that your strange AV is checking all the time. Advanced originals like Goonies do sometimes heavy stuff in the script so that might be the reason why rom based games don't show any slowdowns.
The golden rule when playing VP is: disable any AV and you should be fine.

Yeah, I usually follow that rule (-_-")

Just weird that it's not an issue with any 10.7 rev I've tried (had the AV since...249ish?) until 338. Maybe the flipper code fix just looks sus.

Anyway, you guys are awesome. I'll keep trying the beta releases and reporting what I run into, and I'll keep trying to make sure my feedback is at least a little helpful!

[wiesshund]

Yeah, I usually follow that rule (-_-")

Just weird that it's not an issue with any 10.7 rev I've tried (had the AV since...249ish?) until 338. Maybe the flipper code fix just looks sus.

 

Admin Powershell window

Add-MpPreference -ExclusionProcess "c:\emulation\VPX\VPinballX.exe"
Add-MpPreference -ExclusionProcess "c:\emulation\VPX\VPinball 10.7.4360.exe"
Add-MpPreference -ExclusionProcess "c:\emulation\FP\BAM\Fploader.exe"
Add-MpPreference -ExclusionProcess "c:\emulation\FP\Future Pinball.exe"
Add-MpPreference -ExclusionPath "c:\emulation\VPX\Tables\*.vbs"
Add-MpPreference -ExclusionPath "c:\emulation\VPX\Scripts\*.vbs"
Add-MpPreference -ExclusionPath "c:\emulation\FP\Scripts\*.vbs"
Add-MpPreference -ExclusionPath "c:\emulation\VPX\VPinballX.exe"
Add-MpPreference -ExclusionPath "c:\emulation\VPX\VPinball 10.7.4360.exe"
Add-MpPreference -ExclusionPath "c:\emulation\FP\Future Pinball.exe"
Add-MpPreference -ExclusionPath "c:\emulation\FP\BAM\Fploader.exe"

Edit to meet your own paths etc of course

If you use windows defender, that should make it leave any VPX and FP stuff alone
And yes, the executables need to be excluded as both process and path like that

 

You might try that?

@wiesshund I did tests on your settings on UT, and wasn't seeing it anymore.  But then I reverted to "High" settings and it wasn't doing it anymore either...so I tested Goonies, not there...
Then I remembered that my wife had me install Webroot antivirus while trying out a new pirate source; I normally disable it, but it is on at startup and I'd restarted yesterday...

 

So that's the issue.  Seems that something 338 does that 310 doesn't is triggering active scanning by the antivirus.  It's not getting flagged, but it's doing something that the antivirus thinks is worth interrupting to check...over and over.  (Edit: Excluding VPX in the AV has no effect, it still kills performance on 338 and not 310.)  This is worse on tables that rely on the code in VPX more heavily, i.e. originals and EM's, since roms offload most things to VPM to run.

 

Just for reference, here's the comparison on usage between 338 running normally and Webroot thinking it's sus AF:

 

https://drive.google...iew?usp=sharing

https://drive.google...iew?usp=sharing

 

OH now i see this LOL

 

Well the EXE has a different sig than the previous
it may be doing something different, maybe

It isnt Defender, but can you tell it to exclude files, paths and processes?
i would have it exclude all of VPX's files, the entire path, and the exe's as a process and the VBS scripts as well.

 

Or tell Le Wife to stop pirating  LOL

[wiesshund]

Revision 338 BUG

 

Upon script errors

Mouse cursor is lost

 

Can point at anything else in windows but VPX, after you click stop or continue

once you click stop or continue you can no longer point at or click any part of VPX

No cursor at all present, can not even blind click

[GlueSniffinEd]

 

 

rev338 is up (Release Candidate):

 

- fix issue for negative start/end angle of flippers, as we cannot predict what the script will do with the flippers after player started

- cleanups

-  fix primitive collision regression

- improve cache usage/memlatency for physics engine

- reduce memory usgae in the physics engine

- fix tab order

 

 

Cool
Do you know when you might get to take a look at the nudge routine?
no matter what way you nudge, the table only nudges up
the apron indicators show it correctly, but the balls do not react correctly, they only nudge up
and you have to crank the nudge values up insanely high to even get the balls to move at all

 

Load nudge table in 10.6.2 and then load it in 10.7 and you will see what i mean
it was ok say about 10 or 15 revisions back 

 

I've got a bare Pinscape setup in my cab and just upgraded to 338.  Nudging is definitely off; I CAN nudge and move the ball left and right, but after a very small threshold, it slams it upward hard.  If I nudge like I normally did on 310, I now tilt most tables.

 

Not a complaint to the DEV team, I appreciate everything you guys do here, just confirmation of a bug report.

 

Thanks again fuzzel!

[wiesshund]

 

 

 

rev338 is up (Release Candidate):

 

- fix issue for negative start/end angle of flippers, as we cannot predict what the script will do with the flippers after player started

- cleanups

-  fix primitive collision regression

- improve cache usage/memlatency for physics engine

- reduce memory usgae in the physics engine

- fix tab order

 

 

Cool
Do you know when you might get to take a look at the nudge routine?
no matter what way you nudge, the table only nudges up
the apron indicators show it correctly, but the balls do not react correctly, they only nudge up
and you have to crank the nudge values up insanely high to even get the balls to move at all

 

Load nudge table in 10.6.2 and then load it in 10.7 and you will see what i mean
it was ok say about 10 or 15 revisions back 

 

I've got a bare Pinscape setup in my cab and just upgraded to 338.  Nudging is definitely off; I CAN nudge and move the ball left and right, but after a very small threshold, it slams it upward hard.  If I nudge like I normally did on 310, I now tilt most tables.

 

Not a complaint to the DEV team, I appreciate everything you guys do here, just confirmation of a bug report.

 

Thanks again fuzzel!

 

Same thing i got, tests out the same on coneheads virtualpin cab also
No nudge until you crank the settings really high, and then you risk a tilt.

if there is any sideways nudge, it must be terrible small
slam nudge full let or right, balls all go UP, never left never right, just UP

 

Digital (Key/Button) nudge i think works, it seems to anyways?

Not 100% sure though

 

Luckily for me, i stick so bad at nudging that i just dont.

[Gravy]

Hi Toxie and Fuzzel,

Not sure if either of you have seen the following thread regarding Anaglyph 3D support? https://www.vpforums...showtopic=48019

Pretty sure you guys are now in a feature freeze for 10.7 but thought I would put in a suggestion for native anaglyph 3D to be added to the list of 3D Stereo modes at some point in the future if you guys are able? I've discovered that anaglyph 3D can currently be done by the end user with Reshade and Superdepth 3D shader but there is a problem with rotated screens as windows needs to be set to portrait for it to work and also the reshade shader takes quite a bit of a framerate hit compared to the inbuilt VPX 3D rendering modes. I suspect it would be more framerate friendly to build it directly into VPX?

There are not too many 3D TV screens around anymore so perhaps Anaglyph 3D is a viable alternative for all users (cheap passive glasses that work with any screen). Green/Magenta 3D and also Blue/Amber 3D lenses seem to give better colour reproduction from what I have read, compared to the old red/blue glasses.

Anyway, thought I'd throw the idea into the ring. A huge thankyou for all your efforts.

[wiesshund]

Hi Toxie and Fuzzel,

Not sure if either of you have seen the following thread regarding Anaglyph 3D support? https://www.vpforums...showtopic=48019

Pretty sure you guys are now in a feature freeze for 10.7 but thought I would put in a suggestion for native anaglyph 3D to be added to the list of 3D Stereo modes at some point in the future if you guys are able? I've discovered that anaglyph 3D can currently be done by the end user with Reshade and Superdepth 3D shader but there is a problem with rotated screens as windows needs to be set to portrait for it to work and also the reshade shader takes quite a bit of a framerate hit compared to the inbuilt VPX 3D rendering modes. I suspect it would be more framerate friendly to build it directly into VPX?

There are not too many 3D TV screens around anymore so perhaps Anaglyph 3D is a viable alternative for all users (cheap passive glasses that work with any screen). Green/Magenta 3D and also Blue/Amber 3D lenses seem to give better colour reproduction from what I have read, compared to the old red/blue glasses.

Anyway, thought I'd throw the idea into the ring. A huge thankyou for all your efforts.

 

I would happily take a beta of red / cyan
Only problem with amber / blue is, it is proprietary, it is owned by Color3D
So you can only get the glasses through them, in paper type only
And i do not know what is entailed in making something that uses it, and not putting yourself at odds with their legal rights
Maybe nothing, but i do not know the answer there

 

Plus is it was simply left open, to define a color value
Like entering in and arbitrary RBG Hex value, or moving some color sliders
   well, you can not sue a private individual from typing in yellow right?

[Gravy]

 

Hi Toxie and Fuzzel,

Not sure if either of you have seen the following thread regarding Anaglyph 3D support? https://www.vpforums...showtopic=48019

Pretty sure you guys are now in a feature freeze for 10.7 but thought I would put in a suggestion for native anaglyph 3D to be added to the list of 3D Stereo modes at some point in the future if you guys are able? I've discovered that anaglyph 3D can currently be done by the end user with Reshade and Superdepth 3D shader but there is a problem with rotated screens as windows needs to be set to portrait for it to work and also the reshade shader takes quite a bit of a framerate hit compared to the inbuilt VPX 3D rendering modes. I suspect it would be more framerate friendly to build it directly into VPX?

There are not too many 3D TV screens around anymore so perhaps Anaglyph 3D is a viable alternative for all users (cheap passive glasses that work with any screen). Green/Magenta 3D and also Blue/Amber 3D lenses seem to give better colour reproduction from what I have read, compared to the old red/blue glasses.

Anyway, thought I'd throw the idea into the ring. A huge thankyou for all your efforts.

 

I would happily take a beta of red / cyan
Only problem with amber / blue is, it is proprietary, it is owned by Color3D
So you can only get the glasses through them, in paper type only
And i do not know what is entailed in making something that uses it, and not putting yourself at odds with their legal rights
Maybe nothing, but i do not know the answer there

 

Plus is it was simply left open, to define a color value
Like entering in and arbitrary RBG Hex value, or moving some color sliders
   well, you can not sue a private individual from typing in yellow right?

 

I think you missed the link in the other thread to the blue / amber  glasses, you need to check this link and then choose the amber / blue option as it lands you on red/green by default. https://www.aliexpre...2840546771.html

The ability to swap the lens to left right , right left either physically or in the 3D settings might get around any legal issues if there are any, not sure. Your suggestion of DIY colour choice by end user is a good one though.

[wiesshund]

I think you missed the link in the other thread to the blue / amber  glasses, you need to check this link and then choose the amber / blue option as it lands you on red/green by default. https://www.aliexpre...2840546771.html

The ability to swap the lens to left right , right left either physically or in the 3D settings might get around any legal issues if there are any, not sure. Your suggestion of DIY colour choice by end user is a good one though.

 

 

 

 

The color on the Alien Express pair looks off
Mind you, one can not tell jack diddly from an online image, but if they are and may very well be, because Colorcode 3D has an active patent on the whole affair, then being able to adjust the colors would be needed
BAM literally calls it colorcode 3D, but i am not the guy to advise someone else on legalities of doing something in their software, i lean to the side of caution, and not walking over someone's retail product.
But like i said, if a user can change the RGB values, no one can tell a private user "No, you may never ever ever pick that shade of yellow EVAH!"
A judge would say get the F___ out of my court.

 

Now Green / Magenta (aka Trioscopic) is also supposed to be very good on color depth, there were some more recent games, and movies done in it.
Coraline was done in it.

 

 

Quoting from Re: Anything but green/magenta: 3D and Stereo Photography Forum: Digital Photography Review (dpreview.com)

On generic chinese Colocode knock offs blue amber glasses

 

I recieved the amber-blue glasses that I had ordered on ebay.

They are not the official colorcode but chinese generic Amber/blue glasses.
I tried Bino, KMplayer and Stereoscopic player. First of all the image is really dark. Like two times darker than what you get with R/C glasses. Wth bino you can adjust the image brightness/ contrast very easily and it helps if you lower the saturation by 30% (retinal rivalry decreases). Still even in Dubois mode in Bino it was really uncomfortable. Too much ghosting of the yellow and bluy edges (in both eyes).

The other players had a bit better results concerning ghosting. In general... much greater color but the 3d effect was less visible. I really dont get why is that. Maybe because of ghosting.

Maybe the filters were not very good thats why they didnt filter out fully the colors. I will thik about ordering a pair of official colorcode glasses from 3dstereo.

I think you are right about Red/cyan being the best. Its soooo much easier on the eyes in dubois mode. I have some generic plastic RC glasses that have nvidia discover written on but I think they are knockoffs. I see some ghosting through those too.

Do you know about the pro-ana glasses? I ve read on forums that they are really the best anaglyph glasses and show almost no crosstalk. Should I try getting those?

 

 

 

I have all kinds of anaglyph glasses (more than 30 variants) and a 4K Passive 3D TV.

Regarding best anaglyph color combination:

1. Some sorts of Red-Cyan (and some Red-Blue) have the the best separation power, so will generate the less image ghosting, but at the cost of loss some colors (it depends on the type of Cyan/Blue filters).
2. Magenta-Green variants are the worst (high crosstalk and color shift).
3. In my opinion Amber-Blue are the best if the images are properly prepared (this is not an easy task). ColorCode 3D Amber-Blue glasses does not provide a perfect extinction on brown filter in some circumstances (by example with high contrast subjects in low light scenes), so I use a custom brown filter with enhanced extinction for my purpose.

These types of glasses are good in the first place for viewing 3D printed material. For viewing on PC/TV screens, polarized glasses (passive) or electronic glasses (active) are the best.

 

 

Chromadepth lenses are clear
But it takes a shit load of work on the backend to do a chromadepth image from what i read
Look it up, but i would probably file 99 that one? people are not going to redo table art to accommodate it, nor would i expect them to.

 

Red / Cyan anaglyph, with proper filters (lenses) seem to be thought of as the most versatile, and also the easiest to implement

 

By the way, anyone trying this just to get a look at what is being talked about
You have to teach your eyes to focus properly, if you have trouble at first, stop, take a break, then try again for a few minutes
keep repeating little by little, it can be hard at first to keep your eyes fixed on the illusion, you will find them trying to focus on the reality instead

At first, just try to get comfortable with looking at the table image with nothing going on, and keep your head still, after a while you should feel your eyes being more cooperative and focusing on the 3d illusion

 

Then practice moving your head around, like you are following the ball
The idea is to teach your eyes to move around the illusion with out losing focus of the illusion

 

Then when that feels pretty good, go and play a table.
Depending on the quality of the glasses used, you may have to bump up the tables environmental scale a bit if it seems too dark
Or you can look at other effects in reshade to try and do that

 

If you are using paper glasses, just realize they may not give the best effect, depending on how cheap and old they are

[skillman]

Hi folks,

  I've been developing a table in 10.7 and am having issues with frequent VPX crashes when trying to play. Sometimes it'll happen within seconds of starting to play the table, other times it might run for 10 minutes or more without issue. I've now tried on two machines with two different GPUs (R290 and a GTX1060). No other tables have this issue. My table is an original - no ROM - and makes extensive use of FlexDMD. I've tried build 310, 328, and 338, and both the standard and debug versions of 328

 

When it crashes, I get a crash.dmp and crash.txt file, but WinDBG won't read the crash.dmp file. I thought switching to the VPX debug version might help but no luck. Is there a way I can glean anything useful from it? The stack trace in crash.txt is almost always the same (and doesn't seem very useful - looks like it's entirely kernel functions)

 

TL;DR

 

Other similar tables work fine - Guardians Of The Galaxy has a ton of VBScript and uses PuP and works without issue. Ghostbusters Slimer makes use of FlexDMD (though not as much as mine) and also works without issue.

 

With the machine ruled out as a possible cause, I believe it's most likely one of two things - either an issue with FlexDMD, or a corrupt VPX file. The reason I am suspicious of my VPX file is that I started getting errors from FlexDMD awhile back when I tried to get it to load images from my VPX file. I hadn't really been doing anything with images up until that point, but I was pretty sure I had done some tests with transparent PNGs from the VPX file. Anyway, when trying to load *any* image from the VPX file, FlexDMD throws an "Index out of range' error trying to load the file. The same file is usable in VPX, as are all of the other images and sound assets. I tried deleting a bunch of assets to shrink the VPX size, and Saving As to create a new VPX file, but always got the same error. I eventually gave up and now get my images from a FlexDMD project dir (like UltraDMD did)

 

That said, if it were a corrupt VPX file, I don't know why the crashes would be so unpredictable. They definitely don't happen at the same time/place every time. They are usually preceded by the table hanging for anywhere from 2 to 20 seconds, with all audio stopped, then maybe a short burst of the audio again before VPX exits. Maybe the .dmp file is being saved during that time?  

 

Anyway, if anyone has any troubleshooting advice, I'm eager to hear it. 

[wiesshund]

Post crash.txt here

[skillman]

Here's a crash.txt from last night:

Crash report VPX rev338 (f3b572a)
============
Process: VPinballX.exe
Reason: 0xE0434352 - UNKNOWN at 0023:7759B502
Thread ID: 0x1004 [4100]

Call stack
==========
7759B502 KERNELBASE.dll RaiseException + 0x62 (0xFFFFFFFFE0434352 0x00000001 0x00000005 0x7187f3a4)
61932931 clr.dll GetMetaDataPublicInterfaceFromInternal + 0x9721 (0x00000000 0xFFFFFFFFC7FB5C0F 0x3C7C460C 0x7187f4fc)
619337D6 clr.dll GetMetaDataPublicInterfaceFromInternal + 0xA5C6 (0x3C82AEB8 0x3C82B674 0x7187F4CC 0x6145e28f)
60F90C39 mscorlib.ni.dll (0x7187F4EC 0x60830B53 0x2F812498 0x7187f4ec)
6145E28F mscorlib.ni.dll (0x3C7CD598 0x3C7C45D0 0x3C7C460C 0x00000000)
394AA1D4  (0x00000000 0x00000000 0x00000042 0x3c826910)
394A7901  (0x3C9BC5B8 0x7187F6E0 0x60818604 0x3c9bc550)
607F2E01 mscorlib.ni.dll (0x3C9BC550 0x00000000 0x00000000 0x00000000)
60818604 mscorlib.ni.dll (0x00000000 0x3C7CD198 0x00000000 0x7187f710)
60818537 mscorlib.ni.dll (0x00000000 0x3C7CD198 0x00000000 0x3c9bc584)
608184F4 mscorlib.ni.dll (0x3C7CD198 0x00000000 0x00000000 0x7187f740)
607F2D5B mscorlib.ni.dll (0x188B3838 0x7187F788 0x6184230A 0x7187f7d0)
6183F066 clr.dll (0x7187F7D0 0x7187F778 0x61932410 0x7187f900)
6184230A clr.dll LogHelp_TerminateOnAssert + 0x92A (0x7187F894 0x3C9BC598 0x00000004 0x7187f7b8)
618485CB clr.dll LogHelp_TerminateOnAssert + 0x6BEB (0x7187F940 0xFFFFFFFFC7FB51C3 0x7187FA44 0x61959590)
61959687 clr.dll PreBindAssemblyEx + 0x3547 (0x7187FAE0 0x188B3838 0x7187FA84 0x7187fa28)
619EE306 clr.dll CorExeMain + 0x6F36 (0xFFFFFFFFC7FB5287 0x00000001 0x7187FA84 0x00000000)
619EE391 clr.dll CorExeMain + 0x6FC1 (0xFFFFFFFFC7FB52D3 0x00000001 0x188B3838 0x00000000)
619EE282 clr.dll CorExeMain + 0x6EB2 (0x00000001 0x00000000 0x00000001 0x61959590)
619EE471 clr.dll CorExeMain + 0x70A1 (0x00000001 0x00000002 0xFFFFFFFFC7FB538B 0x61959450)
61959538 clr.dll PreBindAssemblyEx + 0x33F8 (0x31327690 0x1DCA7E0F 0x7187FBA4 0x00000000)
61904C27 clr.dll CoUninitializeEE + 0x135F7 (0x17D17CD8 0x7727FA10 0x7187FF2C 0x77857a9e)
7727FA29 KERNEL32.DLL BaseThreadInitThunk + 0x19 (0x17D17CD8 0x1DCA7A4B 0x00000000 0x00000000)
77857A9E ntdll.dll RtlGetAppContainerNamedObjectPath + 0x11E (0xFFFFFFFFFFFFFFFF 0x77878A6D 0x00000000 0x00000000)ntdll.
Environment
===========
Date/time: 17/11/2021, 22:28:41:805
Number of CPUs: 8
Processor type: 586
System: Windows 10 (10.0 19042)

Memory status
=============
Total Reserved: 268360K (262M) bytes
Total Commited: 1733880K (1693M) bytes
Total Free: 2192000K (2140M) bytes
Largest Free: 2046016K (1998M) bytes

Registers
=========
EAX=7187F2E0 EBX=00000005 ECX=00000005 EDX=00000000
ESI=7187F3A4 EDI=00000001 EBP=7187FF3C ESP=7187FF34 EIP=77857A6E
FLG=00000212 CS=0023 DS=002B SS=002B ES=002B FS=0053 GS=002B


Failed to save minidump.

[Thalamus]

I think you've summed up your issue quite well @skillman. Seems something outside of vpx itself is crashing and since you've tried two different machines; maybe the next step is to shared that file with a few more and see if it is a common issue even outside of your tests.

[wiesshund]

mscorlib.ni.dll is part of .NET framework

clr.dll is also .NET Framework

 

The other 3 are windows core DLL's

 

I would first maybe try reinstalling .NET 4.8
and maybe .NET 3.5 also

[skillman]

mscorlib.ni.dll is part of .NET framework

clr.dll is also .NET Framework

 

The other 3 are windows core DLL's

 

I would first maybe try reinstalling .NET 4.8
and maybe .NET 3.5 also

 

I was hopeful, but sadly that wasn't it. .Net 4.8 is baked into Windows 10, so there's no uninstalling and reinstalling it. And I already have the latest version of 3.5 also installed. It was released in 2007, so I don't think there's any recent updates.

 

From what I've read about Windows stack traces, the symbol names can be misleading, as the offset after it indicates how far away from the symbol the actual stack entry was. In my case, some of them are a _long_ way away - likely not in the same code at all. I assume what I'm missing is a proper symbol table for all of the code, but I really don't know - I've never done Windows development before. Maybe the missing symbol table is also why WinDBG won't load the .dmp file. 

[wiesshund]

Not sure but the crash log might be more clear to toxie