3889 replies to this topic

[dyopp21]

@ Vampirolatino2
 
My playfield monitor is connected via a hdmi cable but is connected to my graphicscard via a hdmi-to-diplayport dongle.
My backglass monitor is connected via a HDMI-to-DVI dongle.
At least I think how it's set up. I'm at work right now so I'm not able tot check my cab at the moment.
Could this be the culprit? How do you have your monitors connected?
I tested vpx on another desktop pc and on this one I see no tearing lines.
 
Correct me if I'm wrong but should vsync not fight the screen tearing? looks like although enabled in nvidia settings vsync is not working.
If I disable in nvidia and enable in vpx I get terrible stutter

 
That is correct. Vsync eliminates or greatly reduce the screen tearing. This was invented for PC gaming that runs more faster than the monitor refresh rate.
Don't quote me on this, but I think is the dongles you are using. Try to connect at least the PLayfield TV or Monitor from HDMI to HDMI. Using only a cable, no converters in place... if you can/have a port in the Video Card.

Before you start going down the connection path as the culprit, I've already tested several connection types in my systems and all have the same tearing.

I did DVI, DVI-D, HDMI (straight), DVI to HDMI dongle and HDMI to DVI dongle. All displayed tearing. I'm pretty confident it's NOT the connection type.

As I said earlier in this thread, I've been playing with this for months. Connection types, multiple GFX cards vs single card, tables that use pinmame vs ones that don't, and of course every different setting I can think of in VPX and the GFX control panel. The only thing I haven't investigated is Intel systems vs AMD (all my systems are AMD).


Sent from my iPhone using Tapatalk

and try to enable PC or Game mode on your TV, maybe that also comes into play.

Also tested it on PC monitor vs TV to rule that out. Also shows up on PC monitor.


Sent from my iPhone using Tapatalk

[vampirolatino2]

Mmm interesting.... well, now that you mention AMD vs Nvidia/Intel ... I think when I had the tearing/lag on flippers were on PCs in Intel/Nvidia too ... so, I think is not that either

 

But trust me that I know what you are saying, for example, if a flipper have a decal on top the decal don't stay in sync with the flipper movement. That is one of the problems right? Or just screen tearing like vsync on games?

[dyopp21]

Give me an example of a flipper with a decal. What table? I've only downloaded 10 or so VPX tables and I don't know if I've seen that issue, or maybe I just don't have a table with a flipper decal.


Sent from my iPhone using Tapatalk

[vampirolatino2]

One from top of my head is for example Austin Powers, VP9.x table. It have a decal on top.

[hauntfreaks]

is this the tearing people are talking about?

first is 1/8 slow motion of VPX

second is VP9 pm5 1/8 slow motion

 

 

[dyopp21]

is this the tearing people are talking about?
first is 1/8 slow motion of VPX
second is VP9 pm5 1/8 slow motion
 

 

That's it. Far more noticeable in VPX than PM5, don't you think? It can show up in different location "bands" depending on your frame rate/vsync settings. With my settings it doesn't affect the flippers, it's farther off to the right, near the right in lane. (On a different table than you showed in your video. I'm working on Blackout VPX beta and it's pretty bad right around the right kicker hole when the ball moves through that area).


Sent from my iPhone using Tapatalk

[Jafjas]

@Hauntfreaks

That's it indeed!
A really "nice" example of the tearing I've got on my tables.
Have you always seen this tearing or is it something you started to notice as of late?

Jasper

[toxie]

something different: all people with problems regarding the playfield reflection might check again with the next build, i did some changes there that helped at least on my cab.

[fuzzel]

rev2279 is up:

 

- playfield reflection handling improved and optimized, please test if this works as before.
- stereo3D should work again
- scripts updated by JimmyFingers (it should be save to use these scripts for VP9 and VPX)
- use a constant detail level for the hit shape of a ramp and fix some null-pointer exceptions
 

[vampirolatino2]

Nice example haunt!! Thanks! Now we need to find a solution.

 

Thanks fuzzel and toxie for the constant development time you spend in vp!

Edited by vampirolatino2, 27 October 2015 - 03:39 PM.

[toxie]

The only thing i can say about the vsync/tearing stuff is that VP10 implements two schemes:

a) if vsync = 1, then it will use the same mechanism as VP9.9.X (or PM5), so it will -always- wait for vsync

b) if vsync = 2, then it will use a new mechanism (only on Win7 and up though, no support for this on XP) that will only wait for vsync if your average framerate is above the refreshrate.

 

Both use official APIs, so no weird hacks or whatever. But it could be worth trying both, as they differ significantly in the implementation.

[kiwi]

Great job guys!
The reflection also works on Win XP.

 

Thanks

 

Max

[dyopp21]

The only thing i can say about the vsync/tearing stuff is that VP10 implements two schemes:

a) if vsync = 1, then it will use the same mechanism as VP9.9.X (or PM5), so it will -always- wait for vsync

b) if vsync = 2, then it will use a new mechanism (only on Win7 and up though, no support for this on XP) that will only wait for vsync if your average framerate is above the refreshrate.

 

Both use official APIs, so no weird hacks or whatever. But it could be worth trying both, as they differ significantly in the implementation.

Toxie I hear what you're saying, but something is different between VP9.9/PM5 and VP10.  I've never had this kind of tearing in the former systems.  Maybe just a touch in the flippers (never to the point of being unacceptable), but when the ball rolls through the tearing band in VP10 it really stands out.  I'm not saying it's the fault of the software or the programmers, or that it is everyone that is experiencing the same thing, but I'm hoping if enough people see it and we start looking for some sort of commonality with those of us that DO have the anomaly, we can fix it.  VPX is the way forward (and beautiful!), but I would hate to think we can't figure this out and make it go away.

 

I have tried both types of vsync many dozens of times.  At the moment neither makes any difference.

[Jafjas]

Hey there!

Dyopp, I think the same thing exactly.
I have tried all the possible settings in vpx but could not solve the tearing mystery.
I love vpx for the superb gameplay and excellent graphics but for me this glitch does kill the joy a bit.
I hope we can solve this mystery together (although I'm not a programmer by any means).
I really thought it was a display problem but because I don't have it on other games like the pinball arcade I now think it has to be something else.
Again, credits to the devs!

Ps. Should we move this problem to a new thread because it's not proven to be vp10 related? Just a thought.

Edited by Jafjas, 27 October 2015 - 07:10 PM.

[dyopp21]

I have a discovery to add to the tearing issue.  Since we were discussing and I'm at work and on my laptop, I downloaded the latest revision and decided to see if I would see it with this Laptop/GFX card combo in the full screen mode (going to an outside monitor I can rotate).  So I gave it a shot and guess what?  ZERO tearing.  In fact I was shocked at how smooth the ball movement was overall considering this thing doesn't have a heap of horsepower.

Here are some deets on what's in my laptop:

 laptop gfx card.PNG   18.68KB   17 downloads

 laptop gen.PNG   18.64KB   17 downloads

 

So, I guess I should concentrate on the settings in my Nvidia card?  My systems have the GTX660 and GTX770.  Is everyone else with the noticeable ball tearing also using Nvidia?
 

Edited by dyopp21, 27 October 2015 - 07:31 PM.

[Jafjas]

Hello,

Yes I have a Nvidia gtx 760 if I remember correctly.
But my pc at work has also an nvidia card but a quadro and no stutter at all.
The only thing I haven't tested is installing the latest drivers.
I updated them maby two months ago so they are not the latest anymore.
Are you guys up to date on the drivers?

Jasper

[Ben Logan]

Just want to say that I'm all for testing VP while at work. In my own personal case, I can't think of a better use of company time. Godspeed gentlemen!

[mpad]

Yeah, less time for Facebook and Amazon

[chepas]

 

- playfield reflection handling improved and optimized, please test if this works as before.

 

It's done something to the alpha layer on a playfield image if using these for holes..

[jimmyfingers]

rev2279 is up:

 

- playfield reflection handling improved and optimized, please test if this works as before.
- stereo3D should work again
- scripts updated by JimmyFingers (it should be save to use these scripts for VP9 and VPX)
- use a constant detail level for the hit shape of a ramp and fix some null-pointer exceptions
 

To add to what scripts / functionality I changed it was mainly the core.vbs and how the vpmNudge.TiltObj array / disabling of those items takes place, but as a result the optional NudgePlugins also had to be updated as they contained the same section of code that would be invoked (by-passing the changes in the core.vbs) if the plug-ins are used / called from a table's script. 

 

From the core.vbs internal documentation section:

- Changed vpmNudge.TiltObj handling to use Bumper.Threshold / Wall.SlingshotThreshold temporary value changes rather than force / SlingshotStrength changes to disable tiltobj array objects

- There existed a bug in VP since at least the 9.x versions where the Wall.SlingshotStrength value being set by scripting during game play did change the value but the slingshot behaviour / "Slingshot Force" (from the editor) of the wall object did not change (i.e. did not have an effect); As a result the attempted disabling of bumpers and slingshots after a tilt event on supported games (that can send a relay for vpmNudge.SolGameOn ) would only work for the bumper objects

- Using thresholds instead also now has added benefit by not actually triggering the related _Hit or _Slingshot routines so animations, sound processing, and other potential nested subroutine calls will also not activate resulting in a better tilt simulation

 

Note:  I made a typo in the internal documenation section and the "Wall.SlingshotStrength" shown above is actually incorrectly referred to in the core.vbs comments section as Wall.SlingshotThreshold (will get that fixed)

 

So you should see a better end result with now finally slingshots being "disabled" as well as both object types (slings and bumpers) no longer still playing sounds or animations on a tilt for at least games that have a power on / tilt relay and are properly coded to forward that callback to vpmNudge.SolGameOn (couple example games / generations are earlier Williams SS - Comet is one at least - and Bally earlier SS - Eight Ball and Playboy being another example).