868 replies to this topic

[atarian]

I was using your settings too Jimmy and I would say it was almost perfect, however I would still get the odd 'jitter' - might be as infrequent as every 2-5 minutes and could be a minor ball glitch. Nothing major just the occasional glitch enough you spoil the illusion that I am playing a real pin. Going to real FS I don't see any glitches at all. I don't think it is my imagination because these tiny glitches always bother me and now looking for them and I don' get them.

For the lag, it feels a bit more responsive, and it feels like it does with Aero off in Windowed FS.

All my experiences are with two screens and a b2s. To run as smooth as possible for any config I always have b2s running as an exe.

Edited by atarian, 04 May 2016 - 08:23 PM.

[jimmyfingers]

Don't get me wrong, I'm not trying to purport my views or settings, I'm just looking for objective data / information on the performance if it exists.  Mostly, when people are / were discussing "performance" I was wondering whether this is in measurable FPS or is it all in visual fluidity / input lag.  Can you guys comment on FPS increase / changes that give us actual numbers?  I know it's a different angle and not related directly to smoothness / input lag, but I just want to confirm what it is exactly people are referring to regarding massive performance changes and also what FPS changes / if any, so that I can compare.

Edited by jimmyfingers, 04 May 2016 - 08:49 PM.

[atarian]

Hi Jimmy - I am sure I speak for most in saying I always appreciate all your analysis - it's your VP10 settings I have been using successfully for ages now, so please ask for any info that helps.

I had my FPS locked to Vsync in the VP video settings (ie: 1) in Windowed FS so I kept that setting for True Full screen too so FPS a consistent 60fps.

Out of interest I just tried a test with the Frame Limiter set to 0 using RoboCop and a B2S in Exe mode.

Windowed FS : 112.0 avg FPS
True Fullscreen: 132 avg fps

So almost 20% improvement. Tomorrow night will try an hour or two actually playing some tables with this FPS setting to see how it compares.

I know some time back you said that you had a noticeable lag with Aero/DC but used it to avoid tearing in Windowed FS. As real FS should eliminate tearing now, and you can now disable Aero/DC are you noticing less lag yourself with these settings? I know it is hard to measure - there needs to almost be a way to measure the delay between the physical noise of the flipper button press and the sound VP plays for a flipper up movement.

Cheers

Edited by atarian, 04 May 2016 - 10:17 PM.

[jimmyfingers]

@atarian - thanks for the kind comment - I wish my boss would appreciate my analysis as much as it seems you do and the ironic thing about that situation is that the word analyst is in my actual job title  Ahh, crazy, scattered, non-communicating employers, you gotta love them.  You're compliment above comes at a key time as I'm getting burnt out in more ways than one.

 

As far as FPS, to clarify there, I definitely play and keep it locked as well for vsync but for testing I disable it to assess the maximum FPS and impact on table changes as well as even VP software revisions.  So, don't take that to mean I recommend at all playing with FPS wide open, it's simply a temporary setting for measurement and....analysis

 

I did mention that for sure there is some level of difference with input lag with Aero / DC / DWM vs. it off but needed / wanted it, yes, set on to get stutter free play and no tearing.  The thing is that it wasn't visually that noticeable really and that's where I'm surprised some people new to VP can see it but also why I still suspect that standard TVs input lag is easliy half of that situation / problem.  As most people know by now, I scrutinize VP in a detailed way and even with that it still took me only until recently to definitively prove that I had a difference in lag between these two settings.  And that was only done mostly in how shots worked / were being made a little better for backhands / back angles and balls rolling through the in-lane since the physmod and VPX days where the flippers physics changes had lost a little bit of that ability for those shots despite all the other gains they made.  

 

As much as i could test for it and did notice it, with all the other settings I had and monitor type, it was still workable for me if I tuned up everything else (USB polling rate, leaf switches, and especially in that my mini-cab HV setup has 120hz, which not only is great for smoother game play / less blur but effectively halves input lag from delayed / buffered frames vs. a 60hz monitor - while vsynced of course).  I have another method coming / to be release that has had some discussion here and there on the forums that will save between 10-20 ms of input lag from how currently all VPM table flippers are triggered.  It is done by bypassing waiting for VPinMAME to return the solenoid firing status and instead rotates / fires the flippers directly from the keydown / button presses.  The previous problems with that is that flippers always stay active but with a certain genre of tables you can leverage other solenoid calls to cease the flippers when appropriate yet gain this more direct fired approach and I have successfully tested a workaround subroutine that will allow a method to work on tables that don't even have the typical "game active" / "game on" status that will allow for this as well plus give the added benefit of being able to more adequately mimic a tilt event using the same synthetic "game on" status derivation / calculation.

 

I will demonstrate the simpler / more basic of the method in both of the first 2 tables for VPX when I finally get them launched, which will hopefully be soon (Firepower and Comet).  The latter method will be for more modern tables like LW3 and AFM and requires a bit more testing but does look promising and will help yield another way we can code to minimize input lag at least within the software layer for VP / VPM.

Edited by jimmyfingers, 04 May 2016 - 10:51 PM.

[unclewilly]

Sorry haven't been around much for this. The fs mode has amazing performance boost.
My only issue is, with windows 10. When i try to exit a table i get the message hold esc or press enter.
But nothing happens when i do either. I have to ctrl alt del. To get the task manager up which closes the player.
Did i miss something i should be doing

[jimmyfingers]

Sorry haven't been around much for this. The fs mode has amazing performance boost.
My only issue is, with windows 10. When i try to exit a table i get the message hold esc or press enter.
But nothing happens when i do either. I have to ctrl alt del. To get the task manager up which closes the player.
Did i miss something i should be doing

At that prompt, hit the down arrow key twice, then the enter key.  It gets me out of that screen (i.e. just can't see what's going on).

[Slydog43]

I have the exit trouble also.  Holding down the esc doesn't seem to do anything, but enter does return to the game right away

[hauntfreaks]

just hit "Q" twice

[atarian]

For me as well the lag is more of a feeling rather than something I can notice visually. The way you can time a shot, etc. Having said that I can adapt between real and virtual pinball quite quickly.

It will be interesting to see jimmys experiments to reduce lag even more.

Hitting Q twice also works for me to exit a table.

[clydedrexler]

I´n agree with Jimmyfingers,

I test 2 tables MedievalMadness and RollerCoaster

And no many differences in FPS...

 

In MM:

With Windowed fullscreen : 100.7 FPS

With NEW full screen: 102.5 FPS

 

In Rollercoster:

With Windowed fullscreen : 118.6 FPS  <-- Higher than the new full screen

With NEW full screen: 117.8 FPS

 

I´m also have the problem with rom find...ddraw, directX ...registry, etc in MM. but for test purposes I change the version

 

Thanks a lot for your hard workl!

 

ClydeDrexler

 

 

 

Attached Files

•  IMG_20160505_134108.jpg   76.35KB   23 downloads
•  IMG_20160505_142055.jpg   75.19KB   23 downloads
•  IMG_20160505_145258.jpg   81.75KB   21 downloads
•  IMG_20160505_144821.jpg   81.19KB   22 downloads

[gigalula]

UW that's exactly for all those little bugs or call it like you want from win10 that's why I finally replaced my win7 backup after more then a month trying to get all to work perfectly there is always something wierd... WIN10 give me strange bug with my AHK script ... computer and cab restarted from sleep itself every day never been able to find the way to prevent this...so I gave up for the now and stick to WIN7.

[hauntfreaks]

the performance gain is for sure there during gameplay.... in every aspec... the guys that are seeing it like myself are not going by #s .... you can see it with your eye and with you fingers... the whole gameplay is brighter, livelier, snappier, smoother and 20 more "er" I can't think of right now.... oh its there.... and we feel it!!  

[clydedrexler]

the performance gain is for sure there during gameplay.... in every aspec... the guys that are seeing it like myself are not going by #s .... you can see it with your eye and with you fingers... the whole gameplay is brighter, livelier, snappier, smoother and 20 more "er" I can't think of right now.... oh its there.... and we feel it!!  

I do more test and I "feel" the gain.
It's no by number of FPS... Is just a sensation and feel.
Salute
ClydeDrexler

Enviado desde mi HUAWEI G7-L03 mediante Tapatalk

[kiwi]

I have a curiosity about the wire ramps,

I have noticed that if I put to 0 the physical walls, the ball not falls from the rails,

and sometimes the ball gets stuck at the exit ramp.

 

Thanks

 

Max

[doogie2301]

Is there any trick to get custom resolutions added via the NVIDIA control panel to show up in the list of available resolutions for true fullscreen?  I am running 1920x1080 @ 120hz, but it looks like only the standard resolutions (up to 60hz) are in the Visual Pinball dialog.

[kiwi]

If you can have "Desktop anti-stretch the ball", since there are tables that don't have the scale X and Y equal, and even use the layback.

 

Max

[jimmyfingers]

Is there any trick to get custom resolutions added via the NVIDIA control panel to show up in the list of available resolutions for true fullscreen?  I am running 1920x1080 @ 120hz, but it looks like only the standard resolutions (up to 60hz) are in the Visual Pinball dialog.

You can manually control that in the registry under the appropriate VP10 branch (should be HKEY_CURRENT_USER\Software\Visual Pinball\VP10\Player "RefreshRate").  Set all the options and screen size that you would want in VPX's video preferences first, then close it out / choose OK (no need to relaunch VP), edit registry and manually enter 120hz.  Try it with the default / non-VPM table first and then look back in the thread about how to get it to work for VPM / VPinMAME tables / ROMs. Keep in mind any change in the video preferences again will overwrite that and you will need to manually re-apply through the registry.

 

This same refresh rate issue (and 120hz specifically) was discussed a few pages back:

http://www.vpforums....=22#entry341421

 

It's useful to look through this and any other topics fully first as quite often the answers to questions and desired information already exist.  Threads that are long and / or detailed like this may take some end-user time to catch-up and sort through, but it's certainly reasonable in comparison to how much more time the developers actually spend simply producing this software and the options that invoke these conversations as well as even the time and effort for other forum members to post their help and findings.

[doogie2301]

 

Is there any trick to get custom resolutions added via the NVIDIA control panel to show up in the list of available resolutions for true fullscreen?  I am running 1920x1080 @ 120hz, but it looks like only the standard resolutions (up to 60hz) are in the Visual Pinball dialog.

You can manually control that in the registry under the appropriate VP10 branch (should be HKEY_CURRENT_USER\Software\Visual Pinball\VP10\Player "RefreshRate").  Set all the options and screen size that you would want in VPX's video preferences first, then close it out / choose OK (no need to relaunch VP), edit registry and manually enter 120hz.  Try it with the default / non-VPM table first and then look back in the thread about how to get it to work for VPM / VPinMAME tables / ROMs. Keep in mind any change in the video preferences again will overwrite that and you will need to manually re-apply through the registry.

 

This same refresh rate issue (and 120hz specifically) was discussed a few pages back:

http://www.vpforums....=22#entry341421

 

It's useful to look through this and any other topics fully first as quite often the answers to questions and desired information already exist.  Threads that are long and / or detailed like this may take some end-user time to catch-up and sort through, but it's certainly reasonable in comparison to how much more time the developers actually spend simply producing this software and the options that invoke these conversations as well as even the time and effort for other forum members to post their help and findings.

 

 

Thank you for the help, jimmyfingers.  I did search this thread for "120hz" before posting my question, and saw your earlier post.  However, there was no specific solution listed, other than some reference to a "registry hack", which I could not find any details on until you just posted now.  So again I appreciate your response, and I'm sorry to have wasted your time.

[fuzzel]

I have a curiosity about the wire ramps,

I have noticed that if I put to 0 the physical walls, the ball not falls from the rails,

and sometimes the ball gets stuck at the exit ramp.

 

Thanks

 

Max

If use use a wire ramp VP adds hard coded wall heights otherwise the ball would fall off a curved wire ramp.

[jimmyfingers]

Thank you for the help, jimmyfingers.  I did search this thread for "120hz" before posting my question, and saw your earlier post.  However, there was no specific solution listed, other than some reference to a "registry hack", which I could not find any details on until you just posted now.  So again I appreciate your response, and I'm sorry to have wasted your time.

 

Even though I was responding to you specifically, I was also mentioning for a wider audience / more general aspect of people not looking through longer threads where issues and certainly hints, ideas, and pointers already exist.  Fair enough that you had apparently looked back / search but it didn't sound from your question that you were aware of the previous post about the 120hz or any building on that with any specific questions about details you may have wanted / needed regarding the "registry hack".  Again, even just expanding on the fact that you had read that part and needed more detail / information would be welcomed for others to assist or economize on their response and attempt to help.  

 

You didn't waste my time and no worries / need to be sorry.  As I say, there was a general message here that wasn't just specifically regarding your question / post but a more general trend that shows itself at times.