357 replies to this topic

[Ben Logan]

Your invisible flippers idea was brilliant. As is Toxie's implementation. Witnessing collaboration like you guys' is one of most inspiring aspects of VP, imo. 

[nFozzy]

You could have the rotatetostart always be instant for both flippers, but then there'd be some visual delay with tap-passes. Up to you I guess.

[toxie]

I hope JF will chime in here, too, at some point.

He has some other brilliant ideas in the making (plus he invented some of the stuff mentioned in here, and then we apparently re-invented it mostly ).

[Ben Logan]

You could have the rotatetostart always be instant for both flippers, but then there'd be some visual delay with tap-passes. Up to you I guess.

 

Interesting idea. Would visable flippers still freeze at start of game, during ball drain, and at end game if both sets were coded for instant rotatetostart? 

[nFozzy]

I haven't tested it, but I assume it would work, I can't think of any tables that force the flippers to stay upright like that.

[Ben Logan]

Nice! I'll try it after work.  

[kiwi]

In Surf'n Safari it's possible baypass the flipper solenoids using only the scripts, a bit of code to add and edit,
this table has a solenoid that says when the flippers are on or off.

[Drybonz]

Thanks to Ben Logan for sending me a copy of a table modded with the invisible flippers.

 

When I leave the new flippers visible I perceive zero difference in between them and the normal flippers... they basically move together.  When I play the table it seems the same to me.  Hitting various ramps felt about the same to me as the timing always does.

 

I don't think this tweak is useful for people who don't have older hardware, but I think it may be useful for people who do.  

 

I believe good gaming hardware and monitor are the best answer to input lag, and this tweak is, at the very least, a nice benchmark for people to tell if they are getting bad flipper lag by comparing the two flippers.

 

I think that people that are noticing a serious input lag, including probably Noah's potential clients that he mentioned, are seeing shortcomings in the hardware they are playing on.

 

All that said, I think the most useful testing feedback will come from people that are getting more input lag than I am, apparently.  I'm still interested to see what you guys come up with.

 

One other thing I will mention is that I wouldn't look forward to plugging that script into 200 tables (or however many I have), and I enjoy tweaking and testing tables... I'm afraid that the general population of players may not adapt to the script.

[gtxjoe]

Drybonz, everyone's setup including yours has a 11-30 msec delay in flipper response due to the way vPinmame works. This flipper script mod eliminates this delay

Do you play real pinball machines? If you don't then that might explain why you do not notice a real difference in n this script change. Whenever I play real pinball I have to remind myself to not flip so soon because I have gotten so used to the slow flipper response in VP

[Drybonz]

Do you play real pinball machines?

 

Yes.

 

I don't notice any difference in the table I tried with the flipper changes.  Others might.  I think it's great if it benefits other people.

[Noah Fentz]

...
 
I think that people that are noticing a serious input lag, including probably Noah's potential clients that he mentioned, are seeing shortcomings in the hardware they are playing on.
...

To be clear, on our hardware, there is no noticeable input lag. This is something we often get complimented on when compared to others.

Instead, flipper rotation speed has been the culprit.

Looking forward to experimenting this weekend on improving that for VPX. Busy week, this week, so that'll be my first opportunity. Will post my findings then.

Sent from my SM-G930T using Tapatalk

[Drybonz]

I feel like there is maybe a couple different ongoing topics... or it is very possible that I am confused.  That happens a lot.

[BorgDog]

Like kiwi's talked of doing on his surf'n safari, I redid my pinball pool and incredible hulk tables tonight to just basically do flippers like an em with a check to see if either the game over or tilt lights were on (if either of those on then no flip)  No reason on the Gottlieb system 1 tables to have the flippers run through pinmame.  I would think other tables of that era could do something similar, and newer ones depends on the table I guess, and some you'll likely just have to do double-flippers and live with slight consequences.

 

Since I have an actual Pinball Pool sitting next to my vp cab I was able to play both side by side, and before making the change where I could tell the difference easily.  After changing the script it is not exactly the same but it is a lot closer and definitely worth doing.

 

I also added the flipper returntostart to both the visible and not visible flippers on the AFM table I've been messing with and hopefully that eliminates Ben's ball in flipper issue, I did not notice it in the couple games I played. nor do I notice any oddities in slower speed movements like tap passes and such.  I also noticed the AFM I have on my cab had coil ramp up set at 5 and lowered it to 2 after doing the other changes and gained a little there as well.

 

Every little bit helps.

[dark]

I hope JF will chime in here, too, at some point.

He has some other brilliant ideas in the making (plus he invented some of the stuff mentioned in here, and then we apparently re-invented it mostly ).

That's good to hear I was wondering what that mastermind has been up to!

[toxie]

The problem with lag is that there are so many components involved, so each part plays a different role, depending on one's setup. So here a quick sum up again:

 

a) The flipper button/keyboard itself (minor)

b) USB polling rate (major: http://www.vpforums....howtopic=29207)

c) VP framerate to trigger the script handling (minor, but the higher the better, e.g. vsync turned off or using NVIDIAs fast sync can help a bit)

d) VP<->VPM script roundtrip (major, also depending on e and f indirectly)

e) Graphics driver (major, maximum prerendered frames tweaking can help)

f) Monitor/TV (major, use game or PC mode, and best is a 144Hz or better one, as it also influences c,d,e)

 

Plus tweaking the flipper setting in VP itself (coil ramp up 0 is fastest, then one can also tweak mass and strength, in combination with EOS Torque to avoid weird physics effects).

[vogliadicane]

Hey guys, I am following this thred with high interest and like, how you work together to trace down the culprit...

 

First of all I also experience the flipper delay in VP10.1, but not in VP9.9 and PhysMod. My setup is rather new e.g. a G-Force 970...

 

No to the important point. Yesterday I made some observation for VP10.1.

Contactors controlled by DOF in my system are absolutely in Sync with button presses (compared to the flipper movement, which is delayed a lot). I would expect, that the contactors as a mecanical component with some real inertia even show a bigger delay, but astonishingly it is actually insignificant, at least very low (I am a musician and really trained to timing).

 

What I wonder from this observation is: Isn't DOF also triggered by VPINMAME so doesn't the signal also have to travel from the button to VPM and back before sent to some of the contactors, which should result in some delay? In other words: shouldn't the contactors be also delayed like the flippers are, IF the way to VPM and back really is the culprit of the delay?

 

Just wanted to thorow this in, maybe it is helpful info.

Edited by vogliadicane, 19 October 2016 - 09:38 AM.

[toxie]

Absolutely. But as mentioned in my post above, there are soooo many components at work here, and depending on each ones setup, they can be worse or better (e.g. how much each component adds in term of milliseconds of lag/latency can differ on each system). Also, a standard 60Hz TV will most likely always be the major bottleneck in combination with the gfx driver (e.g. how many frames does the driver and the TV buffer internally until one sees it).

Plus, on top of all that, our brain in combination with visual and audible systems also plays a role (and that one is also highly non-linear, as it will compensate for some effects).

 

As for the differences when using VP9, PM5 and VPX: Could you try for example Dr.Dude for VPX, as it features relatively fast flippers?

Also, do you limit the framerate differently for them (e.g. do you use vsync and if not what is the average framerate for the different engines when you test)?

[Slydog43]

I have an off the wall idea.  Whenever a table gets started, a quick flipper test is performed on the table. ie you have to flip at 5 balls on each flipper on some sort of fun shooting gallery of sorts.  I know this is really a joke but stay with me.  As I stated earlier my brain quickly looses the perception of flipper lag.  Then when the game starts your brain will be in tune with VP and not real pinball.  I move back and forth when I play and barely notice flipper lag and almost never notice after just a few shots.   

[vogliadicane]

 

As for the differences when using VP9, PM5 and VPX: Could you try for example Dr.Dude for VPX, as it features relatively fast flippers?

Also, do you limit the framerate differently for them (e.g. do you use vsync and if not what is the average framerate for the different engines when you test)?

 

Actually I had tested Dr. Dude after you mentioned it here, and I do feel a lag for flipper movement similar to other tables.

And as I mentioned, these lags are not noticable at all for VP9.9.2 and PM5, so I don't think the TV's lag contributes a lot.

There is surely also a small lag in VP9 and PM5, but it is absolutely tolerable. Such as a real table surely also does NOT have a zero lag; coils must ramp up before movement e.g.

 

The other stuff I will look at, when I'm back home tonight.

 

All I can say now is, that I'm using same settings for VP9, PM5 and VPX, so I suspect, the lag comes more from something within VPX and not the hardware or setup. But as I said, I'll check for and give you the info.

Edited by vogliadicane, 19 October 2016 - 12:13 PM.

[toxie]

Weird. So do you use vsync or some other frame limiter on any of them? How do the FPS differ between the three?

 

Cause the physics themselves are pretty much the same between PM5 and VPX (except for a whole lot of fixes, but none of these should influence the perceived speed of the flippers).