51 replies to this topic

[Thalamus]

It is normally a recommended setting to put it to performance. Honestly I haven't bothered to do that myself though, I prefer the fans to not start if the table isn't demanding it. But, it is definitely a good idea to try out. I know that weisshund has mentioned to use vsync, you can always try my current setting being 2 for the fast sync and 2 for the pre-render. But, don't try to run that with pup-packs.

[kiwi]

Nvidia control panel: I put the power management mode as adaptive, or maximum performance,

I don't know if the terms I just used are the same in the Nvidia panel in English, with this setting the game is much smoother.

I do not know if it is due to my system,

but I had to use this option because after about fifteen seconds that I have launched the table,

the consumption of GPU doubles for no reason and can create stuttering,

perhaps with more powerful GPUs this is not noticed .

[DKong]

Here is what I have discovered so far:

 

FPS Limiter/Vsync: 1

Max Frames: 1

With these VPX video settings, I get FPS: 60.0 (59.4 avg.) 

 

FPS Limiter/Vsync: 1

Max Frames: 2

With these VPX video settings, I get FPS: 60.0 (59.5 avg.) 

 

FPS Limiter/Vsync: 2

Max Frames: 1

With these VPX video settings, I get FPS: 60.0 (61.3 avg.) 

 

FPS Limiter/Vsync: 2

Max Frames: 2

With these VPX video settings, I get FPS: 60.0 (60.6 avg.) - some other tables like Flupper's Totan with this setting get a 60.0 (62.4 avg.)

 

These are with the suggested Nvidia settings above. The only thing I have not tested yet is the experimental registry change and I plan to try that as well.

 

Should I only be focusing on the first number for FPS when running a table? So, as long as it says FPS: 60.0 any of these settings should look/play similar?

Or should I go with the one showing an average FPS closest to the FPS 60.0, so in my tests, FPS Limiter/Vsync: 2 and Max Frames: 2 would be the best for my setup?

 

If the average FPS number is important, is it better to have it just over 60 FPS or just under that number?

[wiesshund]

There is no magic number

you are looking for what eliminates visual ball stutter
that is going to vary, as i showed you in my video?

if none of the things you are trying so far as any effect
Then it may not just be visual stutter you are trying to hunt down, but real physical ball stutter

 

And then you will have to try some of the other things listed about, to see if you find a smoking gun

[DKong]

Okay thanks. The tough thing is waiting and watching for the ball stutter. I don't notice it every time I play certain tables, but it does happen often enough which is why I'm looking around to see if I can get that "butter smooth" that I hear about so often in the forums.

 

When others say they're rig is running at 60 FPS locked, are they referring to the first number that shows up in F11 (the 60.0 I'm getting) or the average number to the right of this number?

[wiesshund]

Okay thanks. The tough thing is waiting and watching for the ball stutter. I don't notice it every time I play certain tables, but it does happen often enough which is why I'm looking around to see if I can get that "butter smooth" that I hear about so often in the forums.

 

When others say they're rig is running at 60 FPS locked, are they referring to the first number that shows up in F11 (the 60.0 I'm getting) or the average number to the right of this number?

 

They are most likely referring to the fact they are running V-Sync with a 60Hz monitor

I dont think anyone is immune from having a frame get interrupted from time to time
hell there are some B2S backglasses out there that will pretty much guarantee it.

And the average number factors in things that may not even be going on during game play
Do anything to kick out one frame, and your average is less than 60 and that is where it will stay, because even 1 frame out of 10,000 still does not = 60 average

 

Did you watch the video i posted?
Look at the frame rates, and watch the ball.

 

The frame rate is not the be all end all of it, you can see it in that video
The FPS is crap, but there is no stutter

The biggest reason for locking the frame rate to vsync is to keep the video card drawing frames in step with how the screen is able to draw them
so you do not have shearing and tearing going on making it look like bad things are happening to the ball, even though the game is running fast

Because the monitor is only drawing some of the frames while the video card is off running at it's own pace.

 

If the monitor can only draw 60 frames every second, and the GPU is throwing between 100 and 130 frames at it, with the GPU fluctuating depending on what is going on
at a given moment, the monitor is missing out on 40 to 70 frames of animation, and since the GPU is fluctuating on it's output, the frames the monitor is shredding up
dont even happen at the same place, it looks bad if you have a very distinct object that is moving quickly and changing directions instantaneously.

 

You are going to have to go through All the stuff that everyone has suggested
Until you identify which is actually causing it
 

[DKong]

I watched the video you posted, thanks for that @wiesshund. My tests showed similar numbers as yours (with vsync off), but I was getting ball and other artifact stuttering with vsync off.

 

I finally had time to test the experimental registry hack @Thalamus mentioned at VPinball. My results were quite different from @DJRobX in that when I give the MinPhysloopTime key a high value (like the 600 he and others have success with), my FPS dropped to 4.5. As you can imagine, the tables were unplayable as everything slowed to a crawl.

When I put a value of between 5 and 29, I get much more favorable FPS during gameplay.

 

Here are my results:

 

Test table Attack from mars (G5K version):

FPS Limiter/Vsync: 1

Max Frames: 1

MinPhysloopTime: 600

Gameplay FPS: 4.5 (4.3 avg.)

 

FPS Limiter/Vsync: 1

Max Frames: 1

MinPhysloopTime: 5

Gameplay FPS: 60.0 (59.2 avg.)

 

FPS Limiter/Vsync: 2

Max Frames: 2

MinPhysloopTime: 9

Gameplay FPS: 60.0 (60.2 avg.)

 

FPS Limiter/Vsync: 2

Max Frames: 2

MinPhysloopTime: 5

Gameplay FPS: 60.0 (62.2 avg.)

 

FPS Limiter/Vsync: 2

Max Frames: 2

MinPhysloopTime: 19

Gameplay FPS: 60.0 (62.2 avg.)

 

FPS Limiter/Vsync: 2

Max Frames: 2

MinPhysloopTime: 10

 

Gameplay FPS: 60.0 (59.9 avg.)

 

Test table Totan (Flupper version):

FPS Limiter/Vsync: 2

Max Frames: 2

MinPhysloopTime: 10

Gameplay FPS: 60.0 (62.1 avg.)

 

Test table Cirqus Voltaire (Knorr version):

FPS Limiter/Vsync: 2

Max Frames: 2

MinPhysloopTime: 9

Gameplay FPS: 60.0 (60.9 avg.)

 

Test table AC/DC (PwrUp version):

FPS Limiter/Vsync: 2

Max Frames: 2

MinPhysloopTime: 9

Gameplay FPS: 60.0 (61.5 avg.)

 

Test table Firepower (3rdaxis version):

FPS Limiter/Vsync: 2

Max Frames: 2

MinPhysloopTime: 9

Gameplay FPS: 60.0 (61.5 avg.)

 

I know you had answered this question above about the FPS being not very important, but since I've done these tests, I would really like to know if gameplay is typically smoother when the Average FPS is slightly above or below 60 (when the FPS is 60.0).

 

Lastly, I tested a PUP Pack table Stranger Things. For this table, I could not leave the FPS Limiter/Vsync: 2 and Max Frames: 2 since the FPS was all over the place. So for this table, I changed Vsync to 1 (in the table options) and left Max Frames at 2 in the VPX global video options.

 

Here are my results with this PUP table:

FPS Limiter/Vsync: 2

Max Frames: 2

MinPhysloopTime: 9

Gameplay FPS: 60.0 (59.9 avg.)

 

@Thalamus you mentioned not using vsync/Max frames each at 2 for PUP pack tables. These both set at 2 globally seems to give the smoothest gameplay and best FPS for most tables. However is this setup the best for FPS for these PUP tables?

 

Thank you guys for your continued help with dialing in VPX for the smoothest gameplay possible. It's greatly appreciated.

[wiesshund]

It will depend on what your monitor likes best
which ever allows it to draw the smoothest

 

so you will have to run some comparisons between vsync 1 and 2
while watching the ball really careful

 

shove a wall with the ball facing side set as slingshot with a power of like 35 and a threshold of 0.5 in front of the drain

that will let you pay attention to the ball, since it will keep throwing it back on the table

 

see which gives the smoothest perception over all.

If you still encounter random stutters
(monitor sync wont be random)

Then move on to checking the other things.

I might start with disabling b2s, and seeing if the stutter goes away
if it does then you know you need to tweak the backglasses for the problem tables

Some of the fantasy backglasses tie into all kinds of lamps and solenoids on the tables to do nifty things in the backglass
but that can also create a lot of load that can slow the table down cause a lot of communication is going on

But you can do things about that.
You can tweak the backglass, try a different backglass, or even edit a backglass and remove all the fantasy lights
and put in simple static lighting, or a self contained animation sequence that uses no B2S communication

[DKong]

I'm not familiar with adding/moving walls or objects in vpx, so that is beyond my ability for now. Part of my tests included running with and without b2s and I did not see a tremendous difference.

I'm going to leave my settings as above vsync/max frames 2/2 for all but the PUP tables and watch them for awhile. 

 

I'll report back if I notice anything of value that might help others. 

Thanks again.

[wiesshund]

I'm not familiar with adding/moving walls or objects in vpx, so that is beyond my ability for now. Part of my tests included running with and without b2s and I did not see a tremendous difference.

I'm going to leave my settings as above vsync/max frames 2/2 for all but the PUP tables and watch them for awhile. 

 

I'll report back if I notice anything of value that might help others. 

Thanks again.

 

Hmm, did not suggest moving walls etc
Do not think the vsync settings have any effect on PUP?

 

Only thing i can think of offhand where PUP comes into the mix, is running 4K PUP videos, as PUP only does software rendering, no hardware
so it is all on the CPU, and processing a 4K video on the CPU is a bit processor intensive, might compete with the table

[htamas]

I'm wondering why does the setting 'Reflect elements on playfield' has such a significant effect on ball stutter with some tables.

There are a few where no matter what I tweak or do, with this activated I get stutter, but as soon as I turn that off, the table plays perfectly smooth.

Some examples coming to mind are Goldchicco's Arena and Al's Garage Band, or wrd1972's latest BBBB version, but there are some more.

This would be OK, except that on a few of those, turning this off also causes certain rendering problems and things disappear. On others, the difference is not really noticeable.

 

I remember some folks mentioning a similar problem. What makes this setting so heavy on performance, and what may be the reason for it?

I'm leaning towards processor power, since my video card is a GTX 1060 (which should be more than enough for a 1080p playfield)  but CPU is only an AMD FX 8320E, so perhaps the CPU bottlenecks the GPU. Still, there are so many tables that seem to be pretty intense, yet they are running perfectly fine, only a small number have this problem with the reflections. Strange...

[wiesshund]

What is BBBB?

 

i was thinking big bang bar, but not enough B's

 

I am also thinking that is not having reflections on a specific table causes parts to disappear, there may be a design issue in the table

that has gone un noticed?

 

my laptop isnt much of a game pc, so i have most video stuff turned down or flat out off

including reflections, i do not believe i have had any parts go missing on the tables i have

 

And your CPU outruns the i5 3230m in my laptop

[htamas]

BBBB = Bugs Bunny's Birthday Ball.

This one is weird because the latest version has the stuttering with reflections on, but the previous one didn't.

 

One example where turning off the playfield reflections causes a problem is Al's Garage Band, where the bottom of the pop bumpers turns black.

I remember seeing some issues on another table without playfield reflections, just can't remember which one it was.

[wiesshund]

BBBB = Bugs Bunny's Birthday Ball.

This one is weird because the latest version has the stuttering with reflections on, but the previous one didn't.

 

One example where turning off the playfield reflections causes a problem is Al's Garage Band, where the bottom of the pop bumpers turns black.

I remember seeing some issues on another table without playfield reflections, just can't remember which one it was.

 

Oh, i have bugs bunny, not sure on which version, but i only put it on like maybe a month ago?

was not noticing any issues with it, but maybe it is not newest version?

11-13-20 is the release date in the table info, so maybe a bit older than i thought

these are the graphics settings i am presently using

[htamas]

This is what I'm using:

 

 Capture.PNG   73.9KB   2 downloads

 

The funny thing is that the post-processed anti-aliasing setting doesn't seem to make a difference, if I disable that, those tables still stutter. Turning off playfield reflections fixes them all. Tried vsync at 1 and 2, no difference.

Your video settings are much more demanding than mine, so that would confirm that my CPU may be the culprit.

Edited by htamas, 10 January 2021 - 03:49 AM.

[wiesshund]

post processing is off because to me, it does not look good
for what small jaggies it removes, the overall blurring effect is not worth it to me

takes away the nice sharp clarity.

 

Runs ok, just do not like the appearance.

 

Hey when the table stutters on you
does it still do it is you disable b2s in the nudge prefs?

I had tried 2 backglasses for it that for some reason were mad awful in the b2s comms area
would up just extracting the nicest graphics parts, and using them to make my own b2s, that did not have lights tied into every possible table function

I mean they looked great with 10,000 lights going crazy but all the b2s comms was not helping the table much.

I guess i could have tried tweaking one of them i guess, but i decided that table really did not call for a bunch of crazy lights
i went with doing colored LEDs instead, and only the actual function lamps tied to the ROM, which i think amounted to like 4 lamps.

 

some backglasses create a heck of a lot of data processing
 

[htamas]

No, disabling B2S doesn't help... still stutter, perhaps a little less than with it, but not much difference (if anything).

[wiesshund]

when i get a chance, i will try bugs on my laptop
and see if there is any ball chatter

CPU wise, yours ranks better, but the numbers are not super far apart

so it might at least make a comparable test

[htamas]

Try Arena, that one stutters way more for me with the reflections enabled.

The FX is not very good regarding single-core performance, Intel is usually better in that.

[wiesshund]

Try Arena, that one stutters way more for me with the reflections enabled.

The FX is not very good regarding single-core performance, Intel is usually better in that.

In it's time period, your CPU was kind of beastly
I was going to get an 8370 at one point, the prices began dropping a lot, but then i guess AMD decided to put a shotgun to them, because they released ryzen 5's cheaper
probably to drive sales to their new flagship

 

I considered (for about a minute) a 9370
Be between the much higher price, and the fact that the thing doubles as an industrial oven, and darned near needs it's own dedicated PSU and it also performs worse at mundane tasks, i decided that was a bad idea.

but did really well at heavily multithreaded stuff, and could heat a barn i mid winter

 

have you ever run a performance monitor or logger and run through something suffering stutters
to see if any particular area of the system is showing high load or utilization?