1002 replies to this topic

[The Loafer]

Mukuste: thanks. In my case, I never unplugged it but the moment I adjusted the dead zone is when I think it made a difference. At the moment, it looks like the gain will need to be put on stronger as it's less effective than before but I really haven't messed with that now that I at least got it working.

Interestingly, while trying to mod for my own use Melon's TZ, I got that similar vacuum effect again, took a look and I had not adjusted the slope. The default I think on his TZ for standard physics is 9. Yeah, don't recommend that setting but if you wanted to see what some of us meant by the ball seemingly accelerating while going down, that is what was happening, so try it and leave it at slope of 9 and see for yourself. I initially thought when I saw that, perhaps I had forgotten to change the slope but I got that effect with the WW that was saved in that thread, so maybe it was a combination of slope needed to be reduced, the dead zone needing to be set and some of the global settings I had setup where messed up (maybe I was setting them in the table when I didn't have to... That is what happens when we setup something in the middle of the night lol). Btw, yeah I was playing the non-BMPR version of drac but am about to test the BMPR-less 8-step version that should be working ok. Hope so, it is a nicer table, with great ball collision sound effects, etc.

The only phys-related commentary that I can add which I believe you are already aware of are:
- in many cases, as mentioned with phys3.exe, the ball starts jumping around when captured on a flipper. This never happens on phys 2, so looking forward to a fix if there is one as it renders many tables unplayable (well, they are playable, just too weird eheh )
- because If I am not mistaken, the physics version and the vp9.9.0 share the same registry, therefore it messes things up with the 120 setting or the nudging settings which need to be different. For this reason, I hope there is another version, or if there is not, hope the vp10 betas will have their own registry settings. Its a must IMHO.

Aside from that right now not much to complain about

UW: the cftbl mod is incredible and plays great. The only flaw is it plays a little easy but some VP tables are like that. Played two games and got the film completed with creature multiball in both BUT I had a great time in both games tried the non-physmod afterwards and after two balls, I stopped playing! It's not that the standard version is bad, its great! It's just that I am now noticing some random tidbits with the flipper physics especially that every now and then just feels wrong, so since I have the modded physics version, no sense in continuing. Looking forward to the reaction of the masses to this.

Thanks melon for an incredible table too, super well done! Experimenting with your TZ, I had a ball fall through the table but as noted above I think I can fix it. Your TZ is awesome too.

[unclewilly]

Can't remember. But I may have left the drain and outline blocking walls on the table. If so just delete those walls.
I'm a sucky player so I need those for testing. I also didn't mess with the nudge so let me know if the force needs to be bumped up on that.

[BigBoss]

I have my sidewinder freestyle set at 60% gain, 20% deadzone and its a little strong. I am guessing closer to 25% gain is better. This is down from 150% in the old versions.

The Dracula has all the bmpr stripped so that wasn't impacting it. I think how analog controller was the problem. I helped a friend set up a cabinet last night and the analog controller was causing the table to feel about 10 pitch. Went I to windows and saw it wasn't centered by slightly down. Re calibrated it and all was well

[slashbot]

This is going to be long

Thanks for the detailed feedback, Rob! It's good to hear that the new feel is growing on you.

It's good to know that unplugging the VirtuaPin controller fixed the weird acceleration you saw. This might be related to the fact that certain nudging devices are 10 times stronger in the physmod than in 9.9, as reported by Slashbot -- or was it 10 times weaker? What is the factor you had to adjust your gain by? I had a look at the nudging code to try and fix this, but it's such a convoluted mess after what the UltraCade guys did to it that I haven't gotten far yet. Maybe someone can tell me if there are different ways to set up nudging with an analog joystick and with an accelerometer? That might be the key.

Just to be sure -- the version of Dracula that you tested in the physicsmod did not use BMPR, right? Because that would seriously mess with the physics.

As for lag: in the physmod, you should generally get away with setting the frame limiter/vsync field to 1 so that you get true vsync at 60 Hz. However, this might increase input lag a bit since the frame times are now longer, so to compensate, you could try changing the Max prerendered frames setting in the Video preferences to 1 if you have it at 2.

If you get stutter or perceived non-smoothness, you should always check the F11 display. Two things to look out for: can your hardware reliably sustain the frame limit that you have set (60 or 120fps)? And do the max frame times deviate noticeably from the average frame times? If these two factors are both ok, then I would expect smooth gameplay in the physmod version. VP9 has bugs in its physics code where you will likely never get fully smooth gameplay at 60fps, I consider 120fps the minimum for that to run well. So, in a way, the physmod version actually takes some load off the GPU since it can now run at half the framerate, though it does tax the CPU significantly more of course.

As for adapting back to 9.9... it's funny how the brain works, isn't it. I can deal with the floatiness after a few games, but the flippers tend to drive me crazy when I go back to VP9, and there are only some very well tuned tables which remain playable in VP9 for me.

For the physicsmod i have to set nudge device 10 times weaker.

Mukuste do u have an analog nudge device, i have a spare usb hud-g board i can send to you?



Sent from my iPad using Tapatalk HD

[mukuste]

Loafer:

 

Ah, so it was the dead zone. What device are you using?

 

At a slope of 9°, I'm not surprised you noticed a certain excessive acceleration in the downward direction

 

I know about the jumping ball and have actually fixed it. One of these days I have to get a new build out... busy times still...

 

The new registry key will probably become necessary; I'm a bit reluctant to make the change because as soon as we did it for 9.9, we got a neverending stream of "bug reports" where people simply didn't update their configuration, but it seems there is no easy fix for the nudge factor; some devices need scaling and others do not or to a different degree. Best would probably be to import the VP9 configuration at the first startup.

 

It's interesting that you mention that the table seems to play easy; I noticed that myself while tweaking some tables. If the physics are more accurate, shouldn't the challenge be closer to the real thing as well? Maybe it's something the table authors will still figure out, like more bounciness needed.

 

 

SlashBot:

 

I don't have a nudge device. It would be nice of you to send it to me if you don't need it. You'd have to send it to Austria; is it large? I can pay for shipping if needed.

Edited by mukuste, 06 May 2014 - 01:25 PM.

[The Loafer]

Mukuste: I own the virtuapin controller sold from this site. For me it works great and I love that it auto calibrates.

[BigBoss]

There's a lot of variance in behavior between some of these accelerometers. I had a nanotech (was horrible, wouldnt calibrate worth crap), an ultimarc u-hid-a which was better, and have instead been using the insides of the MS sidewinder freestyle. These are fantastic. Very reliable and steady in the idle state. Unfortunately, they're no longer being made so you have to scrounge up one. I have two spares right now. I wish you were close to me. I'd loan you a real pinball machine for a year or so to iron out the physics.

[freezy]

If you're talking about U-HID G support, that'd be excellent! 

[Slydog43]

I think we need to donate a pincab to mukuste for all his incredible work.

[ringorian]

If you could add a second joystick which could be used would be great .. The nanotech is to nervous and i cant use my joywarrior quakecatcher for it since its also noticed as a joystick ., any chance mukuste ?

[slashbot]

Mukuste Yes its the U-HID G device (http://www.u-hid.com/home/uhidg.php), i now have a Microsoft Freestyle Pro and works better for me, but the U-HID G should do the job for testing purposes just PM me your adres and i'll send it to you, but i cant give you support for it lol.

Also give me your email so i can send the the latest firware for it in case you want to reprogram it.

[DJRobX]

The nanotech isn't reallly "nervous".    If the centers are jumping around a lot, the min and max values are too close.    The bigger problem is that it doesn't auto-center itself, so it's impossible to calibrate properly without some outside help.    It will drift out of calibration as your cab warms up, or possibly even from heat as your CPU works harder when tables are being played.

 

I wrote a utility to monitor the values and reset the centers when stable (a few seconds of low movement), allowing me to set a much tighter dead zone.    I've also injected this logic into my local VP source code (so it doesn't drift mid-game).   It works extremely well once these checks are in place.   

 

The irony about the sidewinder freestyle pro is that I had one in my garage pile for what seems like forever.  I think I finally took it to E-waste recycling 6 months ago because I figured I'd never want to bother with something that needs that old joystick port.

[BigBoss]

The nanotech device I have here flutters all around like crazy when idle - so bad that sometimes it registers *extreme* hits when nothing is happening. I thought the thing was defective so I got another and it also does the same. The sidewinder just sits idle when idle. Your auto centering idea sounds really cool. Why not submit it to the source tree? Cant it work for any x/y axis nudge device?

[JohnnyDoe]

The nanotech device I have here flutters all around like crazy when idle - so bad that sometimes it registers *extreme* hits when nothing is happening. I thought the thing was defective so I got another and it also does the same. The sidewinder just sits idle when idle. Your auto centering idea sounds really cool. Why not submit it to the source tree? Cant it work for any x/y axis nudge device?

Yeah, I got one too. Could'nt have been a worse buy, since I paid through my nose in shipping and tax and sh#t.

[Pinball999]

You guys can check the first page of my build, you'll find a link to dowload an .exe that helped me to fix the centering issue.

[mukuste]

Download a test build of the physics mod and the accompanying default table below.

Press F11 to get a visualization of ball spin.

Old VP tables will not work with this version unmodified!

 

Changes since physmod3:

• New keyboard nudging.  Table displacement and velocity is now modeled in a physical way, and table swings back after a nudge. New parameter Nudge Time in table Physics settings (on backdrop): determines the duration of the swing and backswing (a setting of 5, meaning 50 ms, seems reasonable). Nudge forces may need to be adjusted (look for Nudge <angle>, <force> calls in the table script). Angle is the direction the table is pushed in! (Note: this is currently buggy, it's better to wait until the next update.)
• Tweaked hold coil logic (strength increased to 2/3 of high power coil, dead zone increased to 3 degrees). Mitigates bounce back of balls hitting the flipper ("trampoline flippers").
• Elasticity Falloff is now configurable for flippers. 0=no falloff, 1=half the elasticity for a collision with a velocity of 1 m/s. The old default was 0.43, maybe a lower value like 0.3 works better.
• The ends of ramp side walls (vertical edges) now properly collide with the ball. Some tables need updates.
• The top and bottom edges of thin (in vertical direction) walls now properly collide with the ball. Fixes bouncing of ball on held flipper on many tables.

This version also has the latest updates from VP 9.9.0.

 

Basic instructions for adapting tables: https://github.com/c...to-vp10-physics

Description of physics parameters: https://github.com/c...ki/VP10-Physics

Attached Files

•  VP_physmod4.zip   686.33KB   143 downloads
•  default-table-physmod.vpt   128KB   63 downloads

Edited by mukuste, 13 May 2014 - 08:09 AM.

[unclewilly]

I hope your move went well.
Can't wait to test out the new nudging.

Is the default table set up with testable parameters for the new nudging.

Really appreciate the work you've done here. Looking forward to your contributions.
It amazes me how far this has come in such a short period of time

[jpsalas]

I agree with UW VP has never played so good as it does now. Thanks for  all your work, mukuste!

(and don't forget the wall target error, the ball only creates a hit event the first time the ball hits the target, the successive hits are ignored )

[Slydog43]

thanks for another physics mod, really making VP play so realistic its just amazing.  Are the keyboard nudging changes going to change the analog nudging at all? in future?  This is something that I think needs to be perfected.

 

everyone should call their mom today and say thanks

Edited by Slydog43, 11 May 2014 - 11:56 AM.

[The Loafer]

Yay! Thanks Mukuste, this is fantastic! I am away all day at ComicCon, actually can't wait to get back home now. Wife's flight from Abu Dhabi better not be late! Lol

Edited by The Loafer, 11 May 2014 - 01:04 PM.