192 replies to this topic

[bord]

I got lucky and never did a layout based on manual drawings. Learned from the mistakes of others. As you know, even minor changes in layout can make a difference in gameplay.

Scanning is the only way to go for project assets unless your project is a true rarity or you have a level of self-loathing that exceeds mine. Model it right once and you're in great shape for VPE or VP11 or whatever else comes next.

Progress looks great, Flupper. One of the greatest games of all time.

[3rdaxis]

When I first started making 3D ramps for VP authors I always wondered how they were figuring out the height values, turned out they were just making educated guesses.  Basically all ramps created without height measurements are likely flawed.  I had measurements for ramps on Gilligan's and Jurassic Park and it made a big difference taking all the guess work out of it. 

That is 100% true. Those blueprints are NOT accurate. Made that mistake once... Once:)

 

 

 

 

Same, never again lol.

That's why I never start a project 3d modeling a machine unless I either own that machine or have direct access too it. it's guarantied impossible to get it right or even close.

 

Even if you have the table and can hold the ramp in your hand it's difficult to get it perfectly right, even with a 3d scanner. (Unless you can 3d scan it "bolted down" to the playfeild) The best you can do is too have the machine and take as many measurements as you can, preferably at critical locations. Slope, descent, balance point (point at which the ball will either continue forward or travel back down the ramp) width, height, etc. 

 

Take for example TAF, despite the fact that I was lent the machine to model (thank you Slydog43) There were only a handful of locations I could get an accurate height measurements on the ramp with it bolted down (see picture below at beige blocks).

But what I was able to get was the balance or "tipping" point of the ramp which I feel is so important and IMPOSSIBLE to know otherwise.  

 

Edited by 3rdaxis, 24 February 2020 - 12:43 PM.

[ClarkKent]

Our latest test building the table in real size results in much better physics without messing with the physics in script. We only added triggers for the tip of the flippers to add mass to the bats to prevent too fast balls. I never understood why building tables in the weird VPX units scale...

Edited by ClarkKent, 24 February 2020 - 06:34 PM.

[JLouLoulou]

None of manual or PF scan are 100% accurate. On Stargate, when i make the sheet metal ramp at the beginning from my real Stargate, the left fit perfectly, but not the right one and the outhole kicker ramp too. This is the same with my wire ramps, which they need to be adapt

Envoyé de mon Redmi Pro en utilisant Tapatalk

[flupper1]

Small update, just to let you know I started in Blender (2.82):

 

 

On my iphone, I can see that the image is a little dark (I have a bright monitor), so here the screenshot with some more exposure:

Edited by flupper1, 04 March 2020 - 09:43 PM.

[Fleep]

Looks supper :-)


Sent from my iPhone using Tapatalk

[JLouLoulou]

Fantastic! I wish to make the same... But I can't... So bad, I'am not familiar with blender. 

[Rawd]

Damn, that is nice.

[apophis]

Looks incredible! Thank you for your efforts! I can't wait to play it.

[senseless]

Everytime I follow a masterpiece WIP I feel really humble and glad we have soo much talent in our hobby. Amazing work Flupper . Really enjoy following this wip!

[CyclonePF]

Those bumper caps look incredible!

[bord]

Looks good, man. What do you use for an environment image?

 

Are you using any new 2.82 features? Lots of cool stuff happening in Blender development lately.

[Ben Logan]

Thanks so much for documenting your progress, flupper. Always such fun to read your WIP threads. 

[flupper1]

@bord:

The Hdri is on this page: http://www.hdrlabs.c...bl/archive.html

and then search for popcorn lobby

 

I also use a overhead soft box light (tip from G5K), something like this (but not this one): https://www.hdri-hub...box-rectangular

 

And I don't think I am using any specific new (well instancing is not new...); I am using E-cycles since that renders much faster. E-cycles costs money, but it has reduced my render times on Totan by 50%

[3rdaxis]

Fantastic work brother. Couldn't have done better myself. 

 

 

 

Sent from my Android using MyFingers®

Edited by 3rdaxis, 06 March 2020 - 09:06 PM.

[Guus]

Wow a another masterpiece is in the making! Thanks for all the great work as always flupper1! We enjoy your tables every week.

[sasa12345]

Amazing uh

[flupper1]

Chapter 5: creating Plastics in Blender

 

In order to get the scanned and cleaned plastic images into Blender I follow a certain process, which I will show you step by step. It starts with the cleaned scanned plastics image from one of my earlier posts, such as this one (this is 1024x1024, the real one is 7000 x 7000 btw):

 

In photoshop I then prepared a black and white outline version of these plastics:

 

This outline contains all holes and the actual plastic contours, even if the plastic is part transparent. This last image, I open in a program called Inkscape, to convert the image to a vector file (.svg filetype). I think I learned this trick years back from Dozer (but not sure who did this first):

 

The converted file (.png -> .svg) is imported into Blender and converted from vector format to a mesh. This mesh is flat and only an outline of the plastics:

 

Unfortunately, the resulting mesh has way to many vertices and some uneveness, due to all the conversion steps. In Blender I first apply a clean up > limited dissolve step:

 

This removes some of the issues but not all, so still some manual work left:

 

In about an hour for all the plastics in this scanned set, I manually clean up the mesh and prepare it for using it in the table:

 

All the plastics:

 

Goal is, to have just enough vertices and geometry to nicely render the table and also, afterwards use the same mesh to export to VPX. If the "Blender render" mesh is too high-poly, it will be a lot of work to prepare the same mesh for VPX (or suffer bad performance in VPX...). With the detail in the shot above, I can easily convert the plastic for VPX, by just removing the holes (the holes will be visible, but part of the texture, not the VPX primitive).

 

Next step is UV mapping (preparing the mesh to display the texture). I do that by using "project from view" UV mapping from a orthographic camera view straight from above the mesh: 

 

Only step left for UV mapping is properly scaling the UV on top of the texture:

 

Finally I move the separate plastics to their required position in the table and apply Blender modifiers to make the plastic have a thickness and nice beveled edges (necessary for rendering):

 

I realize I did use some Blender/3D words you might not know, but I still hope you can follow my steps. Anyway, this is the result when rendering some of the plastics at 8k resolution (I will use 4k resolution for VPX, but just to show the details...):

 

 

This concludes this chapter. And now some full 4k renders with the plastics (already in the exact POV I will use in VPX), lights off:

 

Lights on:

 

And lights off without the plastics, to see what is underneath the plastics:

 

And some render info: 4k lights on render time is about 45 minutes, lights off 30 minutes on my 1080Ti. I use E-cycles to render, which is faster then standard cycles (but did cost me money...), and a denoiser step is used at the end (also e-cycles).

 

Next chapter: 

https://www.vpforums...e=6#entry461866

Edited by flupper1, 08 November 2020 - 04:58 PM.

[watacaractr]

Looks great so far!  

[Fleep]

Looks incredible bro


Sent from my iPhone using Tapatalk