Blender Modeling and Real World

[pwillard]

Clearly we always need to find middle-ground when converting what we see in the real world into a workable model. One of the things that always bites me is remembering that we like to visually have things line up... related to couplers.

I just recently needed to create a cushioned air hose solution, so I needed to take out my notes.

So here are the values I use.

Coupler Height: 0.89m
Coupler Hose Joint: 0.43m
Coupler Hose Joint Hanging: (uncoupled) 0.23 (seldom used by me, actually)

The AirHose joint should be located in the center space of the coupler opening.

Screenshot 2025-01-08 090205.jpg

[EldoradoRailroad]

Yeah, lining up of couplers is a bit of a hit and miss. I have noticed that some payware cheats by squatting the body of the engine against the trucks which look plain wrong. But if you lift the body the couplers are OFF in terms of height by quite a bit. I think you have to accept a compromise with accuracy. It has been a few years but I used a spare coupler in Blender to align the coupler of the current project. That alignment is rarely perfect, but if it out by a few cm (some times a bit more!), "a man on a bicycle" would not notice. But you and I would of course.

[pwillard]

wipshot2.jpg
Another part of this is deciding WHAT will get modeled and what will not. The temptation is to model the plug doors used on these cars, but then photo-painters will be gnashing their teeth since they can't just paste a photograph on the car. Its a tough choice... especially if not many photos of the car even exist. These cars, for example, had a short life due in part to GATX being sued by Cryo-Trans over patent infringements... so its not like you can go out and photograph one now.

[Erick_Cantu]

AAR standard for coupler height is 32.5 inches to 34.5 inches from the top of rail. I set mine to 34 even unless the geometry of the frame rails requires a different height (typically a bit lower).

In terms of detail, I solved this issue by having separate product lines. This allows me to preserve my sanity with simple models for photo paints between more complex projects. I don't do a lot of super high detail models anymore because I find it tiresome.

[pwillard]

I guess my couplers are roughly correct, if not a few centimeters too high since I adopted ~35" as my value from the ground (not actually from rail height)

Honestly, I don't enjoy making models that are "too simple" or so generic that they can be "many things." But then, yeah, you can get bogged down with extra details, and the model becomes a "never done".

So yeah, I'm still trying to figure out how I want to implement the box car doors.

Screenshot 2025-01-10 081043.jpg

[scottb613]

Hi Pete,

Looking good.


Knowing Erick's attention to detail - I just use his cars as my standard when adjusting coupler height.

Regards,
Scott

[pwillard]

The car I'm working on had a lifespan of roughly 5 years, and the CryoTrans cars are similar, so this could morph into CryoTrans cars. While GATX cryogenic cars exited service in 1995 after losing the legal battle, this car design was scrapped once it became economically non-viable when the cost of C02 became an issue.

Many cars were also branded with companies like McCains, Carnation, and Simplot logos. These might be interesting in OpenRails consists near some TigerTrains reefers , which are the more modern CryoTrans reefers.

[pwillard]

So I've only just started on the plug door, and I've decided that these things are complicated.

Screenshot 2025-01-10 154311.jpg

[eric]

Further complicate this by animating the door open/close including the little roller track wheels?

[pwillard]

Hah. um no.
That is the path to insanity. I'm already too close to that.

[Erick_Cantu]

Knowing Erick's attention to detail - I just use his cars as my standard when adjusting coupler height.

I make mistakes, too - I think it's best to live by the mantra that "all models are foobies until proven otherwise."

[mrmosky]

Nice model!
In my opinion we should do what satisfies us, as far as modelling is concerned and not try too hard to guess what others might like.
Having said that, we may not be too far from having better grahics in OR with glTF support. At that time, efforts that we are putting in now may pay off.

[pwillard]

Well, I cannot wait until we can start using PBR materials, if that's even in the cards.

BTW: This project is based on Magazine articles "Model Railroading Sept 1993" and "Mainline Modeler Sept 1992"

I also noticed that the FIRST run of cars had a different coupler cushioning method.

Also, a collection of reference bits like:

Screenshot 2025-01-11 095318.jpg

[superheatedsteam]

Chris Jakeman has advised in a post at Elvas that glTF will be reviewed as a top priority after 1.6 is released and see no reason why the interval after 1.6 need be lengthy.

Has anyone here dabbled with glTF or materials? I am using the free BlenderKit add-on as it has over 3,800 procedural materials which saves me having to create them from scratch and helps with prototyping.

However, getting materials to export using native Blender work flow is a monumental pain. I have only recently found a paid add-on called SimpleBake that is well worth the 20USD. A handful of reviews on you tube was enough to convince me of its worth and I was not disappointed.

[pwillard]

If OpenRails eventually supports glTF, we won't need a custom Exporter anymore, as we could export directly from Blender, which would be nice. I imagine the existing S file format might not be good enough for the layered materials approach, though so that export would remain the same. The improvements in realism would be super nice, though.