How to tweak simulation parameters intuitively?

[magneto]

Hi,

 

Anyone finds a lot of DOP parameters not very intuitive to tweak? I am not saying it's badly designed but it's not very straightforward to reach the right values you need for the simulation to behave a certain way.

 

For example Wire simulation DOPs. The parameter names are also not very descriptive. I assume they are named to be physically accurate but for example "Linear Spring Constraint" doesn't easily stick in my mind, compared to if it was named "Stretching Resistance".

 

Any feel the same? Any tips to get better at faster tweaking to get the desired effect?

 

 

Thanks

[kleer001]

I've run into this effect. All I could do was keep up a window of the help page, go back and forth until I kind figured what I needed to change.

 

A set of example ranges for materials would be worth diamonds. Spaghetti, twine, hair, nylon, silk, bungie cord, etc... It's certainly scale and geometry dependent, but that would be part of its charm.

[Sierra62]

I completely agree, it can take a lot time to know what each variable does, and the naming conventions (or help cards) don't always help.  Personally  I have saved out a lot of different presets for a variety of materials.  Granted I have always found tweaking the wire solver settings akin to voodoo, I keep sticking in new variables hoping this time it will give me the control I want. 

[ryew]

Unfortunately, the help files are nowhere near a useful as they once were.  Where you could once find example scenes and descriptive attribute explanations, now you find no example scenes (or worse, outdated/obsolete example scenes) and bare minimum explanations for attributes (or even none in some cases).  

 

The growing lack of effective documentation of new or completely overhauled features and workflows (let's face it, the vimeo demonstrations of clicking a shelf button don't even remotely cut it) make for increasingly frustrating Houdini sessions that leave me looking for other solutions outside of this package

[Diego A Grimaldi]

Maybe we should write this as a request for H15 or H16??

[Farmfield]

Sliders.

 

[firecracker....................................<->............................Birth of the known universe] 

 

 

To be honest, I think the idea of shelf tools as a starting point are very suiting for the type of application Houdini is - I really don't think Houdini will ever be easy or intuitive unless you get an enough grip on how the software ticks, then more and more things will start to make sense and a lot will start to feel pretty intuitive. 

Edited August 6, 2015 by Farmfield

[magneto]

Sliders.

 

[firecracker....................................<->............................Birth of the known universe] 

 

 

To be honest, I think the idea of shelf tools as a starting point are very suiting for the type of application Houdini is - I really don't think Houdini will ever be easy or intuitive unless you get an enough grip on how the software ticks, then more and more things will start to make sense and a lot will start to feel pretty intuitive. 

 

I think that's a different thing. The ability to customize the behavior of your setup using other nodes etc after the initial properties are established will always be up to your skills, but to get the initial behavior right using the default nodes should be faster and more intuitive.

 

In Wire Object example, can we have scale independent parameters that sets Linear Spring Constraint, etc internally? Maybe a normalized range that covers the most useful range of materials while still allowing to go beyond the normalized range? Doesn't Disney's PBR shader work like this? It's much easier to reach a good value in a normalized range than an arbitrary one that can be anything.

 

Using normalized parameters you can set these properties very quickly and then customize your setup in any way you want using other nodes etc.

 

The latter is what makes Houdini unique and powerful but the former step is still far from optimal IMO.

 

Discuss