Scaling 3D models. Best practices?

Hitfilm is unitless. Coordinates are just some number without physical meaning. Using the models the auto normalize feature states the model is sized such that it fit inside a 500 unit cube.

So what if I want one model bigger than other. What is the best practice to do this?

With auto normalize I can always use the Transform/scale property to do this relative scaling. Are there consequences to this? Can it be better for quality if the model is scaled in the model importer versus the transform property? The models are vector but does this affect texturing quality in any manner?

If not using auto normalize, are there consequences to display quality if a model is sized but at the default mode importer viewer scale it is smaller than what the auto normalize feature generates?

For the sake of argument, assume there are consequences. It seems different models will have whatever unit scale they want. e.g. At the same unit settings, a Death Star and an X-wing model at the default zoom/scale factor in the importer viewer have completely different sizes. Therefore I have to adjust the scale settings to find the unit setting where a model looks a certain size at default view. Cancel the import. Then once I know that for all models, only then can I use the scale factor to adjust each one relative to the other on import and have the smallest one be at least a certain size like auto normalize.

When not using auto normalize, what on Earth can the various units possibly mean? Meter, Inch, Foot etc. I see no definition of how real world units translate to Hitfilm units. Maybe models in their files can have physical unit dimensions. Every unit listed is a physical unit. What the heck is a "pixel". I guess for a fixed camera resolution, lens and distance one can translate to a pixel.

That is every question I can think of on this topic at this moment.

 

Comments

  • Stargazer54Stargazer54 Moderator
    edited January 2016

    I have to second what @NormanPCN has brought up here.   Bringing in 3D models comes with a little bit of guess work.   I just tried an experiment.  I brought in a model that was scaled at 7' in Lightwave and set the unit scale to 1 unit = 1 foot on import into HF.   Then I brought in a UH60 helicopter model that is scaled at about 65 feet from tip of rotor blade to end of main blade out past the nose of the main fuselage.  Again setting unit scale to be 1 unit = 1 foot.

    The relative sizes laid one on top the other look to be the same in LW and HF.   So as long as you start with the first model at the correct unit size on import and all subsequent models are at the same scale then you should be at the correct scale in HF.

    But, I am confused about what the units in HF really mean after import.  For example, the 3D camera defaults to 700 units from the origin.  So does that mean that once the first import has been done, then the unit scale for the camera units will adjust to the same scale (as well as other units for X,Y,Z)?  Is the camera now 700 feet from the origin?

    That does not appear to be the case.  If I put the UH60 with its main blade tip at the origin and with the model pointed down the Z axis, the tip of the tail rotor nearly reaches the camera (looking from the Top view).   So in this case the camera is not 700 feet from the origin as expected.  

    Looking at the Transform properties for the camera I can see that it is now at 2205 units from the origin.  So it seems that the internal HF units are arbitrary.  Very confusing.

Sign in to comment