A word or two about images in Blender

Hello, I’m not new here, I’ve been visiting elYsiun forum for quite some time but just got registered (why? I’ve got 16kBps modem conn - that explains). The subjects of this thread are images, more specificaly, images in Blender. I’m not a great modeler (I’ve used Blender occasionaly, couple of hours a month since 1998 I think but I really got a grip on it in a past few months… no results yet), but the thing that annoys me the most is the Background Image feature in Blender. Why? Well, Back in my beginnnig years I had P150, 32MB (upgrade form P75, 16MB) and when the BG pic feature was introduced (v. 1.67?) I tried it and soon after a few zoom ins it was clear that it was impossible to work with - you wouldn’t want to play a game with 1.2-1.5 fps. So I dumped it. Then I got P350, 128MB. “Let’s model somethin’ usin’ a picture ref’rence!”, I said. Cold shower - nothing changed, it still went like doom 3/high details on an average office computer - in short terms, a slide show.
I thought things will change when I bought a celeron 1.7GHz machine with 256 MB’s of DDRRAM. Still the same old story. So I ask: wtf? Why couldn’t they [the Blender team] make it like in any other 3D program; 3dsmax is 10 times more demanding on behalf of resources than Blender and still, references work fine? Yet another contradictory: when placing UV’s, everything is bareable, it works smooth. Why it isn’t the same in 3d view?
And another thing. When taking pictures of a human face for use as a reference in a 3D prog, theye are in perspective, right? Well, shouldn’t they be… let’s say de-perspectivized™? I mean, you’re using perspective reference to make a perspective model but still place verticies in ortho-normal(or is it orthogonal?) space? Just thought of it, and now I think I know why my models look so staring (from to stare: look with eyes wide open to something, usually in confusion)

Just a suggestion…well three actually. Try loading some .blend files for reference and look at the settings. Try set your camera to 200 when modelling faces - cancels the perspective distort. Try have fun :wink:

It’s best to use a photo taken with the camera quite far away from the head, using a lens of long focal length. This will prevent or minimise the distortion. ( especially from the front view)
If not, the thing nearer the camera (the nose) looks bigger than it really is, and the stuff further away (ears) look too small.
Portrait photographers know this, so they normally use a longish lens.

In 3d apps you should model with no perspective at all.
When you view or render your model through the blender camera view,
it will have the correct perspective, depending on the ‘Lens’ setting.
I think the number in this setting corresponds to the focal length of a 35mm camera.

So if your head models look distorted like you say, you could move the camera way back and set the lens to 200 or so. That would maybe look less distorted, but obviously wouldn’t really solve the problem…

I would always advise anyone doing head models to use reference photos which are appropriate (i.e. taken with a longer lens)



I really hate the Blender developers :< for the making of slow background pics. :<

I’ve experienced that problem with Blender running in Windows using an ATI card. In Linux it works just fine, or an nVidia card in any OS.

You can try loading a lower resolution resolution image.

You can also try to UV map the reference picture to a plane.
On my computer the background images are slow but planes with references work like a charm.
There’s a script called rotobox that can do the UV mapping for you.