How to model from eye or memory

Currently, I can only model from blueprints. There must be Front, side and/or top. How can I start to model from reference images?

From memory you don’t. Even the simplest things need references because memory is bad with forms and proportions, can’t draw modeling notes on it, and it doesn’t inform about the variations of the object.

Modeling from references is the same as trying to build a real object based on images and data. First concentrating on the proportions, then forms, and then part separation. If those aren’t known, are too vague or too complex to interpret from the references, the modeling workflow starts from figuring those out in the 3D space with various possible methods such as blockout, boolean operations, kit bashing, or sculpting.

Then modeling can move on to the structures that includes the forms, part separation, and fulfils the requirements from the pipeline and the targeted end use. If the forms are already in the 3D view, the structure can be modeled on top.

Well in addition to what JA12 said: Are You able to draw from eye? That’s an essential question. Most of the so called “awsome” modllers, have the “eye-of-an-painter”. What does it mean? If you draw on paper you have to find out, what are the basic shapes? How are they placed in the real-world? How are the propotions? (If you see those art-students in town, practicing, you mostly will notice, that they hold their pen in front of there eyes - arms long - aiming on the object they are drawing. What do they do? they get the “messurement of propotions”.)
What are the next “underlying” shapes? How do they fit into the propotions of the “underlying base (canvas) forms”.

The rule here is from simple (like a box) to detail (like windows) -(as far as you talk abnout drawing a building)

If you know those rules, then there are fewer issues to solve, when you go into 3D modelling, because the process is quiet similar, without the “translation” of 3D perspectives into a 2D picture.

But! If you model in 3D with a reference picture - so 2D - then you have to do the “translation” vice-versa. :slight_smile: It’s like turning the car and go the way back you came. But as you know, the sense of view is quiet differnt, like that, when you came up the other way around. :slight_smile:

The eye has some advanced capabilties in a real world enviroment, but you have to learn to “trust your eyes” and “get the feeling” to push this into a 3D enviroment on modelling.

But sorry JA12: I do not completely agree to your statement: “…by memory, is not possible…” It is possible. Concept artists are drawing by imagination - so what else is it, than “memory”? Okay, the “memory-part” is not fitting a single, special, (f.e.) face with all details. But the memory consists of seperated details, which they put together into f.e. a new character. The “reference” in those cases are stored in brain. And why should this not be possible also in modelling? I think, those awful creature sculpter are a well example for capabilities like this.

By modeling I mean technical modeling consisting of correct proportions, forms, and proper structures that take into account the requirements from the pipeline and targeted end use.

Things done as preparation for modeling that only concentrate on proportions and forms I consider as design, as those are interpretations of the modeler and done because the reference material doesn’t inform more accurate reproduction.

Concept artists design based on references, their visual library, and the design parameters. The end result of that is correct because it’s a new design, where as modeling can’t be accurate to the design when the modeler can’t see it. That’s also the reason in-house concept artists also support modelers, producing orthographic projections and modeling sheets of a design.

So could modeling be done accurately without a reference? Yes, if the modeler is extremely familiar with the subject matter and the construction is straightforward enough to get the features reproduced, and there are enough memorized data points to do it. But then you only have to look at beginner models of a coffee mug to know how easy that is. Those are often the wrong size and the walls look like they were made out of thick wood or concrete, and that’s a memory reference of an object that is literally under their noses multiple times a day, every day.

The point is that design and modeling are very different and one shouldn’t model until the design is in visual form. Both design and modeling are busy enough tasks that they should be kept separate with subject matters of any complexity. Concept artists using 3D don’t bother with topology, and master modelers with design sense could try but will fail in some aspects of design or modeling, most probably design, if they jump straight to modeling.

Then, you first create blueprints from eye/memory.
It will also help you to define & refine the blanks.

It’s a standard procedure (taught in primary schools even)… as working from an idea, a concept (in engineering, architecture, theater, film, games…). After many hours grinding (gathering experience), you’ll learn and be able to hold on that mental image (vision) predict any caveats, problems and go from a sketch directly to 3D… image, animation, print…
eg.

Ok Ja12, with this I can agree. Sorry, the lonesome statement as it was, seemed a bit to untrue to me.

And I did not state, that it’s an easy stuff… :no: … and as in anything else… it’s training, training, training and then again training, training - what? My old scoolmaster of photografy said: “The eye, boy! Your eyes!”.