>>5765754NTA but to from what I understand you build accuracy by combining copying reference (2d shape training) with construction (3d shape training).
For example if I want to learn to draw a fire hydrant from imagination at any angle the procedure would be:
- Draw fire hydrants from reference, note the variables between each design, the proportions, the materials, what areas get worn and damaged over time, how do you connect a fire hose, common colors, what does a broken hydrant look like (classic water spurting from the top trope?)
- Pick a design and break it down into simple 3d shapes so you can draw it from any angle. Your fundamental perspective training should allow you to do this.
- Draw the simplified 3d form version from memory and then try to add details from your time using reference and your fundamental knowledge of lighting & materials to make it look realistic or effective in whatever style you want.
- If you fail to add convincing details then go back to reference and study it more, alternating between reference and putting away ref to use your memory & 3d form version.
When you are adding objects to your visual library like this you are essentially trying to strip away everything that isn't covered by your fundamental training. Like you are trying to compress an object into a tiny space in your brain that is easy to access instead of storing a complicated 3d object with 6 billion polygons.
Instead of trying to remember a perfect photograph of a fire hydrant you just remember something like
"A cylindrical object 1/4 as wide as it is tall. Top 1/4 is capped like a hat. 2nd highest quarter has two opposing cylinders coming out the side. 3rd highest quarter has a cylinder perpendicular to the last quarters cyclinders. bottom quarter has a design like a revolver. usually glossy red and bolted onto the sidewalk."
But more concise because you are using multiple layers of fundies training and different memory types to supplement it.
