View Full Version : The Modeling Process: A Consensus
09 September 2003, 10:46 PM
I want to start a thread discussing different workflow methods. I often wonder what steps other artists take in creating a model from start to finish. A few of the questions I have are as follows:
1.) Do you sketch out a concept before creating the model or are you more apt to just sit down and conceptualize in the software app?
2.) If you do make a sketch first, how often do actually put that image in the background of the viewport as a guide for your modeling?
3.) What initial pose do you find works best for rigging? Arms straight out to the side, arms down at 45 degree angles, or arms almost straight down?
4.) Is it wise to create a model in its final pose if there will be no animation added?
5.) What's the avereage time it takes you (in approx. hours) to create a model from start to finish? Please respond to this by mentioning if it's low-poly or high-poly.
6.) Do you prefer the "form-down" technique or "detail-up"? In other words, do you like to start with very few polys and add them as needed or do you create many details and optimize later?
Please don't feel obligated to answer all of the questions if you don't want to.
I hope that this thread will be pinned to the top of this forum, so I'd like others to respond with other good question I may have missed. I will edit my initial post and add that question if it fits with the topic.
09 September 2003, 03:28 AM
reference images are more important than most think... it give the right proportion and helps when pulling and pushing CVs )i like patch modeling a lot)
1) i create some refernce images on a grided paper, this way everything is properly alligned
2)i have the refence images until i do not need them anymore, which is when rigging !!
3) arms straight out works fine (or should i say best)
4)no it's not wise to crate a character in final pose... what if you decide on a different pose at some point in time
for a nurbs patch model it takes (me) 1 week or so, for some high details (without texturing)
09 September 2003, 03:04 AM
Even though I asked the questions, I feel I should contribute some of my own answers to them.
1.) I always do a sketch of what I want to build. Albeit, usually a rough sketch.
2.) I rarely put the images into my viewport backgrounds. My reason for this is because I haven't set out to model something highly realistic. If I was going for accuracy I would do so. So far any character models I've built have been very stylistic.
3.) I've recently learned to no longer model my character with it's arms straight out to the side. Next time I'm gonna pose them at 45 degree angles down. This way the shoulder deformation should be more predictable and getting the correct arm length should be easier.
4.) I don't think it's a good idea to model in the final pose. Like anoe_nomus said, "what if you decide on a different pose...". Although, I read an article by Bay Raitt one time and he sort of encouraged it.
5.) I wish I kept track of my hours modeling, but I'd say it takes me a good 40 hrs to pump out a 4000 poly model. As time goes on and I get in more practice this figure should hopefully go down.
09 September 2003, 02:58 PM
1.) I've always got some sort of reference image. Sometimes it nice and simple/rough; other times I go all out, sketching the muscles, making sure everything's clean, etc.
2.) Usually I'll start out with an image in the viewports, then get rid of it once I have things roughed out.
3.) I always have the arms at 45 degrees.
5.) I used to model fairly quick (never timed it), but recently I've became obsessed with edgeloops and box-modeling. Now I model at a snail's pace...
09 September 2003, 09:21 PM
Yes. I've recently changed my modeling habits to where I'm always aiming for quads and edgeloops. This does take longer, but it's the best way to poygon model.
09 September 2003, 05:20 PM
1.) Yes I do sketch out a concept. I'm a traditional guy first, and pulling things out of my head in max bursts my head.
2.) Always, always, always. I don't like having to think about volumes while I model.
3.) Anywhere between 50 to 25 I guess...I dunno, I just try them out at different angles in increments of 5 degress, so they're easy to rig. But never straight out.
4.) I don't think I've ever come across a situation like this.
5.) I guess about 30? That's a model, rig, and unwrap. No anims, textures are a whole other thing. I'm horribly slow. Real-time models of course.
09 September 2003, 12:08 AM
can i add one more question to the lot?
do you prefer "form-down" technique or "detail-up"
09 September 2003, 05:14 AM
Good question pencil_head. It's been added. Assuming I understood the terminology. Personally I've never heard the terms "form-down" and "detail-up" but please check to see if my translation is correct.
To answer that question I would say that I mainly detail-up. But form down in the polishing stage.
09 September 2003, 07:31 AM
In reply to the "Is it better to model in the final pose?" question, I have one word: symmetry. Its always easier to model one hand spread out and just mirror it than to model two hands in different weird gestures.
I tried to model an Austin Powers figure once in dance-like pose and half way through I started bleeding from my eyes and ears.
Just rig it and pose it later.
01 January 2006, 04:00 AM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.