Learning to model(too many tutorials)

[digital_babu]

Hi Jackmar,

Recently I found myself at the same point as you, and what has helped me an immense deal in my workflow and general modelling skill was taking some time to get a better grasp of underlying principles.

In my case one of my breakthroughs was (after reading a tutorial wich I sadly can't remember anymore) postponing modelling and focusing on planning the model out in photoshop. I'm assuming you are talking about organic modelling in this respect, but I guess it's applicable for any kind of modelling.

So what I did was trying to imagine what are the most defining lines (creases and indents) and draw them out on the ref. And from there try to draw quads and create a topology even before modelling.
Another thing I did was buy the 3d-total dvd "Freedom of Teach - Maya Training Organic modelling - 1: Human Anatomy" wich teaches you how to build up a character by following muscle anatomy. The great thing about this tutorial is that it doesn't take you step by step in laying out every poly, but assumes you already have understanding of modelling. So the tutorial forces you to think for yourself on many occasions.

Oh and what I sometimes do if I get stuck in modelling the correct flow is to replicate the situation in 2d plane, and try to fiddle it untill I understand what steps are needed to get the right topology.

Hope this helps,

Digital_Babu

P.S.: I'm in no way authorative on the subject since I don't even yet have had the courage to post any wips on cgtalk. I'm still not confident enough to finish a model :| :).

[newellteapot]

These advices are good, but I also would like to add something. Modeling is how I make a living, so I think I can help.

1)Start with simple, everyday objects
I know you really would like to do the Xwing, but please start with something like a cup and saucer (not a flying one :))

2)Look at the object and divide it in simpler, smaller shapes.
Example, a mug is a cylinder with a torus attached to it.

3)Model one of those parts at a time
Possibly bearing in mind you will have to attach the simpler parts in a topologically clean way

4)Use box modeling, avoid booleans, nurbs etc (for the time being)

5)Post here the result with a wireframe and ask for critique, lots of people will help you :)

6)To check wether a topology of a character is ok, make a quick rig and see how it deforms, if it deforms badly go back to the modeling

7)Avoid triangles and pentagons etc, stick to quads (this is controversial but this is my view)

Hope this helps :)

Monica

[Jackmar]

Thankyou for your replies!

This is the kind of information that really helps me. so much more valuable than step by step tutorials.

I've tried drawing out the topology first and this helps me a huge amount! This is probably the most usefull technique I have learned(so far) about organic modeling.

[tb582]

I guess I'm on the beginning end of what you guys are talking about... I have done a few little things in Maya with balls bouncing around ect...But my modeling skills are defiantly lacking something, can anyone recommend a good beginning book that explains how to model well? IE good geometry etc?

[B-Mac]

Well, I am assuming you are using max.

I know you don't want tutorials, but this one featured on 3dtotal: http://67.15.36.49/team/Tutorials/toaster/toaster_01.asp

It focus's more on learning and using the tools within max and the object you are going towards isnt that important.

I would say learn the tools of the program. Then you can figure out how you would use it on your aliens arm (lets just say you are making an alien arm :P) and how to make it a little thinner and define the muscle more. which tools you would use to get that effect. there is no tutorial on everything you want to make, so learning the tools to apply them to your model is a thing you should be concentrating.

hope it helped

[salmonmoose]

One of the tutorials I've always loved is 'Thinking outside the box with paul steed' it's pretty old but unlike many tutorials, he makes mistakes and leaves them in the tutorial, so it shows you to a degree the thought process of how to attack a model. The result is pretty dated (the model is a character out of Quake3:Arena) but I think as far as tutorials go it was a turning point from paint by numbers to really being free.

Most tutorials give the impression that there is only one way to come at a task, because the author has only showed the way they did it that worked. I'm sure I'm not the only one who re-models some stuff a few times just to get it right.

http://www.loonygames.com/content/1.13/totb/

[stefanosoik]

Most tutorials give the impression that there is only one way to come at a task, because the author has only showed the way they did it that worked.
http://www.loonygames.com/content/1.13/totb/

I agree with that. When i started studying modelling i tried a lot of tutorials and i was always going step by step...the result was to have something without undestanding what i did and why. We must undestand the technique and not the steps to memorize them.

[xDarkvisionx]

The advice you just gave on figuring out WHY the next step exists greatly enlightened me. Thank you. I used to go step by step by step not even thinking about why!!

[FR33K]

i started doing human modeling 3 months ago and also was one of the user just following the steps of the tutorial. But now i stopped and was reflecting about myself and what i have actually done or learned. A much better way or as i do it now when i do human modeling is firstly get an imagination how every muscle of the human body is looking and the second how to make a human look like a musculous man(study books). This I achieve by finding my own way and technics like how to cut or where i cut. I experiement very much with the c4d tools using them in an other way the pros are using them maybe i find better way to optimize something.
an example: doing a rough human body could also be done with a spline drawn in the side view , extrude it , close the caps and throw in the symmetry object so yu have very fast a rough model. To subdivide the caps parts you just make edge loop cuts so you get a subdivided mesh you can work with. Nobody did that before but it has no disadvantages. So find your own way and if you think its then show it the community to hear what they think;-)

[yenvalmar]

theres not really a right way to do it, but theres definitely better and worse results. whatever methods work for you to make clean efficient accurate topology, thats the best method. a lot of people find polygon box modelling and working mainly with extrusions and cutting faces and smoothing the result to be pretty intuitive, i'd say, but then there are all sorts of freaks out there who make great models in ways i couldnt imagine. find what works for you as far as process but just make sure you know what a proper result is to aim for. there is definitely no substitute for expernience and trying the same thing over and over, i think anyone ive ever talked to in cg will tell you its a constant learning process and the more expert people are the more they tend to disagree about the best way to go.. which to most of us is part of what makes it fun, if thats not cool you may be in the wrong career..

i still try to do everything differently from time to time even if i feel confident about one approach because often there turn out to be benefits to learning other methods to do something which become apparent in another situation later on.. for example when i felt comfortable with polygons i did a project with a nurbs character, i would never use nurbs for a character like that again because i think polys is faster but it did teach me about nurbs all sorts of things i have used in many other situations where i saw it would be faster than polys in the years since then so all in all i think it was a great lesson.

[BadG3r]

Well, being in the same situation I can tell you this:

Take the tutorial, follow it step by step. If it uses for example "Cleanup" (in Maya, im not familiar with other packages so far) and doesn't explain why or what. Make a note and comeback later and read the manual. Go through the web, ask in forums (but not before you did the other steps).

After a tutorial you hopefully learned something new. Try to recreate the model from scratch. If you stuck watch the part again (or reread it).

The next point is. Start something on your own.
I did a steam locomotive. Well that was really nice and like the model but i wouldn't consider it mine or my modeling skills. So right now I model an Apple Cinema HD Display. Increase the difficulty by increasing thelevel of detail.
Trust me, a display sounds easy, but there are a lot of snares where you get stuck and stumble. But somehow you manage them. AAAANNNNNDDDDDD.... you get used to the tools: What do I use for what and I didn't like that because. So you approach your own workflow.

Best thing is to keep the manual open. A lot of explanation there. Well just a manual should be ...


Something that helped me on getting used to modeling (I always thought I have to do everything out of ONE cube or so, stupid me) was this book.
http://www.amazon.com/Introducing-Maya-7-3D-Beginners/dp/0782144349

So, summarizing what some had already said mixed with my own experience, I hope you'll get along.
And Im sure, you will.

BadG3r

[Jackmar]

Thank you so much for the replies!

You guys are awsome!

These replies were more than I could hope for.

[cadmonkey33]

Hi Jackmar,
In my case one of my breakthroughs was (after reading a tutorial wich I sadly can't remember anymore) postponing modelling and focusing on planning the model out in photoshop.

this one maybe? v good set of tutorials.
http://forums.ego-farms.com/index.php?topic=81

[Snikske]

i agree totally on working with paintovers, if you can't get a grip, print it out, take a pencil and start drawing quads, it is so fast and it will help you controlling the mesh, and all the rest is just buttons and shortcuts.

[snouter]

I'm exactly where you are. Here's how I'm trying to handle it.

Learn to fail in a wise way, with an eye towards what you are trying to accomplish. Learn to analyze your failure for exactly why its a failure. Repetition for refinement and intuition.

Here is my first attempt to model my car.

The first model is an absolute mess. But from it I learned why I was trying to place polygons somewhere and maybe to stop trying to push too hard too early. Build a better foundation.

I stopped with that file and started again. I found that I was able to get the basics cranked out in a much shorter time, and overall my model stayed more organized. The techniques became a little more second nature, and I was fighting the software less, and thinking in 3d more. The second model is overall a litte more coherent.

I'm through with the second model. I'm about to try a third time. This time the problems I'm trying to solve are even further down the food chain. I expect to crack the basic model out even faster, and this time, I will apply everying I learned from the first two runs to advance even further. My car has some very soft curves and bulges and I have a plan of attack to deal with them. Getting control over the basic model will help me work towards that. I now see the whole thing as one giant mesh and I have a more intuitive understanding of how and why I need to do certain things with the mesh to get it to behave a certain way.

Now it's becoming a two step modeling proccess that is greatly aiding my understanding. I'm speed modeling the first part the more times I do it, which means those tricks and techniques are becoming fairly rote and well ingrained, and then I'm slowing down when I get to where I bogged down last time.

I also have somewhat simple goals.

The first model was just to even see how the techniques worked at all.

The second model applied everything I learned from the first model, and the overall result is clearly much better organized.

Starting again from scratch, with the third model, I aim to make the second model even more receptive to where I need to have it to start pulling off the curves and bulges. If I can show the same level of improvement between 3 and 2 as I did with 2 and 1, then, I'll about be in a position to make a real go for it.

For me the repetition equals refinement, and it really gets the skills ingrained. It's also less frustrating. I expect to bog down at some point, I just need it to be further along than last time.

[CGTalk Moderation]

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.