Suggestions on Sub-division modeling approach?

Hey all so I was hoping somebody can comment on their own experience, or give me some advice on subdivision modeling. I’m working on modeling some quadrupeds. I’m thinking I really need to change my approach. What I’m doing now works okay. Eventually I’m able to achieve the shapes I want, but it takes time.

Basically my approach is to first gather reference images. Specifically the silhouette of the animal. A side view. I insert the image into the background of my viewport in my modeling program (Maya in this case).
Then I will insert shapes and roughly block them out based on the major parts of the body in the reference image, like the torso, neck, legs and head. I usually use spheres for this. Then I will proceed to combine the shapes (based on ref image) by deleting faces where necessary (where the morphed spheres meet), and snap and merge vertexes together until all of the shapes are one cohesive model.

This approach has worked for me, but it takes a lot of time, because there is a lot of manual pushing and pulling of faces, edges and vertexes. After finally getting one cohesive model, and a good base mesh to work with, I’ll go in and start adding edge loops to increase the poly count and improve the topology. Especially around the areas that were merged together.

After I have a cohesive model that matches the silhouette in my reference image, I’ll go and start adding the details like the eyes, nose, mouth, feet etc.

Like I said, this approach works, but I feel like I could be using a better method. Something that is less time consuming, and something that introduces less topology issues.

I find with hard surface modeling, you can usually stick to just extruding faces and edges to achieve the shapes you want. However, organic modeling seems to need a different approach. I’m not sure I have found the best one yet.

What do you guys think, how could I be doing this better?

Do you have some screenshots of something you’ve modeled this way?
That’s not the way you’d usually want to go about doing it. With a typical sub-d workflow you start with a simple shape like a box and then press 3 to go to smoothing preview mode which will subdivide the mesh–a cube will become close to the shape of a sphere by doing this. You would then edit the shape to be a bit like the body and then add loops and extrude polygons out to create the neck/head and other limbs, it’s just basic modeling. But by using the smoothing it allows you to work with few vertices which control the shape of the smoothing.

I don’t have screenshots at the moment (I’m at work). I can put some up later though. You know, I think I need to relearn modeling. I have been doing things the hard way. Tonight I’m going to try modeling a dog, and I’ll try doing it with a square. It didn’t occur to me that I could simply smooth the model to get the roundish shape I was looking for. What I have trouble with is getting from the main torso to the limbs by simply extruding faces. So I always tried modeling them separately and attaching them later. I actually saw somebody model a deer like that in a tutorial and it seemed to work out for them. So I just started modeling a lot of things that way. A bad habit. I knew I wasn’t doing it right! Thanks for the reply.

Pressing the 3 key in maya will be “smooth preview mode.” So its only how it will look, and it adjusts polycount. So you can actually have a 500 poly model that displays you future smoothed mesh of say 3000 poly. Press 1 to return to normal view. As you move verts and polys around you can change the way the mesh smooths in its preview. So if i make a box and press 3 it will turn into oblong sphere, but it still only has 4 sides. Its makes things so much easier. Its a good idea to go to mesh display and soften edges also. Then you will get an excellent idea of exactly how it will look smoothed. If doing character modelling I use those features a lot. Its better than ctrl z. Its not actually smoothed until you press 1 and then go into the toolbar and smooth.

Yea I’m going to go through some tutorials this weekend to learn more. It’s somewhat annoying because I like to spend more time developing than going through tutorials. However I’m sure the time investment will pay off. Thanks for the suggestions!

Hey thanks for those simple suggestions. I watched some tutorials this weekend and it helped immensely.