A dinosaur thing WIP. Critique welcome.

Hello Blender Artists community. First post!

For the past unknown amount of years all I’ve known about blender is how to extrude horrible looking shapes out of a cube. About a month ago I had enough of wallowing about in my ignorance for a subject that I was interested in. So here I am.

Here I’ve modelled and textured a dinosaur of my own design. Which still needs a lot of work. Mostly on texturing I believe.




I don’t even know where to begin with the eyes.

Any constructive criticism is welcome and I’ll post any changes that are done to it.

I had a problem with the fingers and toes. After I had modelled them separate from the rest of the mesh and then parented them to the main mesh, they always stayed a different shade of grey. Textures also would not apply properly.
After some searching it seemed that the cause was that the faces on that part of the mesh were inside out. The solution was to recalculate the normals.
I can’t say I fully understand it.


After fixing that problem I decided that the normal map, which was made from the image texture with a Gimp plug-in, needed work. The skin seemed raised above the scales more often than I would have liked.



I opened the image in Gimp. Then I de-saturated the colours in the image and then saved over the top of the original image like an idiot.


Now my dinosaur is in black and white.

When I said that this was a ‘work in progress’ I was talking about myself.

I have returned. To fix old mistakes and to make new ones. All gain involves pain.

One part that I was not happy with (and still not happy with) was the mouth. The texture on the inside of the mouth looked like some kind of dried fruit, a sultana perhaps; also known as a small seedless raisin, a grape which has been long forgotten after being stomped out of the wine press and left to dehydrate in the harsh unforgiving sun. Whoever discovered the raisin probably forgot to harvest his grapes. But I digress.
The mouth looked like this:



And so I had to tone it down a bit. Now it looks more like a dried fig.

I also worked on the skin because during moment of pure stupidity it had become monochromatic (see previous post). The image below is the node setup for the skin.



This is the largest collection of nodes I’ve done so far. Some of them probably make little difference. More than half of the nodes are there to add some roughness to the scales, which is a collection of noise textures and maths nodes which are then fed into a MixRGB which has fac input from ColorRamp, the ColorRamp gets Fac input from the image texture. The image texture is the monochromatic image I messed up earlier.


I also used a similar node setup on the claws. Which were just white before.


One thing I have not figured out yet is how to make a seamless transition from the material inside the mouth to the skin material on the outside of the mouth. Currently there is a line of demarcation between the two different materials.




I think you should use more reference. It doesn’t have to be one particular species, it can be your design, but for the different body parts and details you should take a look how body parts of real dinosaurs or reptilia look and how they are connected.

For me it looks unbalanced, the front part is much heavier than the tail, it would fall over. The tail should be much bigger to balance the heavy front.

Also I think you went too early into texture detail. It could work as a cartoon character, but for a realistic dinosaur, it should have more detail like eyebrows sculpted, nostrils, maybe some big skin folds or scales, leg muscles, etc. You did this kind of detail for fingernails and teeth (their attachment to the body is good, not just stuck into the body, but you sculpted fitting cavities in the body), you should go this way for other body parts too.

The attachment of the arms to the body looks unnatural, like separate clay parts stuck one onto another, not like connected with bones and muscles.

The foot has an almost human shape, especially the heel, which is not proper for this kind of animal. Note that the heel is in the air, it’s the “backward knee” in the hind leg, while the counterpart to the human knee is somewhere around his belly. It’s the same for many mammals (eg. dogs, cats, horses), all birds, and this kind of dinosaur - the hind leg bones “shifted” in their function compared to the human legs. Like a human standing on his toes. Look at skeletons.

Your texture looks the same everywhere, there’s no variation. Look at a krokodile, or at a naked hen, or a turtle. It has different skin on different parts of the body. Sometimes scales, sometimes skin folds, sometimes just some skin pores. Different relief, different colour.

Anyway for a first attempt at character design and modelling, it’s not bad. If at this stage you don’t want to add the details I wrote about, I’d suggest to go more into a cartoon/non-realistic direction with the shaders and textures.

The faces have two sides. Imagine cloth fabric which has print just on one side. Wrong normals mean that the print is turned inwards. Recalculating normals means flipping faces so that all the prints are outside, kind of. It’s important if you use normal maps, displacement or bump - the maps must know which direction they should bump out. Also for transparent or translucent shading, it helps Blender to know which side is inside the dinosaur and which part is outside. And important for viewport shading, as you see in your screenshots.

Thank you ania for taking the time to write your advice, and that excellent explanation for recalculating the normals.

After some internal struggle I decided to go back to sculpting and modelling the thing. After you made the anatomical errors obvious I realised that I had based the entire design off a quick 3 minute sketch I had drawn without looking at any references. Then for the next week I spent trying to model that quick outburst of artistic laziness.

I’ve increased the length of the tail to bring the center of gravity towards the legs. Even though it has knuckle dragging long arms it’s not meant to be a quadruped.


I have also changed the feet a bit. Sculpting them with dyntopo which I have only recently discovered.


The mental processes behind the new foot design was that birds are related to dinosaurs and so a demented two-toed chicken foot seemed to be the most logical thing to do.

That’s all I’ve done so far.

A little update on the modelling



A bit more dyntopo modelling on the body.
Moving on to the texturing now.

I will suggest turning on backface culling, it helps to differentiate which side is the normal facing.



Thank you xiankang. I’m already finding that useful.
Blender has so many buttons. I wonder if I’ll ever know what they all do.


I’ve now done many other things since my last post.
I’ve re-textured the body, made eyes and rigged the thing. I’ve bought a graphics card ( GTX 1070 ).
Rigging this model I found difficult because automatic weights didn’t work for some reason. Could this be because of using dynamic topology while sculpting and then not doing any re-topology? All I did was select the entire mesh in edit mode and then I invoked ‘tris to quads’, because I read that quads are best for rigging.
I think I need to learn something about re-topology, I sure there are better uses for my processor time than rigging something that has 621280 faces.
Anyway for this little educational project I did many things for the first time, and then much to my disappointment/relief, found a better and faster way to do it later. But now I think it is time to move on from this model because I feel I’m starting to stagnate on it; the amount I’m learning from it is declining. I would like to take what I have learnt from it and apply it to a clean slate.

I also started experimenting with filmic-blender. So my final renders for this model will be using filmic.


Strange pose. Prehistoric chicken with long arms.