So, I don’t know why creating a new topic makes people want to answer my questions more as a guy named Kurtis said, but I’ll do it. If I have new questions I’ll edit them in.
When a reference image isn’t symmetrical (for example: a human head image, but with the nose NOT symmetrical to my view) it’s very hard for me to model over it. Since I’ve been watching tutorials where all the images they use are 100% symmetrical to their own view, modeling without one of those is difficult for me. Right now I’m modeling a jaguar (a big cat)
The images I’m using aren’t symmetrical, but I’m practicing modeling over the reference images.
It’s much harder than I expected, since the images aren’t symmetrical, you would have to rotate the model to match the image’s angle, but you can’t see all the vertices if you do that, thus making it hard. And, the side image is pretty useless for me, I have to rely 90% on the front view image.
So should I just forget about finding 100% symmetrical images and learn to model without them? it’s very hard for me right now to model without them but it may be because that I’m a newbie. The first model I made was with symmetrical images, and it was pretty easy even as a beginner who only knew the basics. I also watched a video of a someone modeling a human head without a reference image and it looked very realistic, so that brings me to the conclusion that I need to just practice more with any images even if they’re not perfectly symmetrical and forget about finding 100% symmetrical ones (obviously I don’t want to spend hours searching for some images that may not exist.)
So is it worth practicing without symmetrical images? would it really make my modeling easier and faster? or is the way I’m doing it right now fine? or should I learn to draw, which I don’t know how long it would take me, and I want to make my models as soon as I can anyway.
Why should I make my models out of quads? I heard someone say it’s “easier to control,” control what and why?
That’s exactly my thought almost every day at work Unfortunately, getting to know how the object looks is the most important part in modeling, so getting good reference is something that you definitely need and cannot skip. Even if you do not model from pictures matching front and side views, you need a lot of pictures or videos, diagrams, blueprints, drawings and other reference data just to see what the shapes are. You should find as much good reference data as possible. Try searching for high resolution photos as well so you can see detail clearly.
You do not have to. It’s just that it’s easier. In modeling you can insert, select and manipulate edge or face loops or rings easily also a lot of modeling operations behave differently depending on the geometry - for example if you have a lot of triangles with very sharp corners ending on the edge that needs to be beveled, the bevel operation will not work as expected. Subdivision modifier also might not be smooth if the geometry is not ‘tidy’. You will learn this naturally when modeling. You should not model in any way that is not comfortable for you - if you need triangles, use them, but be aware you need to think about the geometry and if they might get in your way later.
How about bring the image which you wanna modeling from it to Gimp or Photoshop. Cut half of face of something. And then copy half of it , from left to right or from right to left. Then You’ll a symmetrical image.
A lot of what you see others doing with Blender comes only with inherent artistic talent. I can’t for the life of me model anything but the simplest non-organic forms without good references to guide me.
I agree totally with MartinZ that you need as many reference images as you can find, and you need to study them frequently both before and during the modelling phase.
I bet if you asked anybody with the ‘inherent artistic talent’ about it you would find out there is no such thing. They all just put so much effort.
I have probably spent somewhere near 10 000 hours doing 3d modeling during my life as far and I feel comfortable with modeling pretty much any shape or object one can think of, but I will laugh if I am asked to model something without seeing it. That is absurd - one needs to know what shapes need to be modeled and the way to find out is to look with ones eyes. That is essential. Even if you are artistically creating something you are joining shapes that you have seen somewhere. You can get yourself very familiar with some shapes so you do not need to look at them any more and just remember stuff and combine different memories on the go. That’s how an artist who has modeled hundreds of faces before can model another one without looking at photos or any other form of reference. But it most definitely always starts with observing the subject. It doesn’t work any other way. Any artist will tell you the same story maybe in different words. Observation is one of the most important things. Almost every time I see a newbie failing and ask what reference photos they are using it turns out I could not do better myself if I only got what they use for reference. It’s the best advice I can share - get good reference photos, look at them. Doesn’t have to be photos, if you catch a jaguar and put it in a cage on you desk while modeling it’s fine too(apart from being illegal and cruel to the animal I mean (but hey, it’s all for art - who am I to judge?.. )).
MartinZ - I believe some people ARE inherently more artistically talented than others, even though skill has a lot to do with it as well. It’s the same with being smart or dumb. Some of it is inherent, some is learned.
I find that the sources that generally come up in google for reference images are pretty useless. You have to be creative in hunting down good sources.
Yeah… Google really will not be enough in some cases. It sometimes helps if you look for high resolution photos like 12 megapixels and more in image search, but the best case would be to be able to see the object and photograph it yourself… It is often really like hunting.
As for talent, I strongly believe there is no such thing. It’s only interests people have. If someone likes drawing for example, they draw, because they like to do it and then they learn to do it well because they do it a lot. I have never seen a person draw well from his first drawing. I have never seen amassing 3d work made by someone who started learning 3d a week ago. It’s an old argument. But all the ‘talented’ people say the same - they worked a lot. Some don’t say they worked hard, because they perceive it as easy and enjoyable, because they like it, but they all spend a lot of time doing the things they are ‘talented’ at. I am only mentioning this because I believe this ‘talent’ concept might be discouraging for someone starting as they often think they have no talent. It’s not true. They don’t have talent yet. 3d modeling is definitely something that does not require some inherent abilities. I know, because I don’t have any myself. The illusion that some people are just born better at this should not stop you or reptiliantroll from mastering it.
I believe some of us are born with an inherent flare for artistry. That’s all I’m saying. I’m NOT saying that some just don’t have it at all. But anyways, you’re right, MartinZ… I spend a lot of time with Blender, and although I’m still quite a noob and struggle with a lot of it, I’m better now than I was a year ago ONLY because I enjoy it enough to keep at it and learn how to improve my skills. I’m NOT a talented anything, but I was able to learn how to model and animate with Blender fairly well ONLY after putting in the time needed to learn and get better.
Sorry for the late reply I was away for a few weeks but I’m back.
So you guys’ advice would be to keep practicing with the reference photos i’ve got? By the way here are the ones I’m using:
Front view image
[ATTACH=CONFIG]502742[/ATTACH]
Side view image
[ATTACH=CONFIG]502743[/ATTACH]
I’m having lots of trouble making this model, it’s hard to model the side view correctly, I have to choose which side to make easier, the front or the side. For example in front view I model the eyes, but in side view they’re not alligned correctly, though I have to deal with that because of what I have in front view. Thus it makes it a pain in the butt to model the side. So my point is: If I practice and observe the shapes in these 2 photos, will I eventually model it almost perfectly? and after lots of practice I may not need to model over the reference images but only look at them. What do you guys suggest I should do? should I practice more and watch more tutorials?
There was no such advice given here by anyone. Where did that come from? I really think you should just read the comments again - you did not understand what I tried to communicate for sure.
That’s why I said “would” and didn’t say “is.” After reading every comment here i came to the conclusion (mostly from Anthony’s posts) that I should just keep practicing with my reference photos, should I do that?
If you can get several shots of the same subject with the same camera, you can make them into a sequence, and motion track it. The motion track points will give you a good spatial reference for building the model.
Otherwise, yeah, put the reference on the screen, position a reference camera in about the right zoom and location, make a duplicate of your geometry for each view, and then buckle down and do it!
You can get good results from even just one reference image:
The best advise I can give you is - if you struggle, get a lot more reference material. Observe more. If you understand shapes better, it is easier to model something successfully, if you do not, look closer. Two images of something that you see up close for the first time is never enough. Practicing something that does not work only makes you learn precisely that - what does not work.
I don’t know how to get several pictures of the same subject unless I catch a jaguar and put it in a cage, haha kidding. :eyebrowlift: Well if I do know how, then how do you put them into a sequence? what is motion tracking? what do you mean by the “reference camera?” sorry I’m new to blender.
" make a duplicate of your geometry for each view, and then buckle down and do it!" Wow, I think that’s the most intelligent advice I’ve heard here, genius idea. I think that fixed my problem of not being able to model both the front and the side view correctly, thank you!
It’s too complicated to explain here, but you can do this same kind of thing, only with still images instead of video:
It only works with multiple pictures taken in the same way, like video. Same sensor size, same zoom, same object. You could try it with multiple different subjects and zoom levels but… who knows what the results would be! An interesting idea, really. Maybe I’ll try it at some point.
Anyway, if you like the results, read up on “Blender motion tracking” and you’ll understand.
Excellent! You can also make several views, so you can model from different angles at the same time. Try Ctrl-Alt-Q and see how that feels.
Oh wait, if you duplicate your geometry for each view, both models would be flat from the opposite view. For example the front viewed from side would be flat. How would you merge these 2 objects together? should you merge them?
Wow I’m trying to make my model next to the reference instead of on it, and it’s incredibly hard to mimic the shape, but I guess it will become easier soon right? because “practice makes perfect.”
Yes, I wouldn’t worry about this being hard too much, this should improve soon. This definitely does get easier and more enjoyable with practice. This way you should be able to use more reference easier as looking at a shape does not require it to be straight photos from the side or the front. It will be very useful to train yourself to understand the shapes just by looking at them. Not that there is anything wrong with modeling over reference - that’s also fine.
dudecon's advise is really nice as well. Modeling from ‘tracked’ images is just wonderful. Tracking calculates the camera lens settings and camera’s positions so you can have reference ‘image planes’ match your model’s perspective exactly so every point matches exactly in every view - imagine that :). I think even though it might seem very complicated for a beginner it might be worth pursuing in addition to other ways of learning, because this gets rid of the struggle with reference completely and gives you opportunity to practice modeling only.
If these images don’t appear in the file I gave you, then how can I post them here so you guys can download them?
Please guide me, how can I improve the topology? how to manipulate vertices and edges easier to make the shape faster? what should I not do? anything that can help me with my model would be very good!
Unfortunately, getting to know how the object looks is the most important part in modeling, so getting good reference is something that you definitely need and cannot skip. Even if you do not model from pictures matching front and side views, you need a lot of pictures or videos, diagrams, blueprints, drawings and other reference data just to see what the shapes are. You should find as much good reference data as possible. Try searching for high resolution photos as well so you can see detail clearly.
What do you guys suggest I should do? should I practice more and watch more tutorials?
