Samir's Creative Ideas & Online Enterprise Blog - Page 38 of 40

There’s a bug in my system

While I was experimenting with the Fiber script, I came across the latest Weekend Challenge (#77) over at Elysiun. I am definitely not quick enough with Blender to try my hand at one of those yet, but #77 had an insect theme to it, and that lead me to think of modelling and insect to go with my newly found grass skills. A few Google Image Searches later, I decided to model some sort of moth.

… easier said than done. While I now knew perfectly well that I could model the wings as extruded bezier curves like I did during one of the tutorials in the manual (twisted metal, 2 posts ago), I wasn’t very clear on how to do the body and limbs. First I tried what I thought would be the easy way. I started with a UVSphere and tried to stretch and flatten it into the approximate shape of an insect torso. A little selective vertice editing in proportional mode, and I had a knob at one end that was supposed to be the head. To see how I was doing, I switched on Sub-Division to add detail, and the whole thing looked like a mess – a blob of clay with stretch marks all over the place. I gave up on that and decided to try again.

I had to try a more controlled approach, with fewer vertices to begin with, so that sub-division could add in the necessary detail. This time I started with mesh circle of 12 vertices and started extruding and sizing to create the basic shape, similar to the sword I modelled many posts ago. At the end I had a very bumpy spindle shaped object with a spherical knob on top. It looked nothing like an insect but it was a start. Now it was a matter of adjusting individual vertices to get rid of all regular geometric arrangements and make it into an organic shape. Many proportional edits later I was satisfied with my bumpy insect torso.

Now came the compound eyes and the limbs. The compound eyes I added in as a UVSphere scaled and rotated into the proper position, and then mirrored into a matching pair. The least complicated way to handle the limbs would have been to make them separate objects, but I decided to make them part of the same mesh to future proof this model for animation, with an armature setup. This meant that I had to extrude faces off the side of the body and model the various segments of the limbs from there. I switched off subdivision to work with the actual faces, and I realized that my model was too efficient. There were a few large faces, and I needed to create smaller ones to extrude from, without rearranging anything. What better excuse to try out Blenders new Knife and Face-loop cut tools? Since I had no idea what these two actually mean I just tried them out one at a time. I selected a quad face in the side view and tried out the knife tool first. This was very easy to understand once I tried it, but it resulted in some new triangular faces. From what I have read these are not supposed to be good when animating an object, so undo, and I tried the face loop cut tool.

Until I tried this out I was ignorant about what a ‘face loop’ was, but the excellent interface of this tool explains it well. Select this tool and move your cursor over the object and yellow lines blink on to show you where a cut can be made on the various face loops (basically a loop of adjacent faces that form a flat strip along the surface of a 3D object). With this is was easy to add detail to the areas I needed to extrude the limbs from. At this point I deleted one side of the model, so that I could mirror it after I had created the limbs for one side. The limbs were simply a lot of hard work with extrusion, scaling, and rotating, both in and out of proportional editing mode. Once the basic modelling was done I sculpted the legs into their natural position, and that’s what I have completed so far, as you can see in the image.

A lot remains to be done, including some antennae and the wings. Most importantly, once the modelling is done I am going to have to tackle UV textures to get the effect I want. From what I have read on the subject, I might find it easier to use an external UV unwrapper than starting with Blender’s built-in system.

Watch this space for updates.

High fiber diet

I can never stick to reading and working through manuals continuously for long periods of time. So, to continue my exploration of Blender I decided to try out something totally different. One of Blender’s greatest features is that it can be scripted with Python to add features that are not built in. The script that has always interested me the most is the Fiber script by RipSting, so I decided to try it out. Once I had installed Python 2.2 and downloaded the script files, it was only a matter of setting the Python path in Blender, and it worked like a charm.

The Fiber script is extremely powerful. The best way to get an idea of what it is capable off is to play around with the example blend file included in the archive. While the patch of grass example this file opens with by default is impressive, there is more to see on the other layers including some basic simulations of human hair. The long list of settings can be intimidating, but try generating fibres by changing one setting at a time, and you can quickly understand what all this means.

I started experimenting with the grass, and was trying to get it to look as true to life as possible. Playing around with the material settings helped a lot. I then found a close-up image of a blade of grass online and used that to create a greyscale bumpmap. Some more tweaking of the texture and bump map settings, and I was quite pleased with the results. The image above was produced after only 30 minutes of seriously playing around with settings, so a more in depth study of this script should produce wonders.

There are also some interesting alternatives to Fiber. Bantam 3D Grass is a free stand-alone program that can produce 3D models of grass and other similar structures. It allows you to design the geometry of individual fibres, so it can be much more powerful for more complex scenes. The models can be exported to standard formats to be used in a wide range of 3D software. For Blender, the Fiber script is just more convenient to play around with at the moment, since you don’t need to switch between programs and you can work in the same interface. Another option that is still in the works is the Beast script being worked on by ezual and Landis. This one looks very promising, but it is more likey to compliment the functionality of Fiber than replace it. My study of Fiber continues.
:satisfied:

Walking biscuit, twisted metal

My adventures with Blender continue. Only now I have switched over to the eternal work-in-progress Blender Documentation Project, which is being worked on in the same collaborative open-source spirit as the software itself. The reason I shifted over from the 2.0 manual is because of the very attractive first tutorial in this new document. It is called Your first animation in 30 minutes. In most manuals this would mean “How to get a cube to move across the screen in a straight line”, but once you get into Blender you must learn to be presented with the unexpected. This is a tutorial to model, texture, and animate a basic gingerbread man model … with armatures! Granted it took me more than a hour to do this but this is certainly not something you expect to get into so quickly in 3D software. Anyone who wants to quickly get the hang of this programme, and to see the possibilities of its animation system must work through this tutorial. Educational and fun.

After creating the basic walk cycle for the gingerbread man I continued with the documentation. Some sections had to skipped because they have been reused from the 2.0 manual. Finally I got to curves, and a tutorial to model a curves-based logo by tracing an image. I worked through this with an old icon I had created back in college. Impressive stuff. Especially the bezier curve tools in Blender are noteworthy. I found myself wishing that Adobe Illustrator and other 2D vector programmes had such easy to use point and curve manipulation logic. Once I extruded it into a 3D shape I didn’t want to stop there, so I pulled and twisted and distorted it into an abstract wonder, part of which you can see in the image. I am still not a master in good, solid lighting, but I’m getting there.

Show off!

How do I present my work? That is always a very complicated question, whether in hard copy or online. A portfolio can be arranged in an infinite number of ways, and an online portfolio can be arranged in an infinite number of ways plus one. Finally it comes down to weighing the pros and cons of the different formats, and just plain personal choice. But arriving at that decision is never easy.

On my old site, the portfolio page had links to subsections like ‘print design’ and ‘illustration’. Each subsection had B/W thumbnails for the work which linked to individual pages per project. The project pages had a short description of the work and one or more smallish, but visible, images of the work. Is this still a good structure? That’s what I need to decide.

When you look around the net, you come across certain standard formats for artist’s portfolios. Some have large pages of thumbnails which lead to large images with short descriptions, some have text links to projects which lead to small images with long text descriptions. And, you find everything between these two extremes.

The first decision to make appears to be between thumbnails and text links. The positive point of thumbnails is that they provide a visual cue to what is on the other side of the link. Unfortunately, thumbnails do not tell the viewer what to expect regarding content. Does it link to a CD cover, a magazine layout, an illustration? It’s impossible to guess. The strong point for text links is that they can be descriptive. For example, if someone is interested only in corporate identity related works, they do not need to browse through irrelevant pages before they reach an appropriate project. But, text links provide no visual cues.

The obvious solution is to use both, but again, HTML technology provides many methods to archive this. You could have text links/descriptions next to thumbnails. You could put in descriptive text into the “ALT” setting for the image so that the text shoes up as a ‘tool tip’ when the mouse pointer stays over it for a second. A third alternative with CSS is to have a text description appearing on the page when the mouse pointer hovers over a thumbnail, or a thumbnail appearing when the mouse hovers over a text link. At the moment I have not selected between these because I have yet to study which solution is easier to manage and smaller in code size, while still keeping in mind the aesthetic quality of the page.

As far as the project pages are concerned, right now I feel medium to large images with minimal text descriptions is the way to go. It might also be a good idea to keep multiple pages for projects with multiple images, rather than keeping a single large scrollable page with many images. This makes sense as it breaks up the bandwidth usage, and the viewer has a choice as to how much they wish to explore. I think text should be kept to the minimum as the work should speak for itself, with explanatory descriptions provided only when essential to the understanding of the work.

Amphora

This is the result of further explorations of the Blender 2.0 Manual. This scene involves the sword which was modelled using the extrude (E key) function as described in the Mesh Modeling chapter, and also the amphora which was created with a combination of the spin and spin duplication buttons as described in the same chapter.

This was meant to be a modelling exercise, and I was quite happy with the results. The remove doubles command is especially useful to create complex extruded shapes. I thought I might as well create a basic scene with these objects, which lead to a string of disastrous renders yesterday. It made me realize that I know absolutely nothing about lighting and texturing. I started today with the same scene but re-thought the lighting, after reading some of the lighting tutorials linked from ‘Blender – an Extended Introduction’. It is definitely an improvement over some of my previous renders, but I still have a long way to go in lighting and texturing.

Right now it is best to stick to the Blender Manual, and get to know the intricacies of the software. Once that is done I can study the more general theories about lighting and materials.

I have no idea what is causing that un-natural ring of light at the base of the amphora, but I am sure I will find out as I learn more about Blender.
:blush:

Splashdown

In spite of all that brainwashing about not judging books by their covers, every one of us goes by first impressions at some level. And why not? First impressions are very important, especially on the web. There are two kinds of people on the net today. There are those with broadband connections, who will see the first page of your site quickly and dismiss you immediately if it is not satisfactory (this is assuming your first page is not a 1MB monster – even broadband has limits). And, there are those who still dial in with modems, 56K at best. These people will need to wait a little just to see your main page. After all that waiting, if the impression you create is not a positive one, why would they spend any more time looking through your site?

Main pages (index.html for us HTML geeks) come in many forms and colours. There is the Yahoo and Amazon school of “lets put every thing on the first page” design. This makes sense for large complex sites, which is why it is almost the standard format for web portals. I’m simply going to ‘skip’ the Flash extravaganzas, because I don’t even consider them an option. With personal sites, front pages come in two basic flavours: with splash page and without splash page.

Pages without a ‘splash screen’ either open with an introduction to the site’s content or with news/what’s new page. I’m not a big fan of the text intro page because it gets dated very quickly. This blog is going to be the news page for the new site, so starting with this page might not be the best idea. Firstly, it is not representative of the content of the rest of the site, and more important is the fact that most viewers might never get to the rest of the site.

The splash page adorned web site also comes in two varieties. First, there is the pretty picture splash page which requires you to click on the image to continue, or sometimes even scroll down and click on an insignificant link to continue – some do not even indicate what needs to be done to continue. Anyway the second and better type of splash page has visual elements to impress and also menu links into the sections of the site. This way you get to have the best of both worlds. On one hand the visitor is not immediately thrown into long text pages, and at the same time they are given an idea of the sites content and sections – something more than “click here to continue”.

The last option is what I have used in my old site, and I plan to use the same in the new one. This time the splash page will be made more intuitive by eliminating a general ‘portfolio’ link and replacing it with the links to the subcategories, as I have mentioned in an earlier post. Overall, this is the best way to go.

Ivory symphony

At long last I have started reading through the Blender 2.0 Manual. I got to the Mesh Modeling chapter and worked through it. Learnt a lot of new information I wasn’t aware of before – resizing an object disproportionately being one (click on the middle mouse button when resizing). Some of these editing features are actually a little more intuitive now in Blender 2.3, and they don’t require as much careful dragging of the mouse.

When I finally got to the part about proportional editing, I simply couldn’t move on. This is another part of Blender I had absolutely no knowledge of before today, and it is a joy to play with. The organic forms you can create with this technique are astounding. The image you see here started off as a plane mesh which I distorted in various ways in proportional editing mode. Since I liked it so much I skipped to the part about subdivision to add some more detail. Even better! 😀

To add a finishing touch, I simply had to render this with radiosity. Blender 2.29 onwards has a new quick radiosity render, which doesn’t require the complex multiple steps to calculate a radiosity solution before the render. It took me a long time to get that working but I finally managed. This is still very new, and detailed tutorials and instructions are not to be found anywhere on the net at present. But, this should be addressed in the upcoming Blender 2.30 guide by the Blender Foundation. Until then, enjoy the image.