Tuesday, November 11, 2014

Blender Quick Tip: Sculpt Mode for FAST Topology Fix

Ok, so maybe it seems like I was only paying attention to ZBrush & forgot all about Blender. But really that's almost impossible since I make most of my base meshes in Blender before taking them to ZBrush. ZBrush is picky about topology though. It likes quads only, regularly spaced and more or less square please. I'm ok with that and normally I at least try to get my edge flow all nice and clean before I leave Blender. 

Blender has improved sooo much over the last couple upgrades! And today I figured out a fast topology fix that means I will never again have a reason not to have near perfect topology on a base mesh. Plus, it's so good and so smart, I thought I should share it. 

Use sculpt mode to clean topology... before sculpting! (seriously)

  • It's only a Tab key away with the new radial menu pop up. 
  • Wire frame works in sculpt mode now (so does mirror modifier btw)
  • Grab Brush gives far greater control than proportional edit ever could.
  • You can adjust size, strength, fall off, switch to smoothing all with easy hotkeys or on the tool shelf.

Even if you never sculpt, try this as a replacement for proportional edit. It's just as fast to get there now and with so much more control over the fall off, you'll get what you want quicker. Plus, proportional edit doesn't let you hold shift to switch to a built in smoothing brush now does it? 

Sunday, August 24, 2014

Asymmetrical Symmetry? Really

So there is this thing about ZBrush and me. I call it an idiot savant program because really it can do things that just have to be seen to be believed. That's the savant part. The idiot part is that you frequently have to put these operations together piece meal from a variety of menus, or find the magic button misnamed and hiding somewhere you would never expect. 

I found a great one today, which does something I would never have thought possible. To use it, you must enable symmetry on your model, which is found on the Transform menu and then turn on the Use Posable Symmetry button right under that. Now ZBrush will let you work on the model symmetrically basing its calculations on the topology of the model rather than an axis. What does that mean? That means that even if your model has been moved off center, or posed (!), ZBrush will still let you make changes to 'matched' parts simultaneously as though they were still symmetrical. It's really amaZing. 

Lookie here:

Monday, August 11, 2014

My Specs for Berry's Meme

Well, let's face it, I was overdue for a blog entry, and computer specs are integral to everything I do here and elsewhere - how I run the software I use to make stuff and Second Life where I make believe the stuff I make is real stuff. 

First off gotta say that although I don't regularly do Strawberry Singh's memes, I often think that when they are fact finding missions, they perform a valuable public service. This one I decided I'd follow through with so that there would be at least one person on a pretty good but by no means great computer who can nevertheless report great performance and experiences in SL.

My Facts & Specs Follow

1. Share any of your computer specs (video card, memory, etc..) 

CPU: Intel(R) Core(TM) i7-3770 CPU @ 3.40GHz (3410.06 MHz)
Memory: 16333 MB
OS Version: Microsoft Windows 7 64-bit Service Pack 1 (Build 7601)
Graphics Card Vendor: NVIDIA Corporation

Graphics Card: GeForce GTX 650/PCIe/SSE2

^ It's about a year old, it was also a custom build. I was pretty strapped for cash, so I kept the graphics card from my then burned to the ground computer, but I made sure that I put in a big old power supply in case I get rich suddenly and can go get top top toppest nvidia card. 

The extra RAM was the biggest improvement factor for me. I only had 2 gigs before this upgrade, and like you could rarely run more than SL and maybe a couple web pages at a time. Now I almost always have Photoshop and Blender open while I am in Second Life with no problems.

Other stuffs: We have some in common - Windows 7, Photoshop CC, Wacom Intuos also the small one, yes, yes and yes!
Blender 2.71, ZBrush4R6 are frequently tag teamed with SL and I use 2 browsers Opera as main and Chrome for some things, both of which usually have multiple tabs open. 

My monitor is at least 6 years old now and has survived 2 different computers and a few random system upgrades. it's an HP w2408 and I have never seen a monitor I thought looked better. It's wide enough to have 2 programs open side by side, meaning that so far I have not felt the need to have an extra monitor... but some day.

2. Which viewer do you use most often? 

The default LL viewer is the only one I ever use per se, but I do take other viewers out for test drives to see whether things look different and if so in what way. Also, students and help seekers are frequently on other viewers, so I like to have the answers to where the buttons or menus are for them. And about once every year, I download and try every single not V1 style viewer just for fun.

3. What is your FPS when you have your graphics on ultra? 

I never actually use the slider, instead I select the stuff I want high or low individually. For instance, I might keep draw distance low, rarely turn shadows on, but would not last a second without Advanced Lighting. 

Anyway, using the slider and the Ultra setting, my FPS was 30 while running through the woods, and a bit higher when I stopped. Maybe that doesn't sound so high, but experience-wise I felt no lag whatsoever. 

With my usual Preference settings that look like this:

I can run through the woods or rip up The Cornfield (speaking of experience!) and my FPS is in the 50-70 range. So I think the shadows are actually what makes the difference. But as I said, the big difference in FPS number does not translate to any real noticeable difference in felt performance or the experience of lag for me. I hope people are not letting the quantifiable overshadow the qualifiable by convincing themselves that a high FPS means they are not lagged and a drop in FPS means they are lagged, or that place is laggy or whatever.

4. How often does Second Life crash for you? Is it usually just a viewer crash or your whole system crashes? What are you usually doing at the moment of the crash? 

Really I almost never crash. But the times I do, it is often because I have too many chat windows open in a crowd of people. So even though I almost never crash, I do sometimes if teaching a largish class.

I also sometimes crash when I am all alone on my platform because I have too many chat windows open and in particular if I have too many text fields open and waiting for entries. For example I will be yapping away in 2-3 groups, 1-3 private IMs, typing in search in inventory and editing/renaming an object, when blammo  I get disconnected. Really, on those occasions, I think the viewer is trying to do me a favor. 

I have only crashed to blue screen a couple times. It was in the past and while on the Firestorm viewer. I think it had to do with the fact that my graphics card was not well served by having vertex buffer objects on. At any rate, it scared me, and worse, once it left me with a video card showing as uninstalled in the system hardware list. 

5. Do you know of any tips or tricks in the settings that would improve performance?

Well, much depends on which part(s) of the performance matters to the person. Draw distance down is my first line of defense and after that if I have a lag spike or whatever I turn off avatars (Ctrl Alt Shift 4.. or corner 4 as I call it). No avatars to render equals huge reduction in lag. But I know a lot of people think avatars, their own included are the most important part of the scene, so maybe that's not the best cure for them. 

I have been noticing more and more lately that keeping my chat windows down to a dull roar also helps with textures staying in focus, which have been a bit of a plague for me for a few months.

I'm not sure if inventory size affects performance or crashing, though I have long suspected it might. The people I find I say WB to the most often seem to have mega inventories 100K-ish in common. Me I keep a quite modest 20K inventory and I have everything sorted inside system folders generally, which someone told me made losing stuff less likely. If we could rename system folders, and I wish we could, my "Photo Album" could be renamed My Filing System.  

Ok, that's enough about that. Thanks for the inspiration Berry. 

Tuesday, April 22, 2014

Blender Tip Smooth Operator

If you want, start humming along with Smooth Operator since apparently setting something to music helps you remember stuff.

Here is a very nice time saving Blender trick I decided to share as a mini one image tutorial. I use the Smooth operator on the Specials menu quite a bit to average out the size of my polys before sculpting, but it was also the perfect trick here to avoid sculpting.  

This shows my somewhat patchy patching job after joining the top and bottom half of a body on an imported obj. I am also working on only half the body to save time, then planning to apply a mirror modifier. I could head into sculpt mode to smooth this out with a brush, but my intended use of a mirror mod means this is not quite as straightforward as it sounds. I would need to apply the mirror mod first, then divide the mesh again if I had further work planned. I could also slowly, using proportional edit maybe, move the verts around until this area made a nice smooth transition. puhleeze.

Instead, use C for circle select to quickly grab some faces (smooth works on edges and verts too, but I find Face Selection mode makes it easiest to see what I'm doing on this). Then press W to bring up Blender's Specials menu, choose Smooth. Now the super timer saver step is press the F6 key to get the Last Operator pop up where you can control the amount of smoothing (number of times that smooth is applied) using the slider. Bonus, in 2.7 you can also choose which axis to apply smoothing to.

Blender gets better and faster all the time.   

Thursday, April 17, 2014

Ebbe Linden & Education In Second Life

Ok, so this is a bit off the topic of content creation, but certainly fits in with the general Some Of My Best Friends Are Pixels thang I got goin', well the pixels part anyway.

So as a frequent reader & long time fan of Strawberry Singh's blog I sometimes almost do one of her infamous and usually super clever Blogger Meme Challenges, but then I don't. This time, I am bustin' out, and taking Flat Ebbe on a tour since it's just too good an opportunity to miss.

That's Flat Ebbe & me at Builders Brewery, a school, help group, sandbox and all round educational nexus in Second Life. I knew that Ebbe would like it here since as I understand it from his public statements so far he is pretty keen to reestablish Second Life as a viable forum and platform for education. 
Me too! 

So in keeping with the meme theme, here are the three things I would like to tell Ebbe:

1. I was genuinely thrilled to hear that education in SL was on someone's agenda again. I hope that the agenda is broad enough to include support for in world educators and learners as well as attracting brick and mortar institutions and students from the outside world. 

Now two of my current pet peeves - both related to content creation and specifically to streaming costs and how Land Impact is calculated. Although I realize this is probably way below Ebbe's pay grade, I bet he knows someone who knows how to fix this kind of thing.  

2. First, the Land Impact algorithm for mesh objects could be improved. It doesn't really give a fair calculation of streaming cost based on volume, because as soon as an object pokes outside of a hypothetical cube, the LI goes up as though the object in question is now occupying a proportionally greater chunk of screen real estate. But in the case of a long skinny object for instance, this is a misreading. Specifically, there is no way that a 500 tri tree trunk should cost more in Land Impact than a 5000 tri tree stump, no matter how tall the trunk in question is.

2b. Second, I guess it probably saved time to make all materials and Alpha Modes invoke a streaming cost calculation, but Alpha Masking, which is super fast and potentially a great lag reducer should NOT raise the LI no matter what surface it is used on. Specifically, as it stands now, changing the Alpha Blend on any sculptie plant (the vast majority of plants in SL are sculpties) would significantly reduce lag and render times in those areas, yet due to streaming costs calculation, would have the effect of doubling the prim allowance required. 

3. Love that you're engaged and engaging, maybe restart the blog as a semi regular what's new. Lot's of tiny stuff changes and often for the better, but goes unnoticed.     

Wednesday, April 16, 2014

Blender Limited Dissolve ~ A love Letter

I confess I feel a bit like I am revealing a dirty secret, but it's such a recent and delicious discovery for me that I wrote a love letter (illustrated of course) to a Blender operator and I'm going to show you my solution to a workflow problem that has bugged me all week.

I've been sculpting and painting these cute little figures in ZBrush but then taking them back to Blender to get ready for SL. Now I don't know if you know this, but ZBrush tends to encourage some pretty high poly work, certainly way beyond what Second Life is prepared to allow.

He's cute, right, but he and his little friends have given me nothing but grief at the finishing up for Second Life stage. I tried exporting subdiv 1 from ZBrush, about 4-5k polys - reasonable for SL, but barely enough to wear their normal maps which are supposed to preserve details such as facial features. Then there were the textures to worry about. I polypainted them in ZBrush, which being ZBrush uses an equation something like 1 texture pixel per poly, so since a 1024 has more than a million pixels, you better have more than a million polys while you paint. But woe to anyone who thinks she can drop that same texture on a subdiv 1 and still expect to see any detail let alone pretty. Subdiv 2 wasn't much better, and now I was spending a significant chunk of time trying to reduce the face count before upload. If he was jewelery or a worn item, of course, SL being SL, I could get away with as much as the uploader would allow... cough. This would not be the responsible thing to do, but desperate times call for desperate measures.

And I think I found the ideal desperate measure - Limited Dissolve in Blender. I am head over heels in love with it. 

First let's be clear - this is not exactly sound retopology practice. I'm not even sure some of what I ended up with deserves the name topology. The only way you might end up wanting to use this over say, removing edge loops or doing retopo was if you happened to have some very high poly models that already have polypainted textures and UV maps, meaning it's a bit late for retopo, and a bit dense to be bothering with removing edge loops, and too soon to decimate.

Now I know what you're thinking, maybe that whole work flow is flawed, maybe I should give up polypainting textures in ZBrush and go back to texture in Blender or Photoshop after retopo but, but, but sob, I just can't.

So if you can't keep from sculpting and polypainting but you still want to strip away enough geometry to make a viable SL trinket, here's my dirty secret love letter to Limited Dissolve.

Ok, if a one button reduction from 171 faces to 5 isn't enough to make you all tingly, I am truly surprised you read this far. If you are feeling pretty excited try to hang on a bit longer so you'll know some of what to expect on the honeymoon.

So far I have no way to predict which areas will work first try, or what sort of angle will produce the desired results. More careful study might answer those questions, but it all happens so fast and it's so easy to tweak, that the only dilemma is what you will say when someone points and laughs at your huge randomly placed ngons.


Now you know my dirty secret love of Limited Dissolve.
Sign me 
Desperately Seeking Less Geometry,

garvie garzo

Thursday, April 10, 2014

Blender First Steps Tutorial

This is a class I wrote for Builders Brewery School in Second Life last summer. It's been popular and it's been tons of fun to teach. I am putting it and the slides on line now for easier access for me, for current students, and for anyone who for whatever reason was unable to attend the inworld class. I hope it's useful for beginner Blender users.

The point of the class was not to teach Blender per se, but just to provide enough familiarity so that students could start box modeling as soon as possible. So many people I talked to while getting ready to do this class told me that they were planning to learn this program or that program so that they could start making mesh for Second Life. I think that's a pretty bad approach myself. Better to just start using a program and learn to do by doing. Anyway, that's what I told the BB students in SL, and these are the slides I showed so that they could get to know their new tool as a tool for beginner modeling.

Allow me to introduce Blender's opening 3D scene (everything inside the orange dotted line). It's not that exciting really, the infamous default cube, and some black stick outline thingies. 
Two things worth noting here: 
First, T is a Typically brilliant Blender shortcut, using the first initial of the item or operator it is a command for. In this case typing t will open the Tool shelf. Pressing T again will close the tool shelf. A single easy to remember letter to toggle open or closed. Blender also usually underlines the relevant shortcut letter for each operation on menus as I have done on the slide.
Second, That tiny icon indicating that the main window is set to the 3D viewport conceals a bunch of other icons for other window/editor options. In Blender you can change the contents of any editor window to any other kind of editor. So you can rebuild the UI to suit your style, your eyesight or your current project.   

The Outliner window looks and sounds a bit technical, but it's really very useful. It will always list everything in your scene even when the object is hidden from view or on a different layer. Right now it lists the 3 things we have in our scene and it highlights the word Cube because the cube is currently selected.

Our first mission is to select the camera and the lamp so that we can move them out of the way. This is a frustrating moment for new Blender users, because a left click seems to do nothing but move that irritating little target around.

In Blender, you right click to select.

To select the camera, right click on it, to add to that selection, hold shift and right click the lamp.

You can also select things from the list in the Outliner window. In this case, a regular left click will highlight them, and holding shift will add to the selection.

And you can select things by press a to select All. A is another toggle style shortcut, A to select All, and a again to deselect All. That last is very handy when modeling, to make sure you don't accidentally have something selected that shouldn't be.

But we don't want all selected, we only need to move the camera and the lamp, so select both the camera and the lamp, by right clicking and holding shift.

Then type m to Move selected objects to a different layer. A pop up will appear wherever your mouse is, to select a different layer, merely click on one of the small grey boxes in the pop up.
Note that layers in Blender are really more like storage lockers or folders and not like layers in graphics.

The 3D View Header is a really useful collection of icons, operations and settings - a sort of command central for Blender operations and manipulations.

First you should notice that we are in Object Mode. Object mode allows us to deal with objects as objects, but not to edit their parts individually. We can move or scale or delete this whole cube, but we cannot do anything to just one of its faces or just one vertex for instance.

Next to it is an icon that shows what 'shading view' of our object we are looking at. The default starter view is solid. Have a look at some of the others, but for now plan to stay in solid view

You will find wireframe very useful in modeling, but since the shortcut for wireframe is so easy (z), I tend to ignore the icon alone and just press z when I want a wireframe view.  

Next to that is the Pivot Point. We will discuss and use and abuse the Pivot Point quite a bit in later project classes. For now just leave it as is.

Next to that is the control to display or turn off the 3D manipulator widget. I keep mine off mostly. You don't actually need it to move things around in Blender, the way you do in SL but if you find it makes things easier, then leave it on.

Under that is a 2-part Tooltip. Like many programs, if you hover over an icon, blender will pop up a little tool tip window telling you what that thing does. As a bonus, it also gives you the relevant commands to use in the Python scripting language in case you want to learn that and possibly become a Blender developer. Personally I mostly find the python stuff almost frightening, so I will show you where in preferences to turn that off if you want to. 

Look for preferences under File menu > User Preferences.

I suggest you think of Blender preferences as settings to try now and again for different projects. You might want to turn the number of undo steps down,
especially if you are on an older computer or a bit short on RAM. You definitely should allow yourself several hours to explore the possibilities in the Add Ons tab. Above are 2 optional settings, the first to turn off the Python scripting tips in tool tips, the second to turn down the number of undo steps.

I think that slide speaks for itself, though students regularly seem to balk at being told they must do something. Fear not, nothing will be overridden by having these on, you won't notice that your mouse or Numpad behaves as though you are not giving them credit or something.

Most people only ever see the default grey factory settings for Blender, since people who make tutorials seem to think you will want your interface to match the one in the tutorial exactly.

In case you want to match mine exactly for later classes, or maybe think grey on grey on grey is a bit depressing, check out some other Blender Theme presets.

mmm eye candy

To uncover all the color options shown, under the Themes tab, first choose a theme, then choose an editor or window from the list you would like to make adjustments for. In the slide I have selected the Ubuntu Ambience theme, then decided to see how things look in the 3D view. You can see that 3d view is selected because it is highlighted in purple.

The main thing I'll be worrying about is the background color of the 3d window. I think black is too dark and too reflective to stare at for hours at a time, and I like to have enough contrast with the floor grid and backing to be useful.
Figuring out which setting affects the bg took some poking, so I have circled it in orange in the image to the left. Look for "Gradient High/Off"

Another thing that really counts for seeing clearly in Blender is making sure that the text can be seen clearly. You will find this setting on the System tab, on a button called DPI:72 (what they were thinking on that one is beyond me). Anyway, press on the button and you can slide the text larger by moving your hand to the right, or smaller by going left. Much better! 

Now Save your new User Settings so that when you start Blender you will get an interface that feels a bit more like home. You can keep improving and changing the look of Blender as you spend more time in there. 

Some things to think about are making face dots or vertices larger, improving contrast between selected and unselected elements, making sure you can quickly and easily see whether a button is on  or off. As you make changes to the UI or your preferences, save them each time here.

Time to go back to the 3D view and see what our cube is up to.
Remember the 3 ways to select things (there are others, but this is plenty for now):

Right click, select from the list in the Outliner window and A to select All.

Select the cube. 

If you have already watched some Blender Tutorials, you will have heard the phrase,"Tab into edit mode" that's because pressing the Tab key will switch between the 2 main modes (usually Object & Edit), 

Things look different in Edit Mode. 

Also some new buttons have appeared on the header bar, because we can do things in Edit mode that we couldn't do in Object Mode.

We'll use these icons for a kind of crash course on the fundamental structure of mesh. Mesh is defined or created first of all by points or Vertices in 3D space, Edges run between 2 points, and Faces form the surfaces in between.

Blender like most 3D editing programs allow you to edit mesh at any of these 3 levels. A simpler way to put that is we can do stuff to edges, to verts and to faces. 

Right now according to the icons, I am in Vertex selection mode. But since every vertex of the cube is selected, the whole cube appears orange.

Type a to deselect All, then right click on a corner to select just one vertex.

Now you can pull just that vertex around and your cube will change shape as the edges and faces connected to that vertex also get dragged along.

If like me, you have turned your manipulator widget off, you might be wondering how to move this vertex or anything around. 

In Blender to move something you press G for Grab, then just move your hand/mouse. 

Now right click to select all 4 vertices on whichever side is facing you. 

Scale (to resize) is S.

Scale can seem a bit tricky. If you start with your hand back from the element your are resizing, you will find it easier to control. You can pull all the way back until the indicator arrows go off the screen and they will come in from the other direction. That is the continuous grab preference in action.

Since I like to point out Blender brilliance and especially its logic when and where I can, I usually encourage people to guess what the keyboard command to Rotate a face is. Since Grab is g, and Scale is s, in a perfect world, Rotate should be r, and it is! Happy dance.

Press A to deselect All. Now let's look at face selection mode.

You can use the buttons on the bottom Header to change selection modes, but you don't have to. You can just press Ctrl Tab to bring up a pop up menu to change Mesh Select Mode

Menus like this pop up wherever your mouse is when you call them, which makes for a very fast workflow.  

In face selection mode the view of our mesh changes slightly. Now we see little square dots on the faces of our mesh. In edge selection mode you won't see anything unless you right click to select an edge or two.

So you can manipulate mesh at the vertex, face and edge levels. Most basic operations work at each of these levels. In most projects in Blender you should plan to use different selection modes based different operations you are doing to your model. But for now, just get comfortable hitting Ctrl Tab to quickly change selection modes.

As important as learning how to make edits to things like cubes, is learning how to navigate around inside the 3D space. You need to be able to see what you are working on from all angles at all times quickly and easily. 

Typically in 3d applications, the object never actually moves and instead the modeler or sculptor is changing her view of the model as she works. This can become so fast and seamless that you may feel you are actually holding the object in your hands as you work on it. But that will only come with practise. It literally takes time to develop the muscle memories for how to navigate inside a 3D program. Think of it as similar to learning the moves in a video game. With enough repetition and some time you wioll eventually get the moves down so well, they will become second nature. Blender is like this, just don't expect a high score the first time you play.

If you have a scroll wheel on your mouse, that is what Blender means by the 3rd or Middle Mouse Button (MMB). Just press on it to treat it like a button instead of a wheel.

If you do not have a 3rd mouse button, good thing you turned on Emulate 3 Button Mouse in preferences, now you can use these substitute commands:

Ctrl Alt together to zoom in or out by moving your hand.

Alt key + left click to toggle or rotate your view.

Alt + Shift + left click to pan the view side to side or up and down.

Panning feels especially like you are picking up the object and moving it into position, but you're not, you are navigating or changing your view. 

By the way, I use a Graphics Pen & Tablet in Blender, and I love it. How to set that up if you have one is on an earlier blog post.

But best of all, and completely required for accurate building is how easy it is to snap to a particular view in Blender. This is where using the NumPad or just plain numbers since we turned on Emulate NumPad, will come in handy.

First off, note that there are 2 main ways of viewing 3D space. In the upper left corner of the 3d viewport you will see either User Persp or User Ortho. 

Perspective view in 3D applications is an attempt to make things look naturalistic by imitating the same kind of distortion we would see based on vanishing points. Orthographic views attempt to display objects without this distortion to give a more geometrically accurate representation of the model or scene. As a general rule, plan to work in orthographic view and maybe occasionally toggle over into perspective to see how things look.

Press 5 to switch to ortho.

Secondly there is apparently more than one entity determining the view. I am referring to the word "User" in User Ortho & User Persp. Blender is pointing out that YOU determined this view. You dragged some stuff around and zoomed all over the place and arbitrarily landed on some view or other and then just left it that way.
It doesn't have to stay this way. 

Press number 1.

Blender snaps to a Front view. Now press 3. Number 3 is to the right of 1 on the NumPad. Blender snaps to the right side view.

7 is on top of 1, pressing 7 will snap to a top view.

You can explore the other numbers to change your view in stepped increments, but knowing only these 3 will probably be enough. 

1 is front; 3 is right; 7 is top.

Adding Ctrl to these will show the opposite view. So ctrl 1 (front) is back. 

If you want to see things from the left side, since you know that right view is 3, then the opposite of this will be ctrl 3, and since you know top is 7, then the bottom view will be ctrl 7. 

How sweet is that?

These snap to views are extremely useful as constraints for operations such as extruding on only one axis. Used with the background grid the snapped ortho views are ideal for easily achieving precise alignments and sizes while you model mesh. You'll get to that as soon as you feel ready to start trying to model something. 

Last step and last shortcut for this introduction is how to get rid of the default cube and leave a nice clean scene to come back to. 

X to delete the default cube... or whatever weird shape you are now looking at.

You can see that x brings up a whole list of possibilities. Deleting vertices will wipe out the mesh entirely, which is what we want. You can experiment with what the others do and ways to make use of them later.

Lastly go to Front Ortho view by pressing 1, 5
and press  Ctrl U to save this empty space as your Start Up File, so next time you open Blender you have a nice clean scene to mess around in.

In closing some general advice - don't worry and try not to panic. You probably cannot actually break Blender, though it will crash now and again. Save often. You can get a new Blender anytime for free if you do somehow manage to break it.

Try not to think of this as a task to memorize shortcuts or menu locations. No one, (except maybe you) thinks that you just sit down and learn Blender the way you learn the contents of a brochure or study for a test. No one was born knowing how to model in 3D software. 

The best thing you can do is just make yourself comfortable in there, and the way to that is spend as much time as often as you can trying to use Blender to do simple little things. 

Go on, pretend it's a game that lets you make anything you can imagine. And have fun!

Here are some tuts & links & stuffs that I think are really good for total Blender beginners, collected on a Pinterest page for easy access. 

Other things I think I should mention:

Expect to have to watch most videos several times, pause as often as you like, make notes or whatever. I mention this because I have noticed that the same people who seem to think that they will learn a program and then make something with it, seem to also think they will watch a video and then know how to do whatever was demonstrated in the video. 

Maybe it's me, but I think that's being a bit unrealistic. Seeing may be believing, but it is not knowing, and definitely not knowing how to do something. If you watched a video of someone playing saxophone, would you expect to open up your new shiny sax and belt out a tune? 

Not likely. 

As I said, it will take time for Blender to soak in. You will learn the commands and how to navigate gradually simply from using the program. Eventually these bits and pieces of info will become muscle memories, and Blender will become a fabulously flexible (and fast!) modeling tool that you can manipulate at will. 

All the best,

Sunday, February 16, 2014

What Do You Mean I Should Make My Own LODS for Upload?

This might have been the subject of the first post for this blog, since it is at least once a week that I find myself trying to explain the process for making LODs for upload to SL. 

I dwell on it in my Blender project class(es) at Builders Brewery (back soon really) so If you want more detail on what LODS are or methods of face reduction, see me there. 

Each of the rabbit models shown will be exported from Blender as an individual collada file with a name that indicates which LOD that model is intended to represent. So the upload process is a matter of NOT choosing the "Generate" button and instead selecting "Load from file" on each drop down and then browsing to the relevant collada.  

I hope this quick look at some screen captures from the rabbit sculpture helps to demystify the process. Please do not think you must copy this exactly, figure out the math I used for the reductions or take what I am showing as a prescription for every mesh. Even I do not follow exactly the same process each time, since each mesh is different. But this is a good example of how to take a low poly medium sized object and get great Land Imact (1 LI), optimal physics and good LODS with NO loss of detail or quality in the finished mesh.

If you are brand new to making mesh for SL, you should look for as many examples of how to optimize and possibly retopologise higher poly meshes as you can find. 

And if you are so new to making or buying mesh you have confused High Poly with High Quality or using Higher LODs than you need as a quick route to a perceived high quality mesh, well then you ARE just confused. But maybe this will help you get over it.    

Tuesday, January 28, 2014

Grow Your Own Tree in Photoshop CC

Photoshop CC continues to update and toss in features almost faster than I can discover them let alone learn what they do, so I usually just try to keep moving when something unexpected pops up on a menu.
But sometimes, straying off the path leads to great new discoveries.

So there I was going on my way trying to find out what was up with getting patterns to follow paths (which is great!), when I discovered that the bottom of the new "Scripted Path" menu had something called "Tree".

I think I expected some way of organizing paths in a hierarchy, maybe a mind map kind of thing, Windows style directory tree. You know, pretty much anything except a tree.


Photoshop CC now has a pretty amazing tree generator. Like wow.
Here's how you do it:

I can't wait till there are more options to this. For instance, maybe it will allow you to generate and mess with 3D trees, maybe you can separate branches and leaves on their own layers, substitute your own textures for leaves and bark and so on. But this is a really cool new feature, perfect for quick background content for games, illustrations and 3D renders.

Monday, January 6, 2014

Alpha Modes & Alpha Channels

So, there I was finishing off the notes for a new class on what's what with SL materials and the new edit window when I realized I didn't have that much to say about Alpha Modes. I mean I kinda knew what they were there for and what they do, but why I would care or how I might actually use them was beyond me. In fact I didn't have much to say about them because I didn't know much about what the Alpha Modes could do.

Time for science and mad experiments then, right. 

Step 1 make a tester texture with an alpha channel.

First discovery was the fact that with a texture saved in .png format, Alpha channel or no, transparency on or off, I couldn't get even the slightest rise out of the Alpha Mode button, not so much as a nod of acknowledgement let alone an engaging conversation.

So if step 1 was to make a texture with an alpha channel, step 2 was to save it as a targa. 

I don't usually use the TGA format, I even advise other people not to bother
with it since the file size is so much larger and it's slightly harder, well more technical anyway, to actually create an alpha channel. BUT there are advantages to extra data sometimes, and I think I am going to make 
experimenting with 32 bit tga and alpha channels a regular gig for awhile.

Ok, first thing you might notice is that the invisibility or secrecy of the alpha channel only lasts until upload time. SL right away pegs this texture as an 'alpha' texture and uses Alpha Blending as the default Alpha Mode.

Wait, wait, wait WHAT is an Alpha Mode you ask?

Well, Alpha blending is what SL has been using all along, at least it's what we mere residents have been using. As for what Alpha blending is, think of there being two basic ways to understand transparency. Either it is see-through or it's not is one way, that's the Alpha Testing or Masking mode, which we will get to in a minute. That's also what Linden plants use, ever notice those nice crisp, but oh so jagged edges? And there is another more common, more usual way to think of transparency, which is a blend of see through and not, to produce a surface that is partially transparent. Think gradients, sheer stockings, or whatever. What it means in computer graphics is smoother transitions between transparent and opaque pixels for smoother edges on low res images. But it also means an increased problem with alpha sorting (which surface is in front is hard for a computer to figure out if it can see through the front one/s). And it can be hard to figure out how to make lighting and shadows work across Alpha blend surfaces.

The SL wiki entry says, "..alpha blending should be used very sparingly, and avoided wherever possible." 
Uh yeah, as soon as I quit making meadows and plants, I will dedicate myself to helping others remove trees, plants and hair textures with alpha from the grid. 

But seriously though. Now that we Alpha Modes in edit, we can use Alpha masking for faster renders, fewer problems with alpha sorting and some other nifty tricks.

This time I am focusing on the nifty.

Ok, notice that with the alpha mode set to None, the alpha channel is disabled and the texture goes back to being it's dull fuzzy blue self. This would be a handy trick for the people with accidental png transparency issues maybe. I guess there might be sometimes where it would be interesting or could save 10 lindens to have a single texture be trans and not trans (?) otherwise it's just odd. But whatever, future plans maybe?

Anyway, here's where we get to Alpha masking. I think this is easier to understand under its other name, Alpha Testing. The test in question is whether an alpha channel's pixel meets a certain level of brightness and if it does, that pixel will be treated as fully opaque or visible and if it does not, it will be discarded, or at least not be rendered. Now that I think of it the whole notion of a fully transparent pixel is a bit of a brain cramp, even IF you get past wondering where discarded pixels go. 

But what's important for us is that Alpha Masking is a yes or no, on or off kind of transparency. It's fast, even with values between 0-255, which is what the Mask cutoff is for. Textures using this kind of transparency are easier to light and they do not have the alpha sorting bug.. at all (!). But you will get jagged edges.

In a later experiment, I will take some old plants textures and redo them using Alpha Masking mode to see if the performance over appearance trade off is worth it.

I might even have gotten around to that today if I hadn't nearly lost my mind playing with Emissive Mask. I fiddled a bit with this when Materials were in beta, but I really did NOT get it. Now I think it is just the coolest thing since err fully transparent pixels.

So here's the deal (I think). Emissive Mask is an SL invention to control the Full Bright yuck effect (or something). At least it COULD work this way people. I have frequently wished for a semi-bright or a full bright divided by half. Like little brighter but not so fake, please, and I think the answer is in this setting. Though all you will really see me doing here is turning it on or off, I hypothesize (I mentioned science when we started right) that a grey alpha channel and Emissive mask as the Alpha Mode is our ticket to a subtler shade pain. 

SL treats full bright as an absence of lighting effects, saying in effect, treat this surface as though it is noon no matter what. Can be handy and saves on prims creating lights maybe. I confess, I certainly use it for signage and boxes of sundries for sale, but let's face it, SL would have been less fugly without fullbright. 
Anyway never mind my opining, here's what the wiki says on Emissive Mask:
Diffuse alpha channels support the concept of an “Emissive Mask”, a special kind of mask that defines which parts of the surface that are capable of having their own lighting that’s independent of the renderer’s lighting pipeline. 

Allow me to translate. The fuzzy blue texture I have been using is now more properly referred to as a "Diffuse" texture, and it indeed has an alpha channel -- the words, fish and flowers combo design you can see applied as an alpha channel in the first slide. Using only the Emissive Mask setting on the fuzzy blue surface, the outline of the otherwise hidden alpha channel appears as a brighter version of the texture. In fact, it's a full bright version of that same fuzzy blue texture where the design is at its whitest. The bold face type and stronger white lines in the fish will now be impervious to external and environmental lighting. (yay n stuff) 

But when I hear the word emissive, I think in terms of glows or glowing not a special kind of mask, so it would only be natural that I would have to see what happens with the Glow setting.

Here's a cool fact. Or maybe it's lots of cool facts lol. I had noticed interesting things with editing transparency in world and the persistence of the formerly not transparent pixels as glowing pixels before, but it never occurred to me to use an alpha channel to control the glow amount or whereabouts of glowing pixels.

This was an oversight!