Site hosted by Angelfire.com: Build your free website today!
 


41.4 Layout (LightWave.exe) - 4
 
Previous   Table of Contents   Home   Next


Layout, CREATING BLANK SCENE USING LOW-POLY-COUNT OBJECT

I load a scene that has three lights and a camera. You might have to create two spotlights, three total, and position the camera as I've described before.

In the scene editor I set the Deinon object to 'Shaded Solid'.

You almost always set a 'Distant Light' at the front left at 80 - 100% (actually would be the back of your 3D world).

You set a 'Spot Light' at the front right at 50, - 70% illumination (depending on how far away the light is).

You set a third light, using 'Spot Light', as a backlight at the rear right at 40 - 50%.


Check the angle of each light by using 'Light View' and selecting each light. They should each point at the object.


You set the camera at an angle to the object at the front-right. Check the angle of the camera and how it looks using 'Camera View'.


Remember to click 'Create Key' after positioning or rotating each light or ensure that 'Auto Key' is selected.

I would recommend, at the beginning, that you NOT TURN ON 'Auto Key' as it's very easy to forget that you have it turned on.

With 'Auto Key' TURNED ON , it will remember if you forget to create a key, but it will also record every mistake you make.

With 'Auto Key' NOT TURNED ON, you can eliminate mistakes just by moving the slider and any changes you did to the item disappear.

Also, though, if you ever select a different item before you've clicked 'Create Key' for an item, all changes you did to that item will instantly be forgotten, so remember that.

IMPORTANT - Think of 'Create Key' as meaning "Remember the changes I did for this item".

I'll use the low-poly-count version of Deinon that I created in chapter 38 by doing a 'Replace->Replace With Object File...'. It won't look great as I'm working on the animation, but it should greatly decrease the time it will take to create the animation.

In the scene editor I hide the Deinon object. The object will turn to a yellow mesh.


When you choose a bone, the object will disappear. When you select the object again, the mesh will reappear.


Layout, DELETING KEYS

In the past, I had rotated the object and changed the camera's position and/or rotation 20 times in 240 frames.


Since I now wish to make him walk, these old camera movements aren't desired.

If I moved the 'Timeline Slider' to frame 70...


Then clicked to bring up the 'Delete Motion Key' requester.


I could leave it on 'Selected Items', make sure I'm on the frame I wished to delete the key of (frame 70 in this case) and click 'OK'.

After deleting an item's key, the yellow line disappears from the timeline.

I could keep doing that for each of the camera's 20 motion keys, but that would be time consuming. There's a faster way to delete lots of motion keys all at once.


Layout, USING GRAPH EDITOR TO DELETE MOTION KEYS

I choose (or press 'F2') to envoke the 'Graph Editor'.

When you set keyframes for items with 'Create Key or 'Auto Key', you are really specifying a set of keys for that item's 'Animation Channels'.

When you change the camera's position by doing a 'Move', for example, three channels will be created called 'Camera Position X', 'Camera Position Y' and 'Camera Position Z'.

When you change the camera's rotation by doing a 'Rotate', three more channels are created called 'Camera Rotation H', 'Camera Rotation P' and 'Camera Rotation B'.

So, when you open the 'Graph Editor', you're actually seeing how LightWave does some of the magic it does.

Each object will have channels for each bone position and rotation, each light will have channels, etc. Thus, when you are doing a 'Create Key', LightWave creates keys in multiple channels at once. If you move an item, then rotate that item, then click 'Create Key', it creates six keys for position of X, Y and Z and rotation Heading, Pitch and Bank.

You reveal how a channel changes through time by double-clicking the animation channel's name in the scrolling 'Scene/Expression List' window at the bottom left of the 'Graph Editor' window. I double-click 'camera'.


This reveals the camera's channels in the large 'Curve Edit Window' at the upper-right.

Each animation channel is displayed as a line on a two-dimensional graph. Since time is always constant, you can visually judge when an item slows down or speeds up by the slope of its curve.

You can also tell whether an item's transition is smooth or too abrupt; and you can adjust the abruptness of that transition.

The upper-left window is called the 'Channel (Curve) bin'. When you first open the Graph Editor, the curves for the currently selected item are displayed.

In the 'Channel (Curve) bin', I click 'Camera Position X' with my left mouse button, highlighting it, I hold the SHIFT down and click the last channel, 'Camera Rotation B'. All camera channels become selected.


I wish to delete all keys except for the channel keys for frame zero so I start dragging a box by clicking the RIGHT mouse button just to the right of the zero key for the first channel. I drag a selection box down and to the right until every key for every channel other than the ones for frame zero are covered. Large arrow indicates mouse drag angle.


I check that all keys which I want to delete are highit yellow and double-check that it's the camera object that's selected as there's no way of undoing the operation after I've done it. As usual, its best to save the scene before you attempt working in the 'Graph Editor'.


I press 'DELETE' to erase all camera keys except the ones on frame zero.

There is now only the yellow line on frame zero for the camera when I look at the timeline.

I don't delete the key on frame zero, or my camera would no longer be pointed at the dinosaur.

I had also created one key at frame 240 for the Deinon object which rotated him 360 degrees, so I select his body, move the slider to frame 240 and press 'Delete Key'.

Now I have a blank animation with lights and a camera, but no movement when I scrub the 'TimeLine Slider'. It's called 'scrubbing' when you drag the 'TimeLine Slider' to the left and right. If you have any animation keys set, you would see the object move in the viewport as you scrub the slider back and forth.

I drag the 'TimeLine Slider' all the way from the first to the last frame to ensure that I didn't forget to delete any keys.

I save that scene with a new name such as 'Deinon_Blank.lws' in case I ever want to go back to a blank version, then again do a 'File->Save Scene As...' and save it with the name 'Deinon_Run1.lws'.


Layout, FRAME RATE / FIELD RATE

You'll generally be working with two or three frame-per-second values. If you're creating an animation which will run off CD, you'll be using 12 - 15 frames per second.

If you're creating an animation which will be recorded to video in America, you'll either use 30 or 60 frames-per-second, depending upon whether your software can interlace the 30 frame per second version.

If you're creating an animation which will be recorded to video in Europe, you'll use 25 frames-per-second.


NTSC Video Standard (American)

NTSC video (the method that televisions in the U.S. use) shows half of the scan lines (every other line) every 30th of a second (red lines in the diagram), and in the next 30th of a second it draws the scan lines that are between those lines (green lines in the diagram)

Each of these 30th of a second scans is called a 'Field'. So, NTSC has a frame rate of 30 frames per second, synched via a sync pulse that comes every 30th of a second, or 60 fields per second.


Interlace

This method of writing every other line onto the screen is called 'interlace'. Read your hardware manual about 'interlace' to see what rate they expect you to run the animation.

60 frames per second will definitely do it, but some boards lets you also write 30 frames per second to video without problems (the board doubles each scanline, duplicating the second field by shifting the first field down a line to achieve the result). This will allow your animation to properly sync to video.




PAL Video Standard (European)

The standard that is used overseas, called the PAL standard, refreshes at 25 frames-per-second, but has slightly more resolution at 720 x 576 pixels as opposed to NTSC's maximum 720 x 480 pixels for an 'overscan' image with a 4 to 3 aspect ratio.


DVDs

Note that PAL and NTSC are broadcast standards. Televisions use these frame/field rates when they grab the signal from the sky.

Realize that there's nothing about a DVD which makes a DVD either have to conform to either the PAL or NTSC standard. The terms PAL and NTSC are applied to DVDs for convenience, and because of historical convention.

DVDs hold data files with compressed sound and picture information on them. This information can be burned on the DVD in either 720 x 576 pixel (PAL DVDs) or 720 x 480 pixel (NTSC DVDs) formats with varied frame rates (24, 25 and 30 frames-per-second are most common).

It's the DVD player, not the DVD itself, which formats the data file for PAL or NTSC display.

(Just as a note... movies, as opposed to video, show at a frame rate of 24 frames-per-second.)

So, what does this all mean to you? NTSC updates the screen more often, so American television renders movement more accurately with less jerkiness.

PAL is a little higher resolution than NTSC is.

To convert one format to the other is problematic. Since NTSC doesn't show at the same frame rate as PAL, you'd have to duplicate 5 missing frames every second to convert PAL to NTSC or drop a frame every fifth of a second (every six frames) to convert NTSC to PAL.

Moral of the story? If you're in America, write it out at 30 or 60 frames per second with 720 x 480 size. If in Europe, write it at 25 frames-per-second with a 720 x 576 size.


Aspect Ratio
Aspect ratio is the ratio of the width of the image to the height of the image. 4 to 3 (also called 1.33 to 1) is the standard.


Overscan Area

Both NTSC and PAL have a display area which is larger than what the typical TV screen can show. The edges that don't actually show on TV screens are called the 'overscan' area.

If you're going to broadcast video which will be displayed on a television, you'll have to ensure that the edges of your animation won't be chopped off, or 'cropped', by allowing for the overscan area.

Some modern sets show the entire overscan area (there's coded information along the bottom of a raster which contains extra information for text display and such so you'll get strange things happen along the bottom line of a TV if it shows you the entire overscan area).

Older TVs sometimes drastically crop the image.

The 'safe area' (area which is guaranteed to display on all television sets when broadcast or displayed from a VCR or DVD) is 640 x 400 pixels wide as opposed to the 720 x 480 overscan area size.


That means that if you're going to create your animation to distribute on VCR tape or DVD which will be viewed on a regular TV rather than on a computer monitor, you should create your animation 720 x 480 in size so it doesn't float in the center of their TV screen, but ensure that the sides or top of the animation don't get cropped.

You should leave 40 pixels free on all sides of your animation [ ((720 - 640) / 2 ) equals 40 and ((480 - 400) / 2) also equals 40 ].

Since I'm creating an AVI animation which will be displayed via a Macromedia Director script on CD, I'll use 15 frames per second.

You set the frame rate by pressing 'o' to bring up the 'General Options' requester.


Experiment to see what works best with your equipment.


Layout, YOUR FRIEND THE STOPWATCH

A stopwatch is inexpensive, can be bought just about anywhere for around $10, and is absolutely essential when doing animation.


You need to start timing things. Use your own body to act out the motions, clicking with the stopwatch to start and stop it (or, you can get your friend to do it for you as you do go through the motions).

I'll use the stopwatch to compute how many frames it will take for the dinosaur to walk one complete step, from his left foot being on the ground to his left foot being on the ground. In animation, this is called a 'Walk Cycle'.

I'll lumber at about the rate I'd think a dinosaur of his size would walk at.

Deinonychus was about 11 feet long and stood 6 foot high, so he's quite massive.

I walk at different rates, thinking of the dinosaur walking with each step. I start the timer just as my left foot hits the ground, then each time my left foot lands I count. When the count reaches 8 walk cycles I stop the stopwatch.

I use more than one step so the accuracy goes up. I could walk 20 or 40 walk cycles for more accuracy. After ten trials, first walking slower (longest time being 13.19 seconds for 8 steps), then faster, I decide upon 9 seconds for 8 walk cycles.

That's the figure I'll use, initially, then. My guess is that it took Deinonychus about 9 seconds to walk 8 walk cycles of left-foot to left-foot steps.


Layout, CONVERTING SECONDS INTO FRAMES-PER-SECOND

So, I now want to know how many frames it will take for Deinon to walk one left-foot to left-foot walk cycle.

I could use SMPTE (Time code which ticks seconds off, rather than frames-per-second, but then you wouldn't learn anything, would you.)

My animation will be going at a rate of 15 frames-per-second.

That means that in one frame, 1/15th of a second goes by, or .066667 seconds.

If I turn that into an equation:

8 walk cycles is to 9 seconds as 1 walk cycle is to 's' seconds.


cross-multiply...


Solve for s.




So, I now know how many seconds one walk cycle should take. One walk cycle should take 1.125 seconds from left foot touching the ground to left foot touching the ground.

Now I can make another equation. f frames is to 1.125 seconds as 15 frames is to 1 second.



cross-multiply...


With math, sometimes I'm not quite as direct as I should be. I can come up with a single equation by substituting 9/8 in the place of s in the 'f' equation.








So:

Total Number of Frames You'll Need = (Frame Rate of Your Animation * Number of Seconds it Took the Dinosaur to Walk That Number of Walk Cycles) / Number of Walk Cycles the Dinosaur Does in That Many Seconds

So, f = 16.875

That means that I would need 16.875 frames for one walk cycle of left-foot to left-foot steps.

There's no such thing as part of a frame, so I'm going to have to round that off.

Also, I'm going to have to divide whatever number of frames that one walk cycle will be by two, because right-foot to the front is half a walk cycle.

16's a good number of frames. That way, it will take 8 frames from where Deinon's right foot is to the back and left foot to the front, to where his left foot will be back and his right foot will be towards the front.

Next, I have to figure out just what the movement will look like. 16 is a magic number because it can be divided by 2, 4 and 8. When creating a cyclic movement, you're going to at least have to have it be divisible by two, but it should also be divisible by 4 as well. 20 or 24 frames would also have been good choices because they're both divisible by 2 and 4.

If a walk cycle is divisible by 2 or 4, then you can figure out a quarter movement or half movement.

Next, I sketch out what the movement of the legs will look like so I'll know which frames I'll have to use as keyframes. The black line represents his right leg and the green line, his left leg.


I'm not going to start the sequence with either foot to the front, though, I'll start the animation with both legs together. Note that I'll be starting at frame 0, so that means it will cycle at frame 16.

The red frames are the ones where it's half-way between one movement and another.

Now that I know this, since his leg movement is so important, I'm going to use the power of a computer to use Photoshop to draw little cel animations so I get the movement right.


Layout, STORYBOARDING

Storyboarding just means creating key frames. Those are frames where things are definite as far as a finished motion goes.

I have created an 8 1/2" x 11" blank storyboard that you can print to your printer to help when storyboarding.

Use a pencil so you can erase your ideas or laminate the pages and use rub-off markers.


The small square is used to write the frame number and the line is used to write a description of the frame.


Layout, ANALYZING MOTION

It's always good to become completely familiar with the 3D Sculpture movements you'll be creating before you actually animate. If you've figured out that when the right leg comes forward, the right bicep should also follow it; or that the right forearm lifts and the hand hinges downward... then it will be easy to replicate that movement once it comes time to animate.

Observation and research are two ways that you use, when animating, to determine how each body part will move. Most animators know Eadweard Muybridge because he shot sequences of people and animals running, walking, jumping and doing strange motions using chronophotography - then he analyzed them frame-by-frame.

He was particularly interested in knowing, for example, whether all four feet of a horse ever left the ground when it ran. In Sacramento, in 1876, he set up 12 stereoscopic cameras along a 50-foot track to prove whether they did, or not.

You, too, can find examples on the internet or in the library of Muybridge's films which you can use for research material when creating sequences.

Two available books that can be very helpful are, "Animals in Motion" and "The Human Figure in Motion", both by Muybridge.

Shooting film, yourself, can be another way that you can analyze movement. Sometimes a VCR lets you single-frame forward to see what's happening. Idealy, you would be able to single-frame both forward and backward.

And, using your own body is another way of determining what something bipedal might be doing. You walk as you think a dinosaur might have walked, then you think of just the thigh as you move.

Realistic animation is achieved when all parts of an object move. When a person walks down the street, their hips move, their body tips forward as each foot is planted before the other, their arms swing back and forth in opposition to their leg movement, breasts bob up and down... shoulders tip forward as the arm swings.

The more movements that you can recreate, the more realistic your animation will be. Orchistrating movements so things feel right is the secret to animating.


Photoshop, CEL ANIMATING WITH PHOTOSHOP

I'll 'cel animate' the first five frames of the animation (I may draw all 16 before I'm done). Cel animation is how many cartoons are done. It consists of drawing each frame, changing movements in tiny increments from image to image to achieve the illusion of motion.

I load the best version of Deinon, point the mouse over the 'Right' viewport and press the zero key to convert the four frame view to a single-frame.


I then the 'Print Scrn' button to copy a picture of the screen to the clipboard (called a 'Screen Grab').

I run Photoshop and press CTRL-n to create a new document. When an image is in the clipboard, Photoshop automatically knows what size that image is, and creates a new document exactly the right size for it. I press CTRL-v to paste the screengrab to the new document.

I select white as the background color, do a 'Layer->Flatten Image' in Photoshop.

I use the 'Magic Wand' tool to select all sections of gray background, holding the SHIFT key down each time I click so selections add. In the tool bar at the top I set the 'Tolerance' to 5 (in Photoshop 7).

I choose 'Select->Modify->Contract' and contract the selection by one pixel to include 1 more pixel of information around his perimeter as the magic wand tool tends to select too much of the dithered border line.


After selecting all of the gray background, I press DELETE to delete to the background color, white.

I use the lasso tool to select the grid, and any other lines such as green camera view or light source lines, pressing DELETE to remove them.


Photoshop, ARM/TAIL/HEAD/BODY/LEG LAYERS

I separate his body to one layer, each leg and arm, his head and his tail to other layers.

Again I use the lasso tool to outline his left leg.

I use SHIFT to add parts to the selection as I work, ALT to subtract selections and ALT-click with the magic wand in the selected white area to subtract any white area I may have selected from the selection.


I press CTRL-x to cut it from the background and CTRL-v to paste it as a new layer.

I select 'Window->Layers', double-click where it says 'Layer 1' and change the new layer's name to 'LeftLeg Step0'


Step zero is the one where both legs are side-by-side.

With the lasso tool, I hold down the CTRL key and drag the leg back to its original position.

I zoom the view with CTRL-ALT-+ and reduce it with CTRL-ALT-minus to best position the leg. Little white should show around the leg.

I hide the new layer by removing the eye symbol to the left of that layer, then select the Background layer by clicking on where it says 'Background' in the Layers requester.


I outline the thin border line that was left from the magic wand tool.


... pressing DELETE to remove it.

I isolate the body from the background by using the magic wand tool and clicking in the white area, then choosing 'Select->Inverse' to make the body be the selected object.

I cut it to the clipboard with CTRL-x, then paste to a new layer with CTRL-v and rename as 'Body_Step0'

A thin outline is still visible on the background layer which I use to align the body layer (dragging with CTRL held down).

I choose the 'Background' layer, do a CTRL-a to select the entire background and press DELETE to make it completely white.

I will now fill-in the spot where the leg was removed. I select the 'Body_Step0' layer.

I select the 'Clone' tool and from the brush size drop-down menu in the toolbar select an appropriately-sized brush (24 pixels big).


I hold down the ALT key. A black target appears. I move the target, then click with the left mouse button to define the area I wish to clone.


I isolate where I wish the belly to be with the lasso tool. This will keep the clone tool from drawing outside the line.


I keep stamping, then redefining the clone area. You can always turn down the Fill amount from 100% to 17% to gradually build up tones. At 17% fill value, select a dark area to darken a section and a light area of the color you'd like to lighten a section.


I'll use the same leg image for both the left and right leg, and in the same way I'll use the same arm image for left and right arm.

I copy the 'LeftLeg_Step0' layer, paste it, reposition it and rename it as 'RightLeg_Step0' I cut off the tail to a new layer, the head and the arm. Then copy copy the arm twice as 'LeftArm_Step0' and 'RightArm_Step0' .

After I'm done filling in the white sections of the body that were left behind when I chopped the pieces off, I end up with this potato-shaped thing which will be the 'Body_Step0' layer.


This will allow me to rotate and modify each of the legs, arms, head, body angle, and tail angle separately.

In the end I will have created the following layers:
'RightArm_Step0'
'RightArm_Step1'
'RightArm_Step2'
'RightArm_Step3'
'RightArm_Step4'

'LeftArm_Step0'
'LeftArm_Step1'
'LeftArm_Step2'
'LeftArm_Step3'
'LeftArm_Step4'

'RightLeg_Step0'
'RightLeg_Step1'
'RightLeg_Step2'
'RightLeg_Step3'
'RightLeg_Step4'

'LeftLeg_Step0'
'LeftLeg_Step1'
'LeftLeg_Step2'
'LeftLeg_Step3'
'LeftLeg_Step4'

'Head_Step0'
'Head_Step1'
'Head_Step2'
'Head_Step3'
'Head_Step4'

'Tail_Step0'
'Tail_Step1'
'Tail_Step2'
'Tail_Step3'
'Tail_Step4'

'Body_Step0'
'Body_Step1'
'Body_Step2'
'Body_Step3'
'Body_Step4'
I also create a layer which has a red horizontal line in it, 4 pixels wide, to represent the ground so I'll know where the feet are relative to it.

Since the right leg has to be behind the body, but the left leg, in front of the body, I arrange the layers to look like this:



Photoshop, ROTATING AND MOVING PARTS

Making sure that I've got the proper part layer selected I turn off every layer but the ones I need turned on at any particular moment (such as just the 'LeftLeg_Step4' layer)

I select the top section of the leg with the lasso tool. Dragging it away from the rest of the leg by holding down the CTRL key as I drag and the background color will appear where the top of the thigh had been.


I rotate the thigh what I think is an appropriate distance by selecting 'Edit->Transform->Rotate'. I rotate the thigh by twisting at one of the corners.

I move the rotated piece by clicking somewhere in the middle of the piece and dragging it to the desired location.

I can continue rotating it to different angles, and moving it (and any other transform I wish to do on it, such as 'distort') until I press ENTER to finish the rotation/movement/distortion of the thigh.


I do the same thing to his calf, using the clone tool to fill in the gap until it feels reasonable that it's the original leg, rotated.


I do it twice more with his shin and the tips of his toes.


I do the same for each arm, leg, body, the head and tail for all five steps.

In the end, this is what Steps 0 through 4 look like:


Previous   Table of Contents   Home   Next