Tyler Paul
Forum Replies Created
-
HERE’S what I came up with.
The matte for the screen effect was made from Fractal Noise set to block and the complexity set to 0, then scaled. Then I made a black grid to match it. The gigantic 3D screen was made with multiple layers of the footage. I used an expression with a slider to adjust the curved bevel around the edges.
Place your footage in a comp. Lable it Top. Make a new mask for your desired shape.
Duplicate the Layer. Lable it Bottom.
Add this expression to the anchor point.
thisComp.layer(index-1).transform.anchorpoint-[0,0,3]On your top layer change the mask expansion to -70
Duplicate the Bottom layer and lable the duplicate middle. (Make sure the layers are in order: top middle bottom)
Create a null layer, lable it curve adjust. Add a slider. Set it’s range from 1-2. Lable it Curve.
On the middle layer add this expression to it’s mask expansion
curve=thisComp.layer(index-1).mask(“Mask 1”).MaskExpansion/thisComp.layer(“Curve Adjust”).effect(“Curve”)(“Slider”);
[curve]Now duplicate te middle layer about 15 times and adjust the curve.
* * *
“Life Should Come With Backround Music”
-Brown Sugar Studios- -
That is so friggan great man. Have you animated it flying around yet? I bet it looks great.
-
That is so friggan great man. Have you animated it flying around yet? I bet it looks great.
-
How ironic that they convert to flash video but don’t allow you too upload it as flash.
-
Awesome! I’m eager to get home and start studing the project file.
-
All that typing and I still missed a couple things. Put an unsharp mask on your draw matte with a radius of 2 and an amount of 100 so to really pull out those nice lines. Also, no matter what size your video is you want to enlarge it to 640 by 480 to get the better detail. Make sure all of your comps are set at 12 FPS. And finally when you render down, prerender your footage, then the combine comp and then render down the post comp to speed up the overall render time. It should be about an hour per minute of footage.
-
I’ll be releasing V3.2 of the toonificator within the next couple days. It’s got some great improvements over the link in my sig so look out for that. Until then if you wanna experiment I’ll give a ya rundown of the concepts I used to create the cartoon and it does a great job of getting rid of the flicker. It doesn’t look like Waking Life so much but rather more like a hand drawn cartoon.
The over all concept is that you have a paint layer which is a solid color matte of your video and combine it with a pen layer which adds the detail drawing and the color grades that the Posterize filter usually creates.
Be warned there is far too many plugins used to include each one but I’ll cover the bases and you should be able to fill in the gaps.
Okay, here we go
First you need to prepare your footage. Drag your video into a comp named footage. Add a noise reduction filter, leave all the settings as is, you want it to wash out your video a little bit. Then do any color correction you think you need. This comp will be used to every step after this.
The paint layer is simple. Drag 2 copies of your footage (comp) in a new comp called paint. Apply a set matte and a brightness contrast to each of them. In the top layer set the matte to key out the darker parts of your footage and turn the brightness down to -60. In the second layer set the matte to key out the brighter parts and set the brightness up to 60. Now add a grey solid below them. The color will look very dull but we’ll pull it out later by increasing the saturation
Now the pen layer. It consists of three parts. The shading (replaces posterize), the detailed drawing (replaces find edges), and the dark fill which adds to the shading. Each of these parts contain a seperate precomposed matte.
For the shading you start by placing your footage in a comp labled shading matte. Reduce the opacity down to 2% and add a white solid below it. Place your shading matte into a new comp and add another white solid. Use a difference matte with a tolerance of 0 and a matching softness of .15. This will create an inverted posterize look with a great reduction in flicker. You do not want the gradient here to cover the entire the radient from black to white. It should appear darker until you invert it. Then it covers all the lighter shades.The next step will add be used to add all the darker shading.
For the Dark Fill layer you take your footage and put in a comp I called the Dark Draw. Add a brightness/contrast filter and set both settings to about 65. This wipes out all the lighter shads and only leaves the darkest part of the video. A fun trick used often in the next steps is placing CC glass onto the pen steps and should be used here. This will distort the details of the cartoon to make it look more hand drawn and imperfect. Make sure you turn the specular shading parameter all the way down and adjust the amount of distortion so that it’s subtle but noticable.
Finally the drawing layer. Instead of the find edges filter, you can create much finer lines with emboss. Drag your footage into a new comp called draw matte. Add a brightness/contrast filter and set them both to about 45. Also add CC glass with your previous settings. Drag two copies of the Draw Matte into a new comp labeled Draw. Add an emboss to the top layer. Set the direction to -45 degrees and the relief to .1 and you may need to turn of the contrast a bit. Copy the emboss paste it onto the second layer and change the angle to 45 degrees. Now make a solid grey layer (128,128,128) and used a difference matte to pull out erase everything but the lines and use alpha levels to clean up the results. You can strengthen the lines as well erase any garbage.
Now you put these three parts together to create the pen details. Start by dragging your shading matte into a new comp labeled pen. Invert it. Add a white solid below it. As I said, it’ll should look pretty light and that’s what you want. Then drag your Dark Draw above the shading. Change the blend mode to color burn. Now add the draw layer at the top. Add a tint effect and set white to black as well. And add another CC Glass.
You pen is done and we’re almost finished.
Drag your paint into a new comp labeled combine. Place a black solid below it. Add your pen layer with the visibility turned off. On the Paint layer add a set matte and set the matte to the pen layer. You should now see a very dull looking cartoon. Drag your Combine layer into a new layer labeled post. Turn the saturation up to 30 using Hue/Saturation. Add color curves and make your darks a little darker and your lights a lot lighter. Last add an Unsharp Mask to make everything look crisp and clean.
Through out these steps I used a variety of plugins to clean up each part if they needed it. CC vector blur with the smoothness and the map softness turned all the way down and it’s amount turned up turns any flickering edges into more of soft diffusion. Alpha levels are great for controlling certain layers for making the pen. And there may be a few more but I don’t recall where I used what. this’ll get ya started if you feel like experimenting. Or maybe just reading this will give ya some ideas to create your way of doing it.
This is what a photograph of my siblings and I looked like when I ran it through. Go ahead and click the sample vid in my sig to get an idea of what the video sorta looks like. (It’s missing some of the steps that I layed out for you here)
-
Very nice. I can’t wait to see the final.
I had already had my cube set at (0,0,0) so that was nice. I also had the sides spinning but Dan’s expression seemed like a better idea so I went with it. The boxes are all set at (0,0,0) and I have used 2 radii values depending on whether they’re a Side or Corner piece. I divided the Degrees by 57.295 to get the radians because I couldn’t get toRadians() to work so that’s my first question.
Now I’m coming up to the part that made making the cube so hard in the first place and that’s keeping track of all the cubes. In programming I’m used to being able to add varaibles to a parameter and that amount sticking. (Example: a=a+1 and a will remain at that value until you tell it otherwise) If this were the case I’d be able to figuare out what boxes made up a face twist them around drop them, Choose a new face and again look at box positions to find the correct boxes for a face twist etc. With expressions as soon as I let go of a box (I change which face I’d like to control) the boxes reset to their original values.
My current idea is each box has an angle control for the 6 faces of the cube. Degrees are added to or subtracted from depending on the turns I tell it to make. This would allow me to keep track of every twist and turn a box has gone through. How can I make the angle controls count up 45 degrees and then stay there. I’m pretty sure I’ll end up with the same problem as I had just moving the boxes and expecting them to stay.
Are their any other ideas?
* * *
“Life Should Come With Backround Music”
-Brown Sugar Studios- -
Very nice. I can’t wait to see the final.
I had already had my cube set at (0,0,0) so that was nice. I also had the sides spinning but Dan’s expression seemed like a better idea so I went with it. The boxes are all set at (0,0,0) and I have used 2 radii values depending on whether they’re a Side or Corner piece. I divided the Degrees by 57.295 to get the radians because I couldn’t get toRadians() to work so that’s my first question.
Now I’m coming up to the part that made making the cube so hard in the first place and that’s keeping track of all the cubes. In programming I’m used to being able to add varaibles to a parameter and that amount sticking. (Example: a=a+1 and a will remain at that value until you tell it otherwise) If this were the case I’d be able to figuare out what boxes made up a face twist them around drop them, Choose a new face and again look at box positions to find the correct boxes for a face twist etc. With expressions as soon as I let go of a box (I change which face I’d like to control) the boxes reset to their original values.
My current idea is each box has an angle control for the 6 faces of the cube. Degrees are added to or subtracted from depending on the turns I tell it to make. This would allow me to keep track of every twist and turn a box has gone through. How can I make the angle controls count up 45 degrees and then stay there. I’m pretty sure I’ll end up with the same problem as I had just moving the boxes and expecting them to stay.
Are their any other ideas?
* * *
“Life Should Come With Backround Music”
-Brown Sugar Studios- -
Plan Nine. Seen it. It’s great. Along with Ed Wood with Johnny Depp. And most recently I got my hands on a copy of Glen or Glenda.
Anyhoo continuing on,
I did think it through and the scale of an object increases exponentially as it nears the camara. That trumps the idea that you need to move in from 50 feet away.
Assuming the model is 10inches wide you can move the camara from a couple yards away to an extreme closeup in 1 second. Slow the footage down 50% so that it’s two seconds long and composite over a large field. The rate of exponential growth of the plane compared to it’s surroundings will make it appear like that baby is really cruising. If that still isn’t enough, then keyframing a basic scale change from 50% to 100% will multiply the speed at which it seems to be going but still preserve the exponential growth as well as the more realistic flight look that should come with moving the camara as opposed to the model.
I’m not going to pretend that I know exactly what I’m talking about here. I’m just brainstorming ideas. I certainly don’t believe you’d have to move your camara from 50 feet away.