Chris Smith
Forum Replies Created
-
-
-
-
Here is a screenshot of Aruna’s Shake session for one shot in Hellboy. All of what you see are nodes. That would take a LOT of precomping in AE to do the same effect.
https://www.vfxtalk.com/forum/attachment.php?attachmentid=924
Every composite program I’ve ever used has awesome advantages. Just some are better suited for specific tasks. For FX, I only like node based.
Case study. You have a matte that defines an area of an image for a blur, color correction, and transparency of certain effects.
In a node tree you just pipe that output of the matte node to the mask inputs of the 3 effect nodes in question. Quick and simple. Sure you can solve your needs in AE, but in this case it would take a bit more planning and not near as flexible when you need to change something.
Yes, layer based meaning the main AE layers. Node based apps use a timeline with keyframes just like AE. But instead of having a ‘stack’ of layers that are always in order from top to bottom, you use a composite node of some type. For example, if you want to ‘screen’ something over another image in Shake, then create a Screen node. And pipe one input into it’s FG channel and the other into it’s BG channel. Then your output is the composite.
AE’s “Normal” transfer mode is the same thing as the ‘Over’ node in many composite programs. So when you create layers in AE they all default to ‘Over’ until you change the transfer mode.
The nodes in masks are also called points or vertexes (if 3D). In AE you can only blur/feather a mask globally. So when you feather the mask, the whole thing goes soft. If a mask allows you to feather based on it’s points, then you can offset points from their sister points to define a feathering amount. In AE you get around this by doing multiple shapes and feathering them differently if needed.
The nodes in most node based apps are very minimal. They do one very simple job and that’s it. For example in Shake, you have a Saturation node. That’s all it does. A Blur node, a pan node, a scale node, a fade node (opacity or level). Many of these functions are built into every layer of AE. So where layer based is awesome is when you are doing a ton of graphics, you don’t want to have to build the functionality of each layer from scratch every time. You would need a lot of nodes to equal the built in capability of just one AE layer.
In DS which is what I use, the nodes are extremely functional. Which threw me off at first because if all I wanted was to desaturate something, I felt gulity piping in an entire color corrector just to tweak one slider. But after I was assured that it only uses render time for what is tweaked, I really grew to like the DS nodes. Like it has a comosite node that basically has unlimited layers with opacity control and transfer mode change in one node. The color keyer has a built in tracker and matte shapes for garbage masking all built in to one single node!
So it’s just like a fork, knife, and spoon. Each system can be clearly argued to being better for a specific tasks. While at the same time they can all can be called eating utensils…er….compositors.
my 2 cents.
Chris Smith
https://www.sugarfilmproduction.com -
Here is a screenshot of Aruna’s Shake session for one shot in Hellboy. All of what you see are nodes. That would take a LOT of precomping in AE to do the same effect.
https://www.vfxtalk.com/forum/attachment.php?attachmentid=924
Every composite program I’ve ever used has awesome advantages. Just some are better suited for specific tasks. For FX, I only like node based.
Case study. You have a matte that defines an area of an image for a blur, color correction, and transparency of certain effects.
In a node tree you just pipe that output of the matte node to the mask inputs of the 3 effect nodes in question. Quick and simple. Sure you can solve your needs in AE, but in this case it would take a bit more planning and not near as flexible when you need to change something.
Yes, layer based meaning the main AE layers. Node based apps use a timeline with keyframes just like AE. But instead of having a ‘stack’ of layers that are always in order from top to bottom, you use a composite node of some type. For example, if you want to ‘screen’ something over another image in Shake, then create a Screen node. And pipe one input into it’s FG channel and the other into it’s BG channel. Then your output is the composite.
AE’s “Normal” transfer mode is the same thing as the ‘Over’ node in many composite programs. So when you create layers in AE they all default to ‘Over’ until you change the transfer mode.
The nodes in masks are also called points or vertexes (if 3D). In AE you can only blur/feather a mask globally. So when you feather the mask, the whole thing goes soft. If a mask allows you to feather based on it’s points, then you can offset points from their sister points to define a feathering amount. In AE you get around this by doing multiple shapes and feathering them differently if needed.
The nodes in most node based apps are very minimal. They do one very simple job and that’s it. For example in Shake, you have a Saturation node. That’s all it does. A Blur node, a pan node, a scale node, a fade node (opacity or level). Many of these functions are built into every layer of AE. So where layer based is awesome is when you are doing a ton of graphics, you don’t want to have to build the functionality of each layer from scratch every time. You would need a lot of nodes to equal the built in capability of just one AE layer.
In DS which is what I use, the nodes are extremely functional. Which threw me off at first because if all I wanted was to desaturate something, I felt gulity piping in an entire color corrector just to tweak one slider. But after I was assured that it only uses render time for what is tweaked, I really grew to like the DS nodes. Like it has a comosite node that basically has unlimited layers with opacity control and transfer mode change in one node. The color keyer has a built in tracker and matte shapes for garbage masking all built in to one single node!
So it’s just like a fork, knife, and spoon. Each system can be clearly argued to being better for a specific tasks. While at the same time they can all can be called eating utensils…er….compositors.
my 2 cents.
Chris Smith
https://www.sugarfilmproduction.com -
Sorry Rich. Lost in translation. They need a sarcastic emoticon 😉 😎
Chris Smith
https://www.sugarfilmproduction.com -
[Rich Rubasch] “It was not a particle generator at all. They keyframed each and every circle’s motion, blur and opacity level. That is the only way you can really get that nice fluid, organic look.”
Have you not ever used Particular? You can set random values for color, opacity, depth, transfer mode etc. As far as blur, use 2 or 3 instances of Particular and have them set to 3 different blur amounts. and mix them.
Chris Smith
https://www.sugarfilmproduction.com -
320×240 is standard. If you know that your end users will have Quicktime 7, I’d use the H.264 codec. If not then try one of the standards like Sorenson 3 at 65% quality.
Chris Smith
https://www.sugarfilmproduction.com -
If you want the particles to conform to a true 3D shape, it’s very easy in Maya. Basically model and animate your object in Maya. Convert the model to a softbody. This makes every vertex turn into a particle. Now just hide your geometry leaving just the particles. That’s it. Now just decide how you want to render your particle. Or instance a snowflake texture as a sprite to these particles.
Chris Smith
https://www.sugarfilmproduction.com -
just change the speed on the layer. For details check your manual or help files.
Chris Smith
https://www.sugarfilmproduction.com