Jump to: Board index » General » Fusion

Shape tweening bitmaps?

Learn about 3D compositing, animation, broadcast design and VFX workflows.
  • Author
  • Message
Offline

Scotter

  • Posts: 2
  • Joined: Tue May 14, 2024 1:33 pm
  • Real Name: Scott Swain

Shape tweening bitmaps?

PostTue May 14, 2024 1:39 pm

Hi -
Newb to Resolve and Fusion.
I have two bitmaps that are different and I want to make a short animation of one morphing to become the other. Is it best to vectorize the images first? Is it necessary? How well does this feature work?
Tips?
Thank you!
Scott
"Scotter" Scott Swain
OceanMedia.net
Offline
User avatar

KrunoSmithy

  • Posts: 280
  • Joined: Fri Oct 20, 2023 11:01 pm
  • Real Name: Kruno Stifter

Re: Shape tweening bitmaps?

PostWed May 15, 2024 11:14 am

Scotter wrote:Hi -
Newb to Resolve and Fusion.
I have two bitmaps that are different and I want to make a short animation of one morphing to become the other.


There is a tool called bitmap in Fusion, used for making alpha masks, I assume that by bitmap you mean raster images not bitmap the tool.

There are many ways to morph shapes in fusion. One method that is quick and easy and you might give a try is this one:

Satisfying Text Morph Animation in Davinci Resolve



I tested that one. Works, well. Here is my attempt.

Just Testing Title Morph In Blackmagic Fusion Resolve




Scotter wrote:Is it best to vectorize the images first? Is it necessary? How well does this feature work? Tips? Thank you! Scott


Weather you use vectors or not, really does not matter. Ultimately its just about the way you set up your node tree. Weather its 3D, 2D, vector or raster, film or photo, you can apply same effects to it. Its just about set up of of how you work with it. Ultimately Fusion is good at going between 2D, 3D, film and photo, raster and vector and doing it more or less seamlessly, but often its not a direct approach its about understanding how to work with each and connecting them in the appropriate way.

Same is for resolution. Fusion is resolution agnostic, so its a good idea not to think about how to make Fusion work like Adobe productions and think about how its meant to be used. Anything generated in Fusion is resolution independent in terms of not having limits resolution, much like vectors. Even paint is vector paint. What gets people confused is reference resolution that limits the rendering of elements in fusion, but its just a reference, and can be adjusted to anything you need. Any text or shape or mask in fusion will scale up indefinitely if you set it up concretely, much like what you would expect of vectors. Any external raster elements such as video or photos, will be limited to the source resolution. If you bring 8K graphics, you can scale it up to 8K, and anything lower and back. Above 8K the source resolution of that particular graphic, you rely on interpolation like any other program. But my point is that I've see to many people think they have to use vector SVG's or Shape nodes in Fusion for resolution independence. And that is just not the case. All things generated in fusion already have vector like properties.

The reason to use SVG's or Shape nodes is not for resolution, its for speed. Shape nodes are faster to render. But they also have many disadvantages because they are meant really for simple 2D style motion graphics, not complex effects. If you only need simple 2D shapes, the shape nodes are perfect. If you need something more, use any other tool in fusion and learn how to use it so you don't worry about resolution.
Offline

Hendrik Proosa

  • Posts: 3106
  • Joined: Wed Aug 22, 2012 6:53 am
  • Location: Estonia

Re: Shape tweening bitmaps?

PostWed May 15, 2024 2:14 pm

Somewhat offtopic. Resolution independence is just a glorified name for using normalized coordinates in some (not all) cases. Whether normalized coords are awesome depends on the context. Having a blur with radius of 0.00372 units is imho less ideal than a radius of 3 pixels. I can easily reason over the latter and know what its effect is, former just obscures the logic from me. Actual image processing in most cases operates with pixel data and absolute coordinates, some specific cases like gpu texture sampling aside (which just does the conversion to abs coords in driver or hardware).
I do stuff
Offline

Scotter

  • Posts: 2
  • Joined: Tue May 14, 2024 1:33 pm
  • Real Name: Scott Swain

Re: Shape tweening bitmaps?

PostWed May 15, 2024 2:15 pm

Thank you!
"Scotter" Scott Swain
OceanMedia.net
Offline
User avatar

Cary Knoop

  • Posts: 1510
  • Joined: Sun Mar 12, 2017 6:35 pm
  • Location: Newark, CA USA

Re: Shape tweening bitmaps?

PostWed May 15, 2024 2:33 pm

Hendrik Proosa wrote:Having a blur with radius of 0.00372 units is imho less ideal than a radius of 3 pixels. I can easily reason over the latter and know what its effect is, former just obscures the logic from me. Actual image processing in most cases operates with pixel data and absolute coordinates, some specific cases like gpu texture sampling aside (which just does the conversion to abs coords in driver or hardware).

Actual image processing has no problem whatsoever to work with non-integer radii.
Offline
User avatar

KrunoSmithy

  • Posts: 280
  • Joined: Fri Oct 20, 2023 11:01 pm
  • Real Name: Kruno Stifter

Re: Shape tweening bitmaps?

PostWed May 15, 2024 4:46 pm

Hendrik Proosa wrote:Somewhat offtopic. Resolution independence is just a glorified name for using normalized coordinates in some (not all) cases. Whether normalized coords are awesome depends on the context.


I suppose so.

Hendrik Proosa wrote: Having a blur with radius of 0.00372 units is imho less ideal than a radius of 3 pixels. I can easily reason over the latter and know what its effect is, former just obscures the logic from me.


I just look at what is happening in the image, if it looks good, I care not what number it says. Numbers are often wild anyway. Its not like f stops on a physical lens or camera. But at the end of the day, its a preference thing. I guess.

Scotter wrote:Thank you!


You are welcome.
Offline

birdseye

  • Posts: 382
  • Joined: Fri Jun 12, 2020 2:36 pm
  • Real Name: Iain Fisher

Re: Shape tweening bitmaps?

PostMon May 20, 2024 12:43 pm

I notice there is a Morph node in Fusion, have you ever used that and how does it work, I'm not seeing it in the manual. I think without using something like the GridWark most morphs I have seen look like fancy dissolves. You could use a Tween node but I guess you would need to get the shapes pretty close. There is another 'trick' with the Disparity node I have never got round to trying,
Offline

birdseye

  • Posts: 382
  • Joined: Fri Jun 12, 2020 2:36 pm
  • Real Name: Iain Fisher

Re: Shape tweening bitmaps?

PostTue May 21, 2024 8:08 am

I think the Disparity node wins by a whisker, probably because it renders in two directions. The Gridwarp could be better but then if it was too good the other nodes would have nothing to do and that's the story I'm sticking with.

https://drive.google.com/file/d/1_qCy_6 ... sp=sharing

Return to Fusion

Who is online

Users browsing this forum: Bryan Ray, Ricardo_2024 and 13 guests