Underwater Caustic Effects with Perlin Noise

[Cardin]

That looks doable. The tricky part, I think, is getting the right combination of turbulence pattern and displacement settings but I think you should be able to achieve the effect you want...

The way I would approach it would be to first create the turbulence using perlin noise or plasma on its own layer. Set it to invisible.

Then, on a duplicate layer of the actual animation clip to be distorted, use Displacement Mapping (under Distortion)and set the Source as the turbulence layer that you created. Since the turbulence layer is hidden, only its effects on the animation will be visible.

[D.T. Nethery]

The way I would approach it would be to first create the turbulence using perlin noise or plasma on its own layer. Set it to invisible.

Then, on a duplicate layer of the actual animation clip to be distorted, use Displacement Mapping (under Distortion)and set the Source as the turbulence layer that you created. Since the turbulence layer is hidden, only its effects on the animation will be visible.

Thank you, thank you ! I had missed the step to set the turbulence layer to invisible (inactive). If I get a successful test I'll post it . (the actual project I want to use this effect on is under NDA , so I can't show it , but I will make tests first , which are shareable . I'll try to remember to document each step I take ... I'm not good about doing that , so sometimes I achieve a good effect by trial & error , then if I have to reproduce the same effect six months late I forgot exactly how I did it ... )

------

UPDATE: Yes, this works. Created the turbulence on it's own layer , then used the perlin noise turbulence as the Source in Displacement Mapping applied to a layer.



------

Here's a second test. This time I did not use Perlin Noise , I used a photo of an actual piece of ripple glass. I had it move across screen from right to left (using KeyFramer) then used it as Source layer in Displacement Mapping applied to the BG layer.

[D.T. Nethery]

David,
You might get something like what you want from the OpticalFlowDeform FX.
sven

Thanks for that suggestion. I'm going to try that next.

[Cardin]

Looks convincing to me

With just a few adjustments, the same method you used could also simulate heat haze for extremely hot temperatures in a blazing fire, desert scene, jet engine afterburners, etc.

[D.T. Nethery]

David,
You might get something like what you want from the OpticalFlowDeform FX.
sven

Thanks for that suggestion. I'm going to try that next.

The Perlin Noise or the ripple glass texture (used with Displacement Mapping) worked for my purposes. I can't get the Optical Flow to work for a rippling/wavy effect... I don't understand it. It seems to be one of those FX (like Inlay Texture) which uses a lot of CPU and/or RAM (?) and makes my computer go very slow. I could barely move my cursor without the "spinning beach ball" coming on for two minutes . If anyone who has faster computer would care to make a tutorial on how to use Optical Flow for a ripple/wave effect as shown above , I'm curious about how it might look.

[Cardin]

The Perlin Noise or the ripple glass texture (used with Displacement Mapping) worked for my purposes. I can't get the Optical Flow to work for a rippling/wavy effect... I don't understand it. It seems to be one of those FX (like Inlay Texture) which uses a lot of CPU and/or RAM (?) and makes my computer go very slow. I could barely move my cursor without the "spinning beach ball" coming on for two minutes . If anyone who has faster computer would care to make a tutorial on how to use Optical Flow for a ripple/wave effect as shown above , I'm curious about how it might look.

Sorry if you already experimented with this, but did you try the Random feature of Optical Flow?

Under the Edit Tab, there is an icon with two crossing arrows. It will set the vector fields randomly as you scrub along the timeline creating a wave/ripple effect. You can adjust the magnitude of the distortion (length) and variance until it feels right. Not sure how to adjust the speed of the distortion except by increasing the exposure of the frames, but this made things a bit choppy. Hopefully, someone with more experience will offer some more insight, as I'm curious too.

Like your system, My computer slows down on some of these fx as well. I usually test things out on a lower res duplicate project which helps out with rendering speed. Then I import the fx stack into the original project. Except when the parameters are resolution dependent, this works out most of the time.

[D.T. Nethery]

The Perlin Noise or the ripple glass texture (used with Displacement Mapping) worked for my purposes. I can't get the Optical Flow to work for a rippling/wavy effect... I don't understand it. It seems to be one of those FX (like Inlay Texture) which uses a lot of CPU and/or RAM (?) and makes my computer go very slow. I could barely move my cursor without the "spinning beach ball" coming on for two minutes . If anyone who has faster computer would care to make a tutorial on how to use Optical Flow for a ripple/wave effect as shown above , I'm curious about how it might look.

Sorry if you already experimented with this, but did you try the Random feature of Optical Flow?

I have not tried Random. I'll give it a try.

I usually test things out on a lower res duplicate project which helps out with rendering speed. Then I import the fx stack into the original project. Except when the parameters are resolution dependent, this works out most of the time.

That's a good idea. I will also try that. Thanks.

[Svengali]

David,
I posted an OpticalFlowDeform tutorial in a new thread so more people might see it.

sven

[slowtiger]

The funny thing about this "underwater effect" is that you don't see that in nature while underwater. The "underwater ripple" is a film convention, like a speed line in comics. So I'd be totally happy with just a ripple glass being moved across the screen, mimicking the age old effect.

[D.T. Nethery]

I usually test things out on a lower res duplicate project which helps out with rendering speed. Then I import the fx stack into the original project. Except when the parameters are resolution dependent, this works out most of the time.

Yes, the problem I run into when I try to test out a "heavy" FX (such as Warp Grid or Multiplane Camera) on a lower-res. duplicate is that the FX Stack parameters don't work when I import the FX Stack into the original (larger) project .

I just posted a Feature Improvement Request asking if someone knows how to write a script that will handle this task automatically.
See here: Need a script to Resize Multiplane Camera project .

EDIT: although from the replies I received in that topic it is apparently not possible with current George scripting .



.

[CartoonMonkey]

This is beautiful water stuff!
Can anyone think of a way to seamlessly tile it? For use in a video game for example?
C

[slowtiger]

Hm, tiling a replacement map? We did something like this by hand (and on paper) in 1995 for "Asterix Conquers America", of course I don't have those drawings now (studio property), but it's totally doable.

[DiletantBabaewich]

Cool Stuff!!!

It's amazing!
Maybe you have an idea how to make waves of the tide?

THANKS!!!

А чем Вас не устраивает прибой сделанный в TVPaint на 2danimator.ru?

[o0Ampy0o]

That is an amazing depiction of ocean waves DiletantBabaewich.

The funny thing about this "underwater effect" is that you don't see that in nature while underwater. The "underwater ripple" is a film convention, like a speed line in comics. So I'd be totally happy with just a ripple glass being moved across the screen, mimicking the age old effect.

That is an interesting point. Some film conventions are more successful than others. Refraction has many stimulating characteristics and is in the realm of dazzling. The experiment examples in this thread are dazzling to me.

I loved Hitchcock's The Birds. However I disliked many of Hitchcock's trademark film conventions. Thinking about comic book speed lines vs. motion blur and how and why(?) ripples have been used to depict underwater conditions I now understand why I dislike some of Hitchcock's trademark conventions. Specifically, they are those which resemble comic book conventions. Their use disengages my involvement in the story and characters.

Regarding distortion ripples used to depict an underwater environment, this ripple effect is borrowed from a natural phenomenon. One could say speed lines are a crude depiction of motion blur. Because of the way the eye and brain work things are blurred and/or out of focus as we direct attention on specific objects. Everything can be clear if we focus on it (and have good vision). Motion blur simulates what happens when we focus on a moving object. Knowing how, what, where and when to focus and blur when depicting a scene is a visual and audio artist's challenge. Distortion like this ripple effect is experienced when things within a body of liquid are observed from outside the body of liquid. There is a surface of the liquid and some degree of light involved creating conditions for refraction. This distortion successfully communicates "underwater" to an audience outside looking in rather than adding a consequence that does not exist. The audience is being tricked into believing something and they are also being provided what they would see given an unstated reality, with a few rare exceptions the audience is always treated as an observer from afar. I hope this makes sense.

[slowtiger]

Ah, Hitchcock! Undoubtly he's a master of the art. Unfortunately he was a master in the first half of the cinematic century, which means his tricks come out of a quite old bag now. It's tempting to imagine what he would've made out of the cinematic language of today. I guess he would adapt a bit of the speed, but not the deliberate confusion of audiences. He would use comic conventions a lot, as he always borrowed from non-filmic conventions, but he would've never been caught with the bad storytelling and editing of certain Marvel films. (There's this brilliant video essay about Kurosawa which at the end just destroys a Marvel film for comparison: https://youtu.be/doaQC-S8de8.)

Many elements of our cinematic language are 100 yrs old now (I already found the ripple glass trick in Guido Seeber's "Der praktische Kameramann" from 1927). Each year something new has been added, which is always nice and fun to hunt down. Some tricks only last one summer and disappear after that, or get overused so fast that any decent filmmaker avoids them like the plague. Others are so very special to one story that they can't be used successfully elsewhere. But in some cases it's like finding the perfect metaphor: after that you can't compare ot to anything else. (After Interstellar (Nolan) no one would dare to depict a black hole different from his version, which was so accurate that it spawned scientific papers and gave new insight.) The amount of clearly distinguishable elements in a film can vary, 95% use it in a way we use street vernacular every day, but 5% or less use it extensively so we, the audience, will notice and may appreciate it like an old poem full of allusions to classical canon.

As to scientific accuracy: ripple glass comes close to what you see if you look through a non-uniform mix of liquids (or gases) of different densities, like water and oil, or hot and cold water, so it's the correct effect to use for any look over some heat source.