# Laplacian node in Flame?

**URL:** <https://forum.logik.tv/t/laplacian-node-in-flame/9005>\
**Category:** Flame Questions\
**Tags:** batch, nuke\
**Created:** [August 24, 2023, 7:01pm UTC](https://forum.logik.tv/t/laplacian-node-in-flame/9005 "2023-08-24T19:01:47Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![onlycarlyouknow](https://sea2.discourse-cdn.com/flex020/user_avatar/forum.logik.tv/onlycarlyouknow/32/2535_2.png) [@onlycarlyouknow](https://forum.logik.tv/u/onlycarlyouknow)\
**Post date:** [August 24, 2023, 7:01pm UTC](https://forum.logik.tv/t/laplacian-node-in-flame/9005/1 "2023-08-24T19:01:47Z")

</div>

Hey, I’m rebuilding a nuke workflow in flame and it uses the Laplacian node. Is there some matchbox that I haven’t found that does this and I’m not seeing it?

I feel like I can just replace this with an erode, a negative and a multiply, but thought I would check in case the math it’s doing is somehow more complex than that.

---

<div class="post-metadata">

**Author:** ![onlycarlyouknow](https://sea2.discourse-cdn.com/flex020/user_avatar/forum.logik.tv/onlycarlyouknow/32/2535_2.png) [@onlycarlyouknow](https://forum.logik.tv/u/onlycarlyouknow)\
**Post date:** [August 24, 2023, 7:58pm UTC](https://forum.logik.tv/t/laplacian-node-in-flame/9005/2 "2023-08-24T19:58:54Z")

</div>

In case any one else wanders into this it’s a gaussian blur, then subtract the original with a comp node.

Not worth a matchbox.

---

<div class="post-metadata">

**Author:** ![PlaceYourBetts](https://sea2.discourse-cdn.com/flex020/user_avatar/forum.logik.tv/placeyourbetts/32/6852_2.png) [@PlaceYourBetts](https://forum.logik.tv/u/PlaceYourBetts)\
**Post date:** [August 29, 2023, 8:11pm UTC](https://forum.logik.tv/t/laplacian-node-in-flame/9005/3 "2023-08-29T20:11:01Z")

</div>

> [@onlycarlyouknow](#):
>
> the Laplacian node

Interesting. I hadn’t heard that term before.  
From [[learn.foundry.com](http://learn.foundry.com)]]([Laplacian](https://learn.foundry.com/nuke/content/reference_guide/filter_nodes/laplacian.html))

> The Laplacian node takes the input image, applies a blur, and then subtracts the original from the blurred input to produce an image useful for edge detection or motion estimation. Apply different smoothing filters to the output to trade off between speed (box) and quality (gaussian).

So a lot like a high-pass filter [How to use Highpass](https://forum.logik.tv/t/how-to-use-highpass/347) or grain theft?

As I do more of my work in sceneLinear now I have moved away from my obscure photoshop blend mode and almost always use blur and subtract.  
The most obvious gotcha is the order of subtraction.