Workable Graphic by crispalomino

Neat how it does a lot.

I'll just leave you off with a friendly critique!
The problem with this filter is that it does way too much. Functionally speaking, you really only want the filter to produce one type of result. In this filter, the filter itself becomes utility for multiple, distinct looks which you've taken the time to explain in the description. However, one of the main thoughts and philosophies of FilterForge as outlined by Crapadilla was to avoid being a jack of all trades filter, which this filter presents 7 distinctly different styles some of which can be grouped to a completely separate filter.

The filter thus far works anyway, but the slider controls for any of the unselected ones are basically inactive and serve no meaning or purpose. This is basically the reason why we don't make filters that attempt multiple styles in 1 filter. Fr om the end user perspective, it's frustrating to see no changes occur when sliders are moved, and since a lot of people don't read descriptions, it's hard to understand (for them) what's going on.

Also note that the output is extremely high-contrast, which I would debate whether this is a desirable appearance or not, although as a designer, I would argue that it's quite overkill. I would dial down the bias parameters or limit them to avoid blown out appearances, which can lead to what appears like an error or troubled spots.

As an advice try randomizing your filter to see how it behaves. When the final output is producing something absolutely unexpected, you'll need to simply limit your sliders a bit to prevent values from going way beyond the scope of what's considerable.

Alright, that's basically me talking about the "stuff to improve on" here's what's cool. You're playing with color. Overall, I can see wh ere you're going with this. I can understand why some of the decisions you made were headed this way. It's just the stuff I discussed above are kinda requiring attention.

And you know... whether you pick this up or not is entirely up to you. I'm just throwing down some thoughts. This is your filter. But it's also a public filter, and given that it should treat people to what you think is awesome but also in your ways make it useable. That's all I got!

I'm in the process of reworking the filter. I've simplified things and will separate things out to different filters, but still grappling with a few nodes. I did however take a look at a snippet that gave me hope, so will post once I figure a few more things out.

Nice filter.
The effects are all good and neat.

What Skybase suggested you has two positive results :
- Reducing the number of components, there are less things that the filter must elaborate every time. This can speed up the filter.
- When there are controls that not always influence the filter final result, it may be hard for the user to understand what the controls are meant for.

But....
( there is always a "but" in all things. Ha! Ha! )
- Sometimes making a new filter only for a small variation from our original filter may make them too similar to each others.
- Other times, without those variations, the filter may look too simple.
So, it is up to you deciding what to do, case by case.

A suggestion I can give you is to make the things as clear as possible simply having a look at the controls. This can be done in many ways. An example :
If you have an image and then you apply several independent effects to it, connected with a Swithch....
- You can call the control of the Switch something like "1,2,3,4 Effect Selector".
- You can call the controls for each effect with names as "1 Effect Blur", " 1 Effect Shadow", and "2 Effect Shape", etc.
- You should order the controls putting the selector on the top of your 4 effects controls, so that the user may guess what is all about.
- You can also write the selector name as " 1,2,3,4 Effect Selector >>>>>" in a way that the name is longer than the names of the other controls below it and it looks like the main control of a following group of controls.