Preserved Topic: Displace percentages and ratios (part 1?)

I went back and re-read what Kai said about the math behind Displace. Let's see if we can clarify it a bit because I find it ambiguous.

Brightness Black=0 Gray=128 White=255
Sample From at 100% Black=-128 Gray=0 White=128

So, what does that mean? It means that White in the Dmap becomes what ever pixel is 128px to the right. And that Black in the Dmap becomes what ever pixel is 128px to the left. So, no matter what size your picture is, the most a pixel will "move" is a maxiumum of 128px at 100%.

Now we have 100% / 128px.

We can use that ratio to scale appropriately. Want to find the Mystery Percentage(M)? No problem.

Let's say we want White to take pixel values from 250px down the road. We use ratios to figure it out.

100% / 128px = M% / 250px

Now solve for M%. We get 195.3125%, rather 195%. That percentage works for both Black and White in the same channel. So, what ever is Black or White, will "move" 250px appropriately.

Why is this good to know? Because 128px is 128px, no matter what the size of the picture being Displaced.

I'll get into some practical examples later. Or at least one practical example.

I'm at it again. Got that bug in my head and just had to get it out. Technical slop from yours truly.

Diplace Math and Ratios

Not the greatest, but shows some of the workings of Displace at the core. Foundation and all that. Once I nail the rotate thing and the linearness of gradients, texture wrapping should prove very interesting.

For that broken glass variation I did, I think Broken China is more appropriate and descriptive.

um....... you have too much time on your hands......
very interesting.

You have read Steve's tutorial at Gurus, right? That contains everything I ever wanted to know about Displace, I'm mathematically minded enough to figure the rest out on my own. If you want to get into using math to get pinpoint accuracy with your effects, then maybe u should right a tutorial... with a warning: Only for hardcore math geeks.

Actually, I'm finding this pretty interesting from the point of view of how someone else approaches Displace. I jumped at it in a much more free form, pictorial manner. I needed to know what the tool did, but wasn't much interested in pixel-precise application of it. That's me. I kinda like getting part way and leaving a little room to be surprised. Warjournal is tightening down the screws, getting the ambiguity out. That's cool to see because it's not my nature entirely. Russell Sparkman wrote an article in PEI Magazine that had the opposing gradients in the red and green channels, so you could select an RGB color and have a sense of which direction it would "push" or "pull" in a DMap. I understood it, but absolutely couldn't work with it - I am able to visualize better with a "channel one" and a "channel two" in my head than I am with an RG color mix. So I think what Warjournal is presenting is just as valid and just as ideosyncratic as my approach. As they say in the hat shop: find the one that fits your head.

Math hurts my brain......

Aside from that I also find it very interesting.

It's nice to see that you can actaully pinpoint what will happen with displace.

Actually, some of the Displace stuff that I am and will be presenting is pretty heady. I've got so much more to say about Displace. So much more.

Tick Tock

Like that. It was done with my stock Rotate Dmap with Adjustment Layer > Invert and some creative gradiant (never forget that a dmap is a PSD). I wanted to use a clock, but I didn't have a pic handy. Sure, it's a bauble, but some of the implications can really be exploited.

Too much to say for right now.

<edit>"implications can really be exploited"? That's almost a silly phrase. Folks, just bear with me through a few more of these posts and I promise some really cool things will start happening. With a little luck, it will actually make sense.</edit>


[This message has been edited by warjournal (edited 01-10-2001).]