Gerstner Waves reloaded :)


#1

hi,
the initial thread is archived, so i didn’t had the chance to point out this tutorial about waves using clothide
water wave with clothide
ps: in the video at the end of the page, the cube forms waves with higher peaks, so if it’s not seen by camera, it will act as an additional wave generator, at least i think so.


#2

In the orig thread, you also linked to a discussion on cg effects in the movie surfs up? I didn’t bookmark that at the time - wish I had.

If you could, please post again.

I don’t understand why the thread was archived so quickly.


#3

George,

Thanks for the link to the tut site! It’s the first time i’ve seen it and looks great!

Richard


#4

hi, this is the only link i’ve posted-> vagues.
there is a simple approach using a combination of wind deformers here.
(scroll to the lower right corner).

ps: clothide -> in which module is located?


#5

The original thread starter had posted a link from surfs up, but I can’t find the thread now… I thought we could search for archived thread, weird…


#6

Here yah go!-Surf’s Up!

file:///Gog’s%20Project/Surfs%20Up%20prod%20Design.webarchive

I’ve read every page (4 ) very informative :drool:

Enjoy

Bill D


#7

Hey Bill!

As you might notice, the file does not compute!

But thanks for the thought - if you can provide the link or ? :bowdown:


#8

Sooner or later, that system would fail miserably. I could already see some weird stretching that looked like cloth, because ultimately it is cloth.

I’m not dismissing the effort, because it actually could work in some simple cases.


#9

Shetland
Sorry this is really it.

http://surfsup.awn.com/?type=article&artID=1 and it works! sorry for the confusion

Bill-D


#10

Now that’s interesting

appreciated

Paul OP 1984


#11

sadly still a bit cloth/stretchy looking

could really do with a translation of the Gerster maths to put into the formula deformer
(assuming thats possible)
maths here…
http://www-viz.tamu.edu/students/jd/658/T_algorithms.html


#12

Not 100% sure I’ve made an accurate translation, but try the attached file.

The occurrences of:

  • 0.8 : controls the magnitude/amplitude, so higher value = more violence
  • 3 : controls the timing, so higher value = more speed
  • 6 : controls the frequency of Y-peaks, so a higher value will give smaller peak-to-peak distance
  • 0.5 : is just a multiplier I added on the end to be able to control things a bit easier

There’s a lot of movement on the X though, if you’re to follow the maths as far as I can see, so you’ll have to hide the edges like I’ve done here, or use a weight map and a restriction tag to limit the waves to the desired area.


#13

…and having done that, you of course have to look into layering several of these with different parameters and multipliers for more interesting results - as attached…


#14

thanks for all you efforts James

My only thought is - that the key part of this wave type seems to be that its peak - ie changes from smooth curve to sharp peak (and then does a strange loop - see previous link) depending on the frequency or amplitude - can’t figure which.

The wave shape is definitely not your standard sinosoidal wave

another pick half way down here

http://http.developer.nvidia.com/GPUGems/gpugems_ch01.html

much as I appreciate the help - not sure this is hitting the button

unless - of course - I’ve missed the obvious (then humble apologies)

regards

Paul


#15

Yep, that’s exactly what happens if you increase the 0.8’s to something above 1.0 (as described in that link) and set my ad-hoc multipliers to 1.0 as well (or just remove them, just to get back to the vanilla formula).

This however creates massive scaling on the X as well, that’s why I changed it a bit. I’m guessing that if you’d be assuming an infinite surface (mathematicians seem to love those anyway, might be a clue) the X-scaling would not be a problem. So, the sharp peaks and weird looping does indeed occur here, with the correct values, it’s just that it turns out a bit uncontrollable when it does.


#16

Off to tinker…thanks again

Paul


#17

Hi James, you’re slightly out there (by one period) but apart from that right. The formula can actually be copied verbatim into the formula effector fields, I’ve attached a modified version that uses coffee to set the fields values according to the basic formula and userdata (all it is is a sinusoidal form in x going against the cosine form in y which creates a pinching towards the peak).


#18

Neither my error nor your fixing it surprises me at all, actually :smiley:


#19

Well - what can I say. Spot on - nailed it!

regards

Paul


#20

I thought I’d quickly test and put it through my basic waves/ocean rig and shader to see what the results came out like, needs about 8 levels of layering before it starts to get interesting, which = slow results, it’s probably quicker and easier to get decent ocean with just displacement and noises (and a little bit of tweaking using morphs).

and here’s the movie http://www.per-anders.net/general/waves06.mov