Realistic water

Started by james adamson, December 09, 2020, 04:06:53 PM

Previous topic - Next topic

james adamson

Hi all.
I have seen some of the amazing water some of you peeps have produced. 
I am finding it quite a challenge to get the detail and for it to not look circa 1994.
I have a test I have done here and my reference. The reference is not ideal as its quite hard to see but its the very fine ripples and subtle flotsam
and the way the light is scattered on the surface. I would like to get it right in TG rather than resorting to post fx. If anyone could give me any pointers that would be great as I have spent far too long at this to only produce
what you see. The pace of the movement in my test I am happy with its supposed to look kinda timpelapsey? I used the 4d thingy but if anyone has a more elegant and controllable solution I am all ears. Also it has been all take from me on this site so far and I do look forward to the day I can offer some solutions. 
Cheers James. 
https://vimeo.com/489105080
https://vimeo.com/489099489

Hannes

I can't open your files... :(

Dune


james adamson

That is odd. I clicked em after I put the post up and they worked.
Strange. I will sort it out now.

james adamson

Weird. I have just clicked them and they play. Could be because they are QT's

james adamson


sboerner

Links are working here. I'll leave the technical suggestions to others (water not yet one of my strong suits) but your efforts look really good so far. I'd say you're off to a nice start – natural-looking water flow with an eddy in the background.

The water surface in the reference is a little hard to see. (We walked across that footbridge on a visit from across the pond late last year. Seems like ages ago now.)

james adamson

Hi. Yeah I changed permissions. Thanks for the feedback.
No small task I have set myself I am rebuilding the shot you see in the reference. That was my textures and hdri capture shoot.
I will stick it on site when done for some feedback. I was supposed be shooting people coming across the bridge to make a timelapsey particle stream of people but Covid ruined me plans. No people!:o. So I have had to mash some stuff together from the web which works as a proof of concept but at some point I am gonna have to go deep and look at some crowd generator gadgets.
Fun and games!

james adamson


WAS

One thing I can suggest is try to play with animation on X and Z rather than Y. Y will give you warpy swirls and stuff coming up from the water which will look strange.

james adamson

Hey.
I animated it with the 4d noise. How would I animate in x and z. I do have a transform input on there as well and I have a large right to left animation with some z.
I guess I am after a bit more control over disturbance. You know how on time lapse water you get those sudden wakes form either 
a boat or a very fast wind patch type thing. Seemingly random disturbances that I can control thats what I think I am after.
Cheers.
James.

james adamson

I just remembered. The reason I used the 4d noise was because it looked very two dimensional with just a transform input.
If you have any suggestions for a more controllable way of adding some disturbance that would be great.

Dune

There's an easy way with 'get frame number' and then some warp by vector displacement, but I don't know the clip from my head. Y would be your choice for whirls, and then indeed the X and a little Z.

james adamson

Hi Dune.
Could you elaborate on this method a bit please? Not sure what the get frame would do and the docs online have not 
enlightened me. Also I have plugged a vector displacement into the water shade so it is the parent and I get the same effect whether
there is a vector function input or no vector function input.
Ta.
James.

Dune

If you have your waves (water shader or PF) pull it through a warp shader. As warper, you use a vector displacement shader, and as function input of that you use a get frame number (no input needed). If you set certain values (needs experimentation) in the vdisp shader, each frame will change the waves. You have to take into account the number of frames in a second and the amount of meters you need to shift waves in respectively X, Y, Z, so some calculation is needed.