Dancing Auroras

Started by WAS, May 06, 2018, 07:23:39 PM

Previous topic - Next topic

WAS

Just some more atmospheric play with clouds. Low quality early render. Still not happy with them. Been trying to use fractals, but they are not "flat" enough like a SSS to create the unvaried height. Trying to overcome it for more varied shapes with a PF.

luvsmuzik

Is this three layers? Density looks a bit strong to me at this moment. Nice self challenge!

WAS

They are a little dense. Though to that effect, dancing auroras are often much denser than lighter auroras seen at lower hemisphere.

It's 3 simple shape shaders piped into each other and into a single cloud layer. Fractalized and warped a bit, of course.

Oshyan

That's quite a good result I'd say! I've experimented with this stuff before myself and it's definitely challenging to get realistic shapes and color. I think yours is probably one of the best results I've seen.

Btw, just keep in mind that aurora are often photographed in long exposures. That makes things look both more blurry and more "dense", as you put it. It gets a bit tricky then when you consider what look you're going for. What it would look like to the naked eye (in most cases *not* like the reference photo you posted above - that's a long exposure), or what it looks like in the many long exposure "wow" photos that people post. Obviously both are valid choices, and since we don't have the kind of motion blur in TG that would make an "exposure time" function useful in this case, it requires different variations of settings to achieve these results.

- Oshyan

WAS

#4
Quote from: Oshyan on May 07, 2018, 04:36:48 PM
That's quite a good result I'd say! I've experimented with this stuff before myself and it's definitely challenging to get realistic shapes and color. I think yours is probably one of the best results I've seen.

Btw, just keep in mind that aurora are often photographed in long exposures. That makes things look both more blurry and more "dense", as you put it. It gets a bit tricky then when you consider what look you're going for. What it would look like to the naked eye (in most cases *not* like the reference photo you posted above - that's a long exposure), or what it looks like in the many long exposure "wow" photos that people post. Obviously both are valid choices, and since we don't have the kind of motion blur in TG that would make an "exposure time" function useful in this case, it requires different variations of settings to achieve these results.

- Oshyan

Exactly. I have a disabled second cloud layer which is an attempt to create those long exposure blurs from shapes which have moved on (IRL) during exposure. The issue here is I need to create a pleasing PF that can be blurred with the smooth filter. Otherwise I only have smooth step and bevel control which has a hard drop-off falloff and doesn't actually transition smoothly in colour for the cloud density input. It also can't be smoothed which is why I need the PF.

The issue with using a PF is it has noise on the Y axis. This basically entirely kills the vertical stretching that is accomplished with the smooth step of the SSS and the depth modulator. Not sure how to overcome this with PFs. Is there a way to use functions to basically flatten/capture a PFs colour (like the 2D colour in a PFs preview window)?

Again though, these type of pure vertical auroras aren't seen anywhere but high northern and southern hemispheres. Rarely captured on cameras because of their locations. What we normally see, and all over Google Images are auroras at lower hemispheres, and are "planned" events, where a solar flare is detected and they know it will bombard the the lower hemispheres and will not be fully deflected (hence the occurrence of an aurora) by our magnetic field. We have more common auroras at the poles all the time that we rarely get to see, and they're absolutely mesmerizing because they dance vertically due to the source of our magnetic fields. At lower hemispheres there is more horizontal distortion due to horizontal magnetic field (relative to the surface)

Even the picture I provided is in Alaska and not perfect auroras, though near too.

Kadri

Quote from: WASasquatch on May 07, 2018, 04:47:22 PM
...
The issue with using a PF is it has noise on the Y axis. This basically entirely kills the vertical stretching that is accomplished with the smooth step of the SSS and the depth modulator. Not sure how to overcome this with PFs. Is there a way to use functions to basically flatten/capture a PFs colour (like the 2D colour in a PFs preview window)?
...

I am not sure i understand the problem.
Using very high values in the Y axis should basically be enough for a certain high range you need at least?

WAS

Quote from: Kadri on May 07, 2018, 06:16:02 PM
Quote from: WASasquatch on May 07, 2018, 04:47:22 PM
...
The issue with using a PF is it has noise on the Y axis. This basically entirely kills the vertical stretching that is accomplished with the smooth step of the SSS and the depth modulator. Not sure how to overcome this with PFs. Is there a way to use functions to basically flatten/capture a PFs colour (like the 2D colour in a PFs preview window)?
...

I am not sure i understand the problem.
Using very high values in the Y axis should basically be enough for a certain high range you need at least?


Not sure why I didn't think of that. Or even extreme squashing perhaps it on the Y axis. >.<

WAS

Some more tests. Got fractals in place of SSS shapes. Just using a SSS to mask in the fractals.

Dune