[Steve Bentley]

Hey Kal, if you’ve been reading that then you’ll love this – be warned it’s a rabbit hole.
https://www.youtube.com/watch?v=EVbdbVhzcM4

As a veteran of cloud tank shots that I’ve gladly left behind – I much prefer directing particles that for the most part behave and stay in their trailer until called for – I do kind of miss the physical effects we used to do. I think you had to be more inventive to get the shot. Now you are at the mercy of how good the software is.

A few things:

We found that having the injection point mounted to some sort of armature helped quite a bit – the little perturbations that you impart with your shaking hand – I defy anyone to squeeze a syringe without a little shaking – will set up pulses in the previously calm water.

Then if you can, add a solenoid system for the injection. I was amazed how repeatable the clouds were. Not matte repeatable but you could get a look, and then refine with more pressure, quicker squeeze, higher temperature water, lower temperature paint etc. But I know this is for fun so that kind of control might kill the buzz.

Think about the ink as the matte not the background. Light it so that you don’t have the super deep shadows in the folds of ink you want, but enough that with good a good colorspace, you can crush the footage to get rid of the shadows and then you have your matte, or crush the other way and you get the deep darks in the folds of the liquid. Any remaining holes in the matte you can garbage matte.

Pick a color of paint that easily warps its Hue to other colors.

Make a drainage system. The biggest buzz kill is to have to clean the darn thing after every take. With a drain/refill system you can do a shot every 5 minutes.

We made this grid thing that we lowered into the water – kind of like the divider thing inside a case of wine that keeps all the bottles separate. It allowed the currents to calm down so much faster and then you slowly remove it and let the currents do a final settle. Kind of like a “straightener” in a fountain water jet. Get that that water still!

Indoor lights are better (due to heat) and some IR film in between the lights and tank help to reduce the convection that the lights will set up due to heat. Use LEDs if your high speed can handle them because there is no warm up needed – keep them off until just before the shot. Some LED systems flicker to change brightness and some are all-on all the time. Choose the latter. Pick as “blue” a white as you can handle – less infrared energy.

The other cool thing you can do with LEDS is time them the flicker to the camera’s frame rate, and you then can do a front light/back light flicker that prints to every other frame and if the frame rate is fast enough , every other frame is your matte (back lit) and alternate frames are your beauty lighting (front lit).

Use a stable base – remember jurassic park? Every time you move, you are imparting subsonics into the water and that starts it moving.

Try a salt water layer or a cold water layer or both- if you put in a cold salt solution first then let warm distilled water run down the side the of the tank, to gently fill. The paint will hit that inversion layer and fan out like its hit something – but the “barrier” is clear, so its very apocalyptic.

Adding glycerine or salt or heavier soluble salts to the water can dramatically change the density, which can get those billows happening sooner out of the nozzle and prevent contamination from taking place too soon so the cloud holds together longer. Using another liquid can have some unique effects – think how oil and water just wont mix. I’m not sure what liquid to recommend, we used to do this with (gulp) Carbon Tetrachloride and TriChloretheline before anyone knew it was toxic. But it prevented any mixing and the edges were crazy sharp. (DON’T DO THIS!) But vodka might work or some other high proof alcohol. Keep the space well ventilated!

Some contents or functionalities here are not available due to your cookie preferences!

This happens because the functionality/content marked as “Google Youtube” uses cookies that you choosed to keep disabled. In order to view this content or use this functionality, please enable cookies: click here to open your cookie preferences.

[Steve Bentley]

Looking at a few of your examples, you might be frustrated by AE’s lack of paint. Oh you can paint in AE, sort of, but its clunky and it’s nowhere near the experience you get from Pshop or Flame or Nuke or even Painter (Painter is actually really good at doing painterly animation! Ah, if we’d only had that back in the NFB days).

Is that Norfolk Southern actually Stop motion? Or a mix of CGI and stop motion, or just all CGI?

[Steve Bentley]

Didn’t mean to come down on that point M, but speaking as one bitten by that very thing, I just wanted to forewarn others. That line between video frames and film/photo frames is blurring due to the adoption of Progressive, so many video camera can now be still or more filmic cameras and you can shoot a true single frame.

One other thing to watch out for and perhaps test: on some cameras we’ve noticed what must be a chip charging issue. The longer the chip in the camera sits between frame exposures, the more the image changes. I don’t know if this is sensor fatigue (kind of like the way old time black and white video would solarize at the margin between high bright and deep blacks – the tubes just got over saturated) or poor power supplies or what, but it can create a flickery look, as each frame has a slight difference in exposure and contrast depending on how long the chip sat between frames. Its subtle and just comparing photos you might not see it – but let it rip at 12, or 24 or 30 fps and it creates this feeling – kind of like when they show a projector playing in the movies and introduce a (false) flicker in post to make sure you know its a projector.

One area to do with stop motion where we still see a difference (at least from a filmy look, and apart from grain structure) is with Go Motion (are we still calling it that?). With the camera moving during the exposure to impart motion blur on the stop motion, there is a definite look difference between the rotary shutter of a film camera and the way a chip is “exposed” in a modern digital camera. Neither is wrong, but if you are after a mid eighties look, and can’t get there ,this is usually the reason. Take a look at some super fast motion as recorded by a gopro. Shot from a motorcycle, telephone poles and trees are these crazy wavy things due to how the chip is offloaded. Its more subtle in Go Motion obviously, and that’s part of the problem – so subtle you can’t always pin point the problem.

[Steve Bentley]

Not “every” camera can do stop motion well. Some, when shooting a single frame will only produce an interlaced still as they are expecting to grab another with the opposite interlace to put right after it. This can work but there is a look to stop motion that the interlace seems to fight- its too smooth or something. (soap opera stop motions anyone?)
It really depends on the camera. I’m with Dave, make sure to use a still camera. You can shoot larger than a standard video frame if you want for maximum sharpness and flexibility in post, and you often have a wider range of gamuts to choose from than a video camera would. I’m not a fan of HD and stop motion together – its unforgiving – every little imperfection or fingerprint or even dust is glaring in HD. With down sampling of a large still in post you can get a more filmy feel.

As for the set up – what comes out of the camera is usually final footage that only needs editing and CC. So Premiere might be better than AE. But if you are shooting separate elements and compositing them together then yes AE might be better.
We like to shoot and have that frame fed directly to the computer. Some cameras come with software that lets you review that footage as it builds up in the computer so you can see how the motion is building. There’s even a few that have an onion skinning feature that let you see the live frame overlayed with the last frame shot (or a few of the last frames shot in ever diminishing density) One of the great features of AE is that if you import a series of frames as the frame numbers grow, they will also update in the footage in the time line. So in the absence of a “player” you could view the evolving footage in AE without constantly adding another frame. And you can set up a comp that will simulate that onion skinning with however many previous frames and at what every density you choose.

[Steve Bentley]

Thanks Brian,
It turned out to be two things. First the scale we are working in huge. I still think the density maps weren’t working properly, but they are now. We’re using a few layers of noise to filter which hairs show up, and I think the scale of the noise wasn’t high enough – there’s no indication what space that noise is being generated in – I think its the object’s space, which seems correct now, but I don’t understand why it wasn’t working before.

The other issue was a carbon based one – the dynamics were still turned on for a few of the generators (we thought we had turned them all off) so depending on what keyframe you were on, you may or may not get the results of a dynamics sim all the way from frame 1. I think if you interrupt the dynamics run-in, C4D gets lost and just does the best it can, which can produce different results from test render to test render.

In the end we turned off the dynamics and went with an evolving noise pattern projected on the wheat “hairs” to simulate that shimmer that wind gives to a field.

[Steve Bentley]

Have you restarted AE since the prefs change?

[Steve Bentley]

This would also be handy for creating already matted elements (or elements that are on a keyable color or white or black) where we then need to use a Straight alpha with. Straight needs the color elements to extend outside the matte area and then the alpha cleans up the edge.

I’m curious how this effect came to be in Nuke. Given the example in the video, is it perhaps a testament to the native keyer in nuke that it doesn’t do soft edges very well? It seems to me that this process should be part of the keyers job. (the software not the person)

Ok, back to being contructive: We’ve done this with trapcode’s shine. Its not perfect but it can do the trick depending on the shot.

[Steve Bentley]

I wasn’t able to load the AE project for some reason (different version perhaps). But the usual gotcha for most people is that you have to interpret the UV pass element in the project panel and go to the Color Management tab and choose preserve RGB. You also must work in a 16 bit or higher comp.

[Steve Bentley]

That’s not to say AE does that pruning well. (has it ever?)
It might also be a permissions issue. Do you, as the user, and AE, as the writing app, have permission to read and write to that drive?

[Steve Bentley]

This is in the same vein so I thought A) it might point out a limitation similar to yours and B) might bring out the answer for both.
We have a large 3D environment inside AE (sorry not large, massive! The rearward objects are some 60,000 units from the camera). It seems were being bitten by something similar.
A solid layer that is angled back away from camera covers a huge area and has a bezier mask on it. Near camera it’s great, but off in the distance that crisp edged mask gets fuzzy, even with continually rasterize turned on.
Is that the supersampling and AA turning, in essence, a circle (scaled down really small) into a sqaure made by four dots?