jed

I nicked the idea from this Daily Mail story. I think the illusion is supposed to be the blue and black lines changing length (as opposed to the star thing). I also had a go at animating the out of step wonky bricks thing (there's a building painted with the pattern in the article), but that's still a wip. Here's a still wonky bricks, only circular - needs large monitor There seems to be a lot more to optical illusions than I first imagined. The size of pic and viewing distance plays a part.

Has anyone got this octo running on dynamics ? Getting the legs in sync is harder than I thought. Here's my wonky 1st attempt -

I saw an interesting optical illusion animated gif in an online newspaper today, so I thought I'd waste an hour or so recreating it. The key seemed to be using a bit of trig to keep the moving parts a constant width. I've added a 'reveal'. this file has a control panel, so you can play with the parameters - https://www.dropbox.com/s/21rgcyj8xgwyyw7/star.c4d?dl=1 might amuse someone

Not sure why your track's slipping - have you got the base80 cylinder the same radius as the track rounding radius (and located it midway in the track) ? For bumpy terrain, you might want to look at a dynamic system. I've had limited success, as long as you don't get too ambitious re steering and speed. Here's a straight line dynamic version https://www.dropbox.com/s/zixwjmfrt21ct4t/trackerdynamic.zip?dl=1 I sent you a pm about some new math ideas I had for the non-dynamic tracker.

I built the Lester Banks version and added one of my steer rigs - https://www.dropbox.com/s/zrpkz5i0wwhc5rh/tracker.c4d?dl=1 does the stationary spin on axis thing

I think this is the XPresso you are looking for https://www.dropbox.com/s/gfa9t0yovbohxcm/basewheel1.7.c4d?dl=1 the wheel rotates correctly for any position or rotation of the parent null @HSrdelic can this brilliant bit of Base80 global matrix magic be put in the big file repository before it's lost forever ? added: I watched the Lester Banks videos. He locates the Base80 wheel at one of the cylinders - I think it makes more sense to put it midway between the front and back cylinders (on each side), especially for the skid-steer spinning on axis thing.

You can add the global position vectors of the 2 anchors, divide by 2 and write that value to the mid-way object. C4D doesn't seem to like dividing vectors, so in this file I multiplied by 0.5. The mix node with data type vector, set at 50% would do the same thing. https://www.dropbox.com/s/txv3mrktudeeqez/anchors.c4d?dl=1 added: I've edited my original file to put the XPresso after the anchors in Object Manager, since strictly speaking the calculation comes after reading the anchor position - can sometimes give a 1 frame lag.

You're welcome.

You could use a memory node set to 1 to 'remember' the value at the previous frame. Position velocity is another way. Position velocity is a vector but (luckily) C4D does the conversion to linear when plugged into a real. In this file I scaled the 2 methods for comparison. With memory, you often get spurious values at frame 0 - you could probably incorporate if F = 0 blah blah to fix this. https://www.dropbox.com/s/pvf6gf5781zug9o/vel.c4d?dl=1

Thanks for the comments. Here's another example - has more permutations, uses MS Gothic font. I think that's enough anagrams for a while.

@Abyss integer division in C4D Python caught me out once. Cinema is still on Python 2.7 where 3/2 = 1 not 1.5 (fixed in v3). I was doing some stuff that used current frame in the math - and it gave all the wrong results. Wasted half a day sorting that one out. Since then I always enter constants as 25.0 etc.

I agree - there's a ton of Python tutorials on YouTube, but the hard part is interfacing it with C4D. I find the official C4D Python SDK extremely difficult, so I tend to grab little bits of code from stuff I don't really understand in posts here, CG Society and the plugin site. I guess one plus point is - ordinary programmers just get to (mainly) crunch numbers, whereas we get to move stuff around.

A recent cafe thread about automatic anagram animation got the old grey matter ticking, and here's the current WIP. Might amuse somebody - Python scene file https://www.dropbox.com/s/yvqtmxjmzbukjkm/carthorse.c4d?dl=1 edit : file needs monospace font such as Courier New - I used Consolas.

This bit of python moves the letters automatically https://www.dropbox.com/s/5pvzpxxad4t5lp9/anagram3.c4d?dl=1

If you use Python, you can turn the word into a list of letters, shuffle the list, and make it into a new word. This file uses the monospaced courier new font - works best with mono. https://www.dropbox.com/s/zbmdr4ywcncogl8/anagram.c4d?dl=1 not a serious suggestion. I'm sure you're looking for something more sophisticated - like solving the anagram.

Local means with respect to its parent as opposed to global - with respect to world co-ordinates. If an object has no parent, these are the same. With nested objects that use position or rotation in a calculation, this can be important. In this file the red sphere is a child of the blue cube, and the cube has a vibrate tag. Although the red sphere is moving in world space, in XPresso it seems to have no velocity because it is not moving with respect to its parent. The yellow sphere is also a child of the cube but has its own vibrate tag and in XPresso it has position velocity. Note that you usually have to enable animation refresh under calculate in the XPresso menu to see values change in real time (the vibrate tags have different seeds so they don't move in sync, which would confuse matters). In C4D, some things are absolute, some relative and some have the option to choose. In the video, I think the guy is multiplying not adding - might affect things. Change it it the node drop down menu. https://www.dropbox.com/s/4hatt1md69n8hbr/relative.c4d?dl=1 BTW position velocity is a vector - has X, Y and Z values. When you plug a vector into a real, C4D does the 3-way pythagoras conversion to a linear value automatically.

If you hang a result node on position velocity, it gives an output - are you actually multiplying by a non-zero number ? I don't have your plugins so I can't see what's supposed to happen, but in the video he seems to be measuring the velocity of the parent of the emitter - you are linking the topmost null. Don't know if this matters - it might be a local vs global co-ordinate thing.

For self build, the easiest route is to buy a 'motherboard bundle' consisting of cpu, motherboard + ram. This way you know the components are compatible. Some bundles even come assembled and tested - it's easy for a noob to damage a cpu with static (or bent pins on the socket) during insertion. A search on Amazon for motherboard bundle should give you some ideas in the range £500 - £600, cheaper if you can make do with last year's chip. IIRC Intel are on 7th generation ie 7***, 6th gen are 6*** etc. I've only experience with consumer chips (not Xeon etc), but it seems like the older Xeons are cheap but not very powerful, newer ones are crazy prices. You have to hit the sweet spot re price/power. Don't be seduced by the number of cores alone - do your research. The new AMD Ryzens are mentioned a lot on forums. The blade server you linked would need a rack mount and psu, not the usual computer case. edit - been reading up on blade servers. They are normally cooled by forcing air around the rack - if you look at the Ebay pic, there's no blowers on the cpus. This would be a problem if you didn't have a rackmount case. Also, being rather old, the cpus would run hot - especially when rendering - so the cooling would have to be spot on..

A quick Google tells me that Windows 7 only goes up to 2 processors - see the performance table here link, and then you'd need Pro edition. For 4 chips, you'll need server software. In my little render farm I've got some no frills i7 Windows 10 boxes, built for < £500 each, as and when I had some cash.

I use a site called cpu benchmark where they list various cpu's 'power'. I've found their scores tally quite well with Cinebench and render times eg a chip that scores 10,000 will render a scene in approx half the time that a chip scoring 5,000 would take. The only reference I could find for Opteron 8378 was for a 2 cpu board - which they scored at 5,382 link. In comparison, a modern quad core Intel i7-7700K scores 12,145 link. Doing the math, your 4 Opterons have about 89% of the power of 1 quad core Intel. This is only theoretical, although I've read on CG Society that with multiple Xeons (more common) you don't always get 4X the render power with 4 cpus - usually a bit less. So your server board is almost as 'good' as a modern quad core i7. You don't say what OS it has. C4D R18 needs a minimum of Windows 7 SP1, which is probably the equivalent of Windows Server 2008 R2. If you're into old hardware, it might be an interesting project but personally I'd steer clear.

Not sure how you would sort color - maybe by Hue from HSV. I just about understand colorspace in XPresso, but not in Python. When I've outputted a color from Python, it's just as c4d.Vector ie def main(): global color R, G, B = 0.75, 0.06, 0.8 color = c4d.Vector(R, G, B) print color.x, color.y, color.z # get RGB

Since you are a new member, maybe you (or anyone else) don't know that Kiwi started this forum, and wrote tons of free tutorials. I learned most of my stuff from him. His piston constraint thing is quite inspired IMHO. IIRC, in his piston tutorial Kiwi jokes about the difficult math way to solve the piston problem - I took that as a challenge. That route is not too hard - just some sine, cosine and a bit of Pythagoras see pic

Try this - new divider when box is multiple of 50. Cloner is moved on X to keep it centered. You should put XPresso and user data on their own nulls. https://www.dropbox.com/s/b1w3qwnhp131w6v/dividers.c4d?dl=1

Scene is too complex for me to understand - was expecting 2 pistons, 2 conrods and 1 crankshaft. edit - this seems to work. I did a bit of axis centering and put each piston in a null at an angle. It seems to matter where the XPresso goes, so try either moving your XP in the OM or adjust the constraint priority. https://www.dropbox.com/s/qruhnkzvy7zdzo8/pistontest.c4d?dl=1

Needs a few Daleks