kameelian
Iterator
Posts: 181
|
|
« on: June 29, 2011, 02:54:13 PM » |
|
How to create 'unique' fractal art (or how not to).
This is in relation to Mandelbulb 1.6.9.6 (but I assume it will be relevant to any newer/older versions)
I'm Sorry if this has turned out a bit lengthy. I guess only the 'keen' will read it all.
There are some general 'tips' at the beginning, but the lengthy bit at the end tries to answer the question of creating unique fractals, so it might be worth persevering with it...maybe...
I am aware that full and detailed documentation/explanatiion for Mandelbulb3D does not exist and what is around is quite sparten. So, in a bid to 'give something back' to this community, I am just adding to the pot of available info, by sharing a few of my working experiences and observations to try to help other newbies. I am in no way an expert, in fact I have only done just over a month with it. I do not understand the maths, this is purely from my own trial and error experiences with the M3D program and if I get something wrong maybe an expert can put me right on it. I did start making a few notes in a bid to remind myself. However, in different formulas, I have learned that what worked once does not necessarily work similarly again - so I stopped making notes. These 'tips' are just some of the more coherant notes to myself. I am not very organised and have got into a right mess doing variations of the same thing as tests and then not knowing which version is the right one(!)
Tip 1)
SAVE OFTEN(!) AND don’t forget to save EACH of the options m3i, m3p, m3l, m3a etc
Tip 2)
Say you have (only) 2 keyframes you want to animate between with 25 intermediate frames. How many frames will you end up with? 2+25 = 27. But NO...you get 26. If you have three keyframes, you might get 51 (not 53). For some reason, M3D does not seem to output the first frame; instead it has already started to create the '2nd' frame (as ani000001). It does not merely give you the two end frames and all those between, So if you want the first frame to actually appear, you will have to add it manually. You DO have a copy of it don't you - as it is a keyframe. (If I want the keyframes to actually appear in the anim, I have, perhaps a bit clumsily, put two identical keyframes with a 'tween' of only 1 frame between them and this does seem to work).
STOP PRESS! I have since been informed that the above is relevant if you use the Quadratic Bezier option in the Subframe Interpolation box of the anim window. Apparently, if you use the Linear method, you do indeed get frame 1 into the anim and all the keyframes are visited - although the end result may not be as aesthetically pleasing. THANKS bib and cKleinhuis for the info(!)
Tip 3)
Now, in addition to this, even if you only want the resultant frames between two keyframes, be aware that even these frames will be influenced by whatever third frame you may add in beyond them. Oh yes. [[Again, I am going to assume that this is not the same for Linear interpolation]]. Try it by merging two specific keyframes with a variety of 'third' keyframes and see what differences you get between the two keyframes you want to animate. I do not know at which point in the tweening, the second keyframe starts to have impact on the previous one - presumably from the start, otherwise all the early frames would be identical, but less obvious is - at what point in 25 subframes, the third frame's influence kicks in. I have not worked this one out. BUT...although you think you are only morphing between two frames, strangely it DOES matter what the 3rd (or more) frame is in the anim list...Maybe an expert can shed some light on this?
This is remarkable, really. If you want to transform between, say, a dozen keyframes and the last one has some movement in it that you want to slowly emerge throughout the animation, it could in theory work - depending upon when it kicks in. However, in practice, it is harder than you'd think to control things. BUT I do know that if the formulas you are manipulating are sufficiently compatible, if you add a third frame after the two you want to morf between, you do get different results, depending upon what that third frame is - again, sometimes useful and sometimes not. Gosh, it's never ending possibilities.
Also, when animating, I have seen many, many times, a comment something like: frame '234 of 210 completed' in the anim window (i.e. it has not stopped rendering at the final frame). How or why this happens (and so often), I have not sussed out. Is it influenced by the number of keyframes indicated in the final keyframe (which will never be implemented unless 'loop' is on)? I don't know.
Tip 4)
Far Plane (in navi window) seems to bring in/send away the fog. Higher numbers bring more detail into the mid and far distance (ie reduces fog) Increase ‘far plane’ (in navi drop down) if things start to disappear as you approach (i.e. the fog predominates).
Tip 5)
Not only DE stop (.8 to.1) but also ‘ray step multiplier’ (0.1) [in 'calculation' of main window], will reduce those grainy renders – sometimes. However, it slows down rendering.
Tip 6)
If you want to continue animating where you left off a previous attempt, be careful that restarting rendering WILL erase your previous annim00001 picture UNLESS you put the new start number (one following the previous final number) in the ‘Start with index’ on the animation window. Simple but easy to forget.
Tip 7)
‘Max Its’ in navi window seems to bring the colour cycle through the image. Numbers like 2000 maintain a more constant colour, 200 brings in rainbow effect (dependant upon ‘light’ settings).
Tip 8 )
On menger3, Scale = 3, CScale X = 4, CScale Y = 1 (seems to) remove al that cutting through the scene. (but is this only on the particular anim?). It seems that, for the in-between frames, the numbers become really fine. E.g. 3 might become 3.24479166666667 and only this one is the exact pic you want.
Finally, in a bid for accuracy, if anyone else has read haltenny's pretty darn good beginners tips with pictures (thanks haltenny), you will see that he said in tutorial 1 that the very small numbers after the decimal point did not really do much and so he recommended getting rid of them and using only numbers before the decimal point. Whilst it is true that you can substitute 10 for 100 and nothing will seem to happen, I can categorically state that, whilst sometimes changing big numbers has no effect at all, at certain times, changing the smaller numbers - even by one number, can have a huge impact - I guess it just depends on where you are in the formula. For example, if you have 1.23456789 and you change it to 1.234, this can make a huge difference - as can changing it to 1.34567 or 0.123456 or -1.23456789 or 12.3456789; you (I) just never seem to know when it's going to be radical. Have a play around and see.
Big thanks to Jesse for a brilliant program. Very clever indeed. M3D Fractal creation seems to get its hooks in you, doesn't it.
Oh, and I still have no idea the differences between formulas and parameters. Thick or what?
---------------------------------------------------------------------------------------------------
- SO then, onto the idea about unique fractal animation creation
There are gazillions of brilliant pieces of fractal art across the internet. But, as you see more of them, you will begin to notice some obvious derivatives of others (some moreso than others), purely based upon the fact that we are all tweaking the same basic formulas. This is not a problem, there is still plenty of scope for everyone, but how on earth do we find something radically new? How do you get to places no one else has been yet? Is it even possible?
I don't know if every image that we can possibly come up with is already encoded to eventually be 'found' by one route or another (they are fractals afer all) - or whether using the program in different ways (adding new formulae) will enable unimagined new stuff to emerge. Probably.
I don't know for sure, I'm only just past my fourth week, so this may be obvious to more experienced users but, as I have not seen anyone asking for advice when this technique goes 'wrong', I shall assume it is not a technique in common use by many(?). I'm sure someone will tell me if I'm stating the obvious or abusing the program. Even if no one else has posted this as an issue - I can't believe that there is no one who has not tried this as an option - or got into this mess as a result of experimenting.
I've heard talk of 'hybrids' and I have assumed that this is what I'm doing(?)
SO... what I have liked to do is load in a chosen parameter set from, say, the clipboard (or an m3i) and put that as the first keyframe for an animation. Then what I do is load in another parameter set from a completely different source and stick that as the second keyframe in the animation window. Then what I try to do is create an 'animation' between the two images in say 20, 50 or 100 in-between frames. Often, what it spews out is a load of garbled rubbish; rapidly changing random shapes, cut through blocks, 'sunsets' or single-coloured screens but, every so often, I get a little gem; a completely brilliant image or two that I will want to use as the starting point of a completely different (and hopefully original) set of animation keyframes. (parameters for even further 'merging' with others later).
However, unlike normal (old-fashioned) morphing or 'tweening', where the images simply merge between the two pictures, you do not not usually get the same outcome in Mandelbulb 3D. Although you CAN morph between two 'pictures', you don't necessarily get an obvious merger between them. Presumably, as they are pictorial representations of maths formulae, it is the formulas that are tweening and they can produce some bizarre outcomes - nothing like the original pictures used for the animation keyframes. Of course, this may turn up your real gem - if you manage to save its various forms [m3i, m3p, m3a etc] (I say 'if you manage' because, more often than not, some very weird results occur at this point).
You would think that if you stopped the animation whilst the subframe you want is in the main window, you would then be able to save the M3I of what you are looking at - and sometimes this is the case.
But there also lies a problem that I have been unable to fathom or to correct:
The weird thing is, whilst you can usually save the picture being displayed in the main window as an accurate bmp, JPG or png, it is evidently not a WYSIWYG. If you save that same picture as a parameter or complete M3i, (or if you do a recalculation of the 3D to increase the image size), it very often gives a completely different picture to the one you are seeing, so do not be surprised if the values that load up next time you use the m3i or m3p are completely different to the bmp/png you expected. It is as though it presents you a picture but preserves the parameters 'behind' what creates it - and these are often very different. This is quite disappointing if you thought you'd saved an excellent picture but all you have is a low res or small bitmap of it - and some totally alien parameters, thinking that you could go to work on it later. Thinking that the issue may be to do with how the formulas left in the formulas box re-impact upon the image in the main window when recalculated, I would expect that if I don't recalculate, I should still be able to save the picture showing in the main window as m3i or to the anim window or navi window - but this is not always so. I have put a zero in the 'iteration count' box on all formulas currently in use and tried to resize the image and recalculate 3D, but then nothing at all happens. I have tried rerunning the anim from the number of my wanted frame and then paused and saved the various files - and sometimes this works.
SOmetimes, somehow I do manage to get the picture as an m3i, but then (at best) the colour might be completely wrong (and either uneditable or at least can't get back to the original colour - and it is amazing how much diffference the colours/shadows make) or it might not get sent properly to the navi or anim window (where I thought I would be able to resize it and send it back to the main window).I have seen the message: 'error: formula option is not valid' in the box at the bottom right of the main window on more than one occasion. So with all options so far, I still can't quickly get to a place where the correct image is editable every time - and I have spent many hours getting nowhere in an already time-consuming hobby - which I love, incidentally - which is why I persevere trying to get the original imagery that I am after. That said, I do often choose ones that are still a bit familiar too.
If you give it a try, you might find this outcome or you might not, because it does not seem to be totally consistent - presumable depending upon the compatibility of the formulas/parameters in the anim(?)
Can anyone who is very much more experienced shed any light upon how to get the image that I see into an editable status?
Whilst I do, of course, also use the program conventionally - and get recognisable results, I think that the above scenario might offer some explanation to some of the difficulties I was experiencing on my first weekend with the Mandelbulb3D program and brought here as questions - that folk seemed to think had not been asked about before.
I'm sorry if this is as garbled as some of the imagery I have created, but if you're still with me to here, is there any advice - or thanks for the berrrilliante tip?
That's all for now folks.
(I've since added a sample 'M3D test' of this issue for anyone who wants to try it)
cheers Kam
|