More realistic balance without real-time physics

Suggest a beneficial feature, or idea's for the future.
Post Reply
Natural_20
Posts: 1636
Joined: Mon Jul 14, 2014 5:25 am

More realistic balance without real-time physics

Post by Natural_20 » Wed Nov 28, 2018 4:02 pm

Creature balance right now seems to be calculated solely from the idle pose, which can result in some weirdness. You should be able to get much more realistic results by calculating the energy cost of standing in a mid-stride pose and then incorporating a fraction of that into the cost of movement. Given the variety of possible gaits, this might require checking a few points in the walk cycle, but real-time calculation should be unnecessary.
Quasar wrote:Dammit, we forgot to sell the psycho drugs and now the chickens are ODing.

User avatar
magmacube_tr
Posts: 501
Joined: Wed Jul 11, 2018 8:01 pm
Location: Now on the depths of the once-volcanic Hasan Mountain

Re: More realistic balance without real-time physics

Post by magmacube_tr » Wed Nov 28, 2018 5:59 pm

Natural_20 wrote:
Wed Nov 28, 2018 4:02 pm
Creature balance right now seems to be calculated solely from the idle pose, which can result in some weirdness. You should be able to get much more realistic results by calculating the energy cost of standing in a mid-stride pose and then incorporating a fraction of that into the cost of movement. Given the variety of possible gaits, this might require checking a few points in the walk cycle, but real-time calculation should be unnecessary.
So your saying that cratures stance should be counted as a energy cost that is connected to the movement :?: Like bipedal cratures movement cost more energy because of their more sensitive balance :?:
Your cube of magma with memes 8-)

Natural_20
Posts: 1636
Joined: Mon Jul 14, 2014 5:25 am

Re: More realistic balance without real-time physics

Post by Natural_20 » Wed Nov 28, 2018 8:41 pm

magmacube_tr wrote:
Wed Nov 28, 2018 5:59 pm
Natural_20 wrote:
Wed Nov 28, 2018 4:02 pm
Creature balance right now seems to be calculated solely from the idle pose, which can result in some weirdness. You should be able to get much more realistic results by calculating the energy cost of standing in a mid-stride pose and then incorporating a fraction of that into the cost of movement. Given the variety of possible gaits, this might require checking a few points in the walk cycle, but real-time calculation should be unnecessary.
So your saying that cratures stance should be counted as a energy cost that is connected to the movement :?: Like bipedal cratures movement cost more energy because of their more sensitive balance :?:
I'm saying that the energy cost of movement should be modified based on how the creature would have to balance at various points in its walk cycle, instead of just how it balances when standing still. There are a lot of creatures that should fall on their faces when they try to move because their legs are splayed too far apart compared to how high their body is raised.
Quasar wrote:Dammit, we forgot to sell the psycho drugs and now the chickens are ODing.

smjjames
Posts: 157
Joined: Sat Dec 15, 2012 11:07 pm

Re: More realistic balance without real-time physics

Post by smjjames » Thu Nov 29, 2018 5:02 am

Isn't there already some real time physics? Theres the streamlining physics which lead to body shapes already familiar from nature (and a tendency to use the ceratopsid type heads. I can't prove it, but it seems like aquatic herbivores favor those type heads sometimes). It should be possible to do realtime physics without straining the CPU. Implementing multithreading will certainly free up resources, or rather, allow resources to be used more efficiently.

Still, having the creatures balance be affected by balance should help things develop more naturally and avoid some of the tricks that Quasar has had to use in order to force them to have all legs be on the ground if quadrupedal for example.

Natural_20
Posts: 1636
Joined: Mon Jul 14, 2014 5:25 am

Re: More realistic balance without real-time physics

Post by Natural_20 » Thu Nov 29, 2018 12:42 pm

smjjames wrote:
Thu Nov 29, 2018 5:02 am
Isn't there already some real time physics? Theres the streamlining physics which lead to body shapes already familiar from nature (and a tendency to use the ceratopsid type heads. I can't prove it, but it seems like aquatic herbivores favor those type heads sometimes). It should be possible to do realtime physics without straining the CPU. Implementing multithreading will certainly free up resources, or rather, allow resources to be used more efficiently.

Still, having the creatures balance be affected by balance should help things develop more naturally and avoid some of the tricks that Quasar has had to use in order to force them to have all legs be on the ground if quadrupedal for example.
The physics aren't real-time, as that would fry your CPU. The creatures' balance is basically calculated at birth and not updated from there. Streamlining is handled via comparison of body part widths and, in the case of heads, a "streamlining" stat for each head type, set manually by Quasar.
Quasar wrote:Dammit, we forgot to sell the psycho drugs and now the chickens are ODing.

User avatar
White parrot
Posts: 1760
Joined: Wed Sep 12, 2012 4:42 pm

Re: More realistic balance without real-time physics

Post by White parrot » Fri Nov 30, 2018 2:15 pm

Natural_20 wrote:
Wed Nov 28, 2018 4:02 pm
You should be able to get much more realistic results by calculating the energy cost of standing in a mid-stride pose and then incorporating a fraction of that into the cost of movement.
I doubt this is the right approach because it doesn't take into account dynamic balance. Fast animals don't stand in mid-stride pose, they are carried away by their momentum: in this fashion, a galloping horse only touch the ground with one hoof at a time. Less pertinently, bicycles are easier to balance on while moving than stopped despite not changing shape in any way between the two states.
Note that dynamic balance is a tricky and risky business (the cost of failure is a fall at potentially high speed, after all) so not all creatures may choose to rely on it. A six-legged creature can easily keep perfectly stable by only moving three legs at a time (the middle one of a side and the front and posterior ones of the other); a four-legged creature could only move one leg at a time but would be quite slow. With even less legs or to gain higher speed, you have to get risky and rely on less and less support points at a given time...
Perhaps it's simply a matter of making mid-stride apparent imbalance less and less relevant as speed increase, but I certainly should read more on gaits and biomechanics.

Natural_20 wrote:
Wed Nov 28, 2018 8:41 pm
There are a lot of creatures that should fall on their faces when they try to move because their legs are splayed too far apart compared to how high their body is raised.
I think it's a bit more complicated by that: I imagine a creature could theoretically propel a leg upward then sweep the other forward before the first leg fall down, repeat. Sounds very energy-consuming, but feasible: the problem here is that they don't move their limbs fast enough to realistically deal with the effects of gravity. Speed is key.
At this point, we shouldn't be surprised by anything nature does. She's like a meth addict whose drug-fueled rampages unfold in slow motion and span millions of years.
Silly Otter wrote:Welcome to the forum.
Please ignore the cultists.

smjjames
Posts: 157
Joined: Sat Dec 15, 2012 11:07 pm

Re: More realistic balance without real-time physics

Post by smjjames » Fri Nov 30, 2018 7:01 pm

Natural_20 wrote:
Thu Nov 29, 2018 12:42 pm
smjjames wrote:
Thu Nov 29, 2018 5:02 am
Isn't there already some real time physics? Theres the streamlining physics which lead to body shapes already familiar from nature (and a tendency to use the ceratopsid type heads. I can't prove it, but it seems like aquatic herbivores favor those type heads sometimes). It should be possible to do realtime physics without straining the CPU. Implementing multithreading will certainly free up resources, or rather, allow resources to be used more efficiently.

Still, having the creatures balance be affected by balance should help things develop more naturally and avoid some of the tricks that Quasar has had to use in order to force them to have all legs be on the ground if quadrupedal for example.
The physics aren't real-time, as that would fry your CPU. The creatures' balance is basically calculated at birth and not updated from there. Streamlining is handled via comparison of body part widths and, in the case of heads, a "streamlining" stat for each head type, set manually by Quasar.
Good point on CPU usage actually, heh. It does seem like the ceratopsid head streamlining setting is off because a frill really isn't going to help with streamlining. In certain positions, maybe, but in most cases it's just going to create drag.

A while back I suspected that it favored it too much and was going to make a post on it, but then the next simulation had it not use the frill, but it still seems like it might be favoring those for underwater a bit much.
White parrot wrote:
Fri Nov 30, 2018 2:15 pm
Natural_20 wrote:
Wed Nov 28, 2018 4:02 pm
You should be able to get much more realistic results by calculating the energy cost of standing in a mid-stride pose and then incorporating a fraction of that into the cost of movement.
I doubt this is the right approach because it doesn't take into account dynamic balance. Fast animals don't stand in mid-stride pose, they are carried away by their momentum: in this fashion, a galloping horse only touch the ground with one hoof at a time. Less pertinently, bicycles are easier to balance on while moving than stopped despite not changing shape in any way between the two states.
Note that dynamic balance is a tricky and risky business (the cost of failure is a fall at potentially high speed, after all) so not all creatures may choose to rely on it. A six-legged creature can easily keep perfectly stable by only moving three legs at a time (the middle one of a side and the front and posterior ones of the other); a four-legged creature could only move one leg at a time but would be quite slow. With even less legs or to gain higher speed, you have to get risky and rely on less and less support points at a given time...
Perhaps it's simply a matter of making mid-stride apparent imbalance less and less relevant as speed increase, but I certainly should read more on gaits and biomechanics.

Natural_20 wrote:
Wed Nov 28, 2018 8:41 pm
There are a lot of creatures that should fall on their faces when they try to move because their legs are splayed too far apart compared to how high their body is raised.
I think it's a bit more complicated by that: I imagine a creature could theoretically propel a leg upward then sweep the other forward before the first leg fall down, repeat. Sounds very energy-consuming, but feasible: the problem here is that they don't move their limbs fast enough to realistically deal with the effects of gravity. Speed is key.
Have we ever had really fast critters in Species? I know there was a thread made years ago where someone unintentionally made the critters faster and faster until the rovers couldn't keep up anymore, but it wasn't a true predator-prey interaction since the critters weren't trying to run away from anything. Might be something worth attempting again since the rovers seem a lot faster now and things have changed since.

Post Reply

Who is online

Users browsing this forum: No registered users and 2 guests