DDraceNetwork

Oh wow, i just realized it yesterday so i thought that it might be a 16.9 bug

If I remember, it was there since DDNet Wiki was a thing

Alexander

I see the tee bob up and down like 1 pixel when moving, is that intended?

I think that's also in vanilla, so intended

12:15

I think it represents the tee walking

it's part of the walking animation

anim = Animation("walk")
anim.body.frames.Add(AnimKeyframe(0.0, 0, 0, 0))
anim.body.frames.Add(AnimKeyframe(0.2, 0,-1, 0))
anim.body.frames.Add(AnimKeyframe(0.4, 0, 0, 0))
anim.body.frames.Add(AnimKeyframe(0.6, 0, 0, 0))
anim.body.frames.Add(AnimKeyframe(0.8, 0,-1, 0))
anim.body.frames.Add(AnimKeyframe(1.0, 0, 0, 0))

for anyone wondering, the numbers are (timestamp, x_offset, y_offset, rotation) the rotation is in degrees

16:01

the offsets are 1/32 of a tile. negative y is upwards in map/world coordinates. so the body moves between the exact tee coordinates to 1/32 upwards and down again repeatedly

16:05

you can sample such animations by providing a time, in this case we don't actually use the amount of time spent walking or anything, but instead the x coordinate to determine where to sample the walking animation. the animation repeats every 3.125 blocks. so at the same x coordinate, you always have the same body position while walking

@Patiga I would say it's all in world units, in which a tile is 32, and walking repeats every 100 units ^^

Ravie

@Patiga I would say it's all in world units, in which a tile is 32, and walking repeats every 100 units ^^

also true, but for me world units aren't intuitive bluekitty

@Patiga I was making a python module for map animations and used twmap for output, and well env_point.content = (p.x / 32, p.y / 32, p.r) justatest

Ravie

@Patiga I was making a python module for map animations and used twmap for output, and well env_point.content = (p.x / 32, p.y / 32, p.r) justatest

yeah thats unfortunate pepeH

I still think/hope that 1 = 1 tile makes it more intuitive/accessible

in my case having 1.0 for a tile made the values for rotation much bigger than position on average and it messed with optimizing the output

1. sounds fine, how did it mess with optimization?

say if you had a quad move 4 tiles and spin 360, the 360 dwarfs the other values

19:43

it works better if they're not wildly different

I don't understand how the length of the values impacts any performance related stuff

19:44

what kind of optimization are you doing?

removing redundant points from the output, if all values are in the same unit then I don't need a different error threshold specially for rotation

hm I understand

this is just a weird me case tho

Patiga

for anyone wondering, the numbers are (timestamp, x_offset, y_offset, rotation) the rotation is in degrees

i had to find all this out manually troll

I did too :p

I guessed all of it justatest