blueprint

[lofty rapids]

i feel like having to go through functions is less efficient then just traversing the code

as a for instance when i program i tend to have just one file

untill its fairly complete

then i'll split it for the sake of it looking better

is there a fast way to go backward in display in blueprints ?

like for instance if i was to make a bunch of functions

if i dbl click it

and goto it

can i quickly jump back to where i was ?

[maiden wadi]

In the same blueprint, yes.

Arrows here.

[lofty rapids]

awesome thank you

i never even noticed them

[maiden wadi]

This is one of Atre: Dominance War's most convoluted BPs at the moment. I'm a function junkie. 😄

[storm orbit]

You can also have multiple event graphs (for better organization)

[lofty rapids]

so nice very organized

[lofty rapids]

right like collapse stuff

[storm orbit]

Well, I'm less a fan of collapsing nodes - you ought to make a function if you need to encapsulate or group some logic. The exception would be a Math Node

[lofty rapids]

i never saw those arrows which is ok, but still its easier for me to visualize whats going on in a single go

i do like functions and i should clean it up, but i do this later because no one is usually even looking at it lol

[narrow kite]

The revamped back end combat calculations are nearly finished. The enemy dmg isn't finished, just the player dmg. It was a little weird on the coding side, but I think I got a grasp on it.

[lofty rapids]

what are you using for the enemy health bars ?

[narrow kite]

All the functions where moved to the 3rd person player BP. So they can be used universally from any circumstance. If a projectile collides with the player it will run these same calcs as a enemy hit or the player stabbing.

[lofty rapids]

theres something about the way these player healthbars are facing ...

how are you rotating them ?

[narrow kite]

I may have slowed down the tick rate

[lofty rapids]

is your healthbar centered ?

[narrow kite]

Maybe to slow. What is the issue? Not updating quick enough? Like offset?

[lofty rapids]

i don't think its pointing in the right direction

[narrow kite]

You mean the direction it's facing or how it's attached / location?

[maiden wadi]

I recall someone doing world space widgets with a vertex shader rotating it. Haven't tried that myself.

[narrow kite]

It may be 2 separate things. On an enemy jump animation I noticed it's not tied to the mesh for example.

The other is the facing direction which is what that code does. So the rotation is updating to face the player

[lofty rapids]

20 to :23

its not pointing correctly

[lofty rapids]

^.^

you can see the rotation is weird, i bet if you play and look left/right the things don't look right

[narrow kite]

I'll remove the modified tick rate and see

[lofty rapids]

it could be that its not centered

i personally wouldn't rotate it directly at the camera i ahd a hard time with that

i rotate it toward the camera but like flat where the char is looking

so it looks basicaly the camera but the y is at the same position

something like that

but since you do look directly at the camera, im going to guess its not centered

[lofty rapids]

i think i see, its far back from the character

its not directly above ?

probably why its rotating like that

[narrow kite]

The placement is on the image above. However, shouldn't it be connected to the mesh asset or something? I can't put it as a child class for example of the mesh by just dragging it.

[lofty rapids]

no its fine but it should be directly above

is it set back a bit ?

because think about that when you rotate

[narrow kite]

Yes probably.

[lofty rapids]

try it centered

directly above the player

so when they rotate it doesn't skew like that

[storm orbit]

Since the widget is parented to the enemy actor, setting the world rotation will get COUNTER rotated by the actor's z-rotation as it turns around. You can subtract the actor's z-rotation from the result before plugging it into the Set World Rotation

[maiden wadi]

Or just set the component as Absolute rotation so that it's parent hierarchy doesn't rotate it.

[lofty rapids]

i think its just set back too far if its dead center it will look way better

[narrow kite]

This is the placement btw. I moved the skeleton slightly for centering

[lofty rapids]

so it is centered already ?

is that your pivot off to the side like taht ?

am i bugging out or is that pivot not centered ?

[cloud token]

Rigged skeleton

[narrow kite]

The rotation in that image is from the skeleton when I moved it's rotation slightly

[lofty rapids]

ya i mean somethings off, you can keep if you like it but i feel like your rotation is off

[lofty rapids]

this would be the problem

[maiden wadi]

Are you talking about the fact that the widget rotates like it's own right side is at the center of the character?

[lofty rapids]

that the center of that rotation widget is not in the center of the capsule

[maiden wadi]

It's because of this.

The widget component is centered, but the widget in it isn't. Too big, and offset.

[lofty rapids]

something off center for sure

[ruby cobalt]

hey I need architectural advice here.

I have a parent class that I use for almost every actor that matters.
The root of that class is a static mesh component, and that's quite handy.
However. I need the skeletal mesh variant of that class.
There is a massive overlap between Static and Skeletal in this case like they are 95% identical.

What's the proper way to build such system that accommodates both Static and Skeletal?

[narrow kite]

I tested this not sure

[maiden wadi]

It's not your math here. It's the widget component's widget location.

[lofty rapids]

its off center

[maiden wadi]

Look here. You have the widget component selected. It's centered on the character. But look at where the healthbar is in relation to the center of the widget component.

[lofty rapids]

those arrows should be in the center of that bar

[narrow kite]

You mean this?

Yeah it's not in the center

[dark drum]

In BP you can't as you can't really change the root of a child BP. Whats the actors purpose? The sort of logic in it?

[dark drum]

Replace the canvas panel with an overlay and then you can center it in context to the widget bounds.

[narrow kite]

I just matched the offset right/left and changed the anchor to center

[lofty rapids]

set it 0.5

and 0.0

[dark drum]

As an FYI, you should avoid using canvas panels unless absolutely necessary. They can be heavy and with you intending to have multiple on screen it would be best to just use an overlay.

[lofty rapids]

alignment 0.5

position 0.0

[narrow kite]

Would need to force size it something that way

[ruby cobalt]

it's used to configure physics control plugin with any kind of asset. It can be the child or parent of any other actors of this class and i set up forces to attach them together in multiple ways.

also got a system that let me switch parents on the fly.

and used as a parent for a spring arm class for camera, lights..

spring arm child version only needs the static mesh.. so maybe i can have the physics control logic in another component that is imported in both my static parent class and skeletal parent class?

[lofty rapids]

do you have size to content ?

[dark drum]

Or just replace the canvas panel with an overlay and center the vertical and horizontal alignment. 👀

[lofty rapids]

ftw

[narrow kite]

Yes, but that doesn't change anything. Let me check on Patty's

[lofty rapids]

definately better

the canvas panel i hear is brutal

[dark drum]

Yea I'd look at using a component you can place onto an actor instead. Then you can have two base actors (one for SM or SK) if still needed.

[dark drum]

It just makes me think of the earlier conversation about performance. 😅

[ruby cobalt]

ok makes sense

[maiden wadi]

I'm vaguely curious how hard it is to override the nav generation code in Unreal...

[dark drum]

I believe both static and skeletal mesh component is a child of 'Mesh Component' in case you didn't know. You can try work with just the mesh component and cast to the specific type if needed in the component.

[ruby cobalt]

yes that's how i set it up in a previous version

[dark drum]

I've often thought about this also. Not enough to actually look though. 😆

[ruby cobalt]

i tried to keep most logic in a more general class.. but my static/ske weren't at the root in that setup

[maiden wadi]

We have a flat map. And I feel like nav generation takes about 150x longer than it should. It's starting to be one of those... really strong considerations.

[ruby cobalt]

so i was pretty much never using actor location in the logic.. everything was based on the components.. and that's one more layer that makes things slightly more confusing all the time.. i kind of like how simple things got since i switch to actor based stuff

[dark drum]

I know with some stuff epic has something in the project settings to change the class that's used but there's nothing under the nav settings for it. I'd imagine it would be a C++ job though.

[lofty rapids]

i like the health bar, looks nice

[narrow kite]

Thanks Patty

How is this?

[cloud token]

That's a wild spider 🥰

Aggressive too

Looks good to me 🥰

[narrow kite]

I tested just swapping their base texture with a lot of different assets and made a whole collection of them. It worked suprisingly well.

Bam! And it's now fire

[cloud token]

Damn.. it looks awesome. For real

I would do that bigger too. Like an elite 😄

Its variety but also same spider.
I like

[narrow kite]

I have played around with some variances like that and it does help. It's more notable in a Monstrous spider which was scaled much larger and noticeable in a game world. Their attack / pattern / element may also match. Such as a burn from this one.

[cloud token]

Did you see free on fab today? Got some nice VFX for you i think 😄

[narrow kite]

I will definitely make use of those. This was also released for free https://www.fab.com/listings/d0fe50c4-6ebe-40d5-b78a-56960832f49e

The limited time vfx from first appearance looks about as good as any games on the market. I'll see in testing.

[narrow kite]

If you use the Ashes of C formula that asset is worth roughly $298,000 😉

[cloud token]

I have no idea what that means 😅

[narrow kite]

A reference to the ridiculous $300,000,000+ Ashes of Creation took from various backing, investors, steam release. Kind of like Star Citizen, but worse. It's shut down with a big controversy. They produced very little for those costs.

[cloud token]

Oh.. i have not heard about that.. i will go read about that 😂

[narrow kite]

Will probably be bad for other parties doing kick starter and indie dev funding. But as a story, crashed MMO, unpaid employees, lawsuits everywhere. If you like MMO drama that's it right now.

[faint pasture]

What are the important commonalities to your base class?

[cloud token]

What a disgrace.. this is one of the reason i stop buying games and decided to make my own game instead 😂

[gentle urchin]

If nothing else, it keeps me busy for countless hours!^

[maiden wadi]

Literally countless hours.

[cloud token]

[thin panther]

It's been a good while since I'd heard of a good ol kickstarter mmo scam. Though it doesn't appear this was a scam im surprised anyone still successfully crowdfunds mmos after the long string of rug pulls

That being said its nice to see that the greed of a few doesn't harm the many too much, i hope it remains this case

[narrow kite]

You might want to look a little deeper into it. There are claims of embezzlement, co mingling personal funds, using funds to buy his own house, sell it and keep it all, paying his husband incredible wages with no work conducted. It might be a scam 😉

And just mysteriously taking millions in funding for himself for unknown and not marked reasons lol

10 years of development, 200+ employees and 100's of millions in revenue.......... to be fair though I heard 9 years of that was their dev team trying to optimize the game. #patty 😉

[maiden wadi]

So... Would you say that what remains of the game is the.... Ashes of creation? 😂

[thin panther]

I've heard the claims and admittedly I've not kept up with anything but I've not seen anything conclusive outside of management disagreements. Has anything new come out?

Everyone stopped talking about it all of a sudden so I kinda assume it wrapped up 🤣

Actually idk why I'm surprised it was funded when people are still buying Scam Citizen ships

[narrow kite]

There are detailed interviews with investors showing texts and many receipts. Steven is doing a counter suit, but he has multiple parties suing him. The full proof will probably be more in the outcome of all the law suits. With forced documents from discovery etc. I'm pretty Steven was lying the entire time, but time will tell. Some of those claims may also be very criminal.

[lost wolf]

the developers are the ones getting fucked over completely even more then the backers

some of them spend 10 years of their life trying to make something nice happen

[narrow kite]

No way they got the best deal possible. Years of being over paid and yielding little results.

[lost wolf]

yeah for sure video game developers are extremely overpaid

im not talking about steven embezzling money, i'm talking about the people doing the actual work

[maiden wadi]

The actual developers were likely not overpaid.

[lost wolf]

it was sarcasm btw we are extremely underpaid compared to other sectors

a programmer does not make video games for the money

[maiden wadi]

If someone was stealing money at the higher level, they were keeping it from everyone else. I'd bet the people actually working on it weren't paid shit. According to what I've read they didn't even get their last paycheck.

[lost wolf]

no notice, no nothing

they have families and shit

[narrow kite]

Scars of Honor talked about their funding. It's not in the US. They stated that it would cost multiple times their cost if produced in the US or equivalent. Which is pretty much widely known. Making Iphones in the US for example would cost a lot more.

[lost wolf]

hiring a programmer with 5 years of experience easily costs 200k

[maiden wadi]

Not sure what your correlation here is?

You think because they're US based that they took more money from people than they would have in the EU?

[narrow kite]

Their dev team wasn't producing though. You really have to take that into account. There is no way they that many people where actually doing legit work with those results.

[frozen hedge]

Why still doing this in blueprints not general lol

[lost wolf]

cause baiting

[maiden wadi]

Has there been a blueprint question that wasn't answered?

Also the general chat is for AI talk and Jeff.

[frozen hedge]

No, I’m just saying in general

[tulip anvil]

yes, why are the blueprints not blue

[narrow kite]

They may be correct, but it's actually been lively for people asking question they get 3 responses!

[lost wolf]

sounds like a bug

i'll upvote it if you make a forum post

[maiden wadi]

You can change your editor colors and make them blue. 😄

[narrow kite]

Ahem!!!

[tulip anvil]

tbh i'd probably pay for a plugin that makes the blueprint editor look like actual blueprint in a couple different themes

[lost wolf]

just imagine blue on white blueprints

it would feel like 2032

[pine carbon]

I just want right angle pin connectors.

[lost wolf]

eletric nodes?

there's a plugin for that

[pine carbon]

Please. I hate how if you get too close to a node above or below you, the connector sine wave goes BEHIND the source.

[lost wolf]

curves are ugly

45 degree angles are the only correct way

manhattan grid

[pine carbon]

I'm working in a team, and I need to make sure that however my wires look also look good for everyone else.

[lost wolf]

i think if you include the plugin then everyone should have it

[maiden wadi]

Here you go. Blue blueprints. 😂

[lost wolf]

not the right kind of blue 😭

[pine carbon]

Perhaps more complicated, but I also wish you could shuffle the order of input and output pins on a per-instance basis.

[lost wolf]

this is what we could have

we could look 100% smarter while working

[pine carbon]

Especially bools. The amount of times I have wished I could move a bool output pin up is innumerable.

[narrow kite]

Dragging pins and up down would be nice. Is that possible? Like input pins added to a function?

[pine carbon]

You can do it for the source, but it'll always be the same on every instance of that node.

[steady night]

i got a corrupted charactermesh0 .... any advice :/???

[narrow kite]

Can those be changed without deleting/adding?

[lost wolf]

you can make multiple node instances if you really really really want it

[lost wolf]

the tasks should be white 😭

[pine carbon]

Yes but that is awful programming lol.

[narrow kite]

Those answer the question. It looks terrible and hard to read........that's why Blue prints are not blue

[maiden wadi]

Fair. 😂 I'm not going that deep.

[lost wolf]

https://tenor.com/view/inception-deeper-go-deeper-we-need-to-go-deeper-leonardo-di-caprio-gif-16756828

[narrow kite]

Can you show an example of the issue with the wiring?

In general you can sort out almost anything to at least be presentable fairly quickly.

Unless you're handing this over to your team mates

[pine carbon]

If someone gave me that I would roll back their commit.

Goodness me.

[lost wolf]

i wouldn't even look at it indeed

[narrow kite]

If your BP looks like a broom, you are doing it right.

[ruby cobalt]

oh so here a bunch of conditions set a name and a switch on name trig the right outcome?

[narrow kite]

That's legit though it's a pop up controller. That's every letter of the alphabet causing that

[lofty rapids]

one giant switch for mankind

[lost wolf]

yandere dev flashbacks

high quality spaghetti

[maiden wadi]

That looks like it could be a map.

[lost wolf]

imagine having go through all the paint of connecting those 1 by 1

[narrow kite]

You DO NOT want to see the chinese version of that

[faint pasture]

that certainly looks like a smell

[pine carbon]

What could you possibly be switching on to make this necessary?

[narrow kite]

Is there a way to update all of these effects at once? (Each has it's own settings). It runs one time, but to change it for multiple could they all be altered at once? Like the matrix option for other assets?

[narrow kite]

In all the online training/examples that's actually how it's provided.

It's separating out every digit/character for pop up text

[pine carbon]

And what do these lead to?

[narrow kite]

The corresponding letter

[pine carbon]

Yeah Authaer was right, this could almost certainly be a map (dictionary).

That would be massively more maintainable if you ever need to do this anywhere else.

[narrow kite]

There are only 2 and the other is much shorter 0-9

[ruby cobalt]

isn't it better to check for names instead of stirng?

[narrow kite]

Not sure what you mean. It's taking the string (word) and converting it to an array. So each index/letter can be sent through "the broom" individually.

[white parrot]

String is key, texture val

[dark drum]

Yea, definitely a map where each letter is the key and a texture as the value.

[white parrot]

Regardless of the ugly-ness of what a this blueprint is, it's way less efficient

[lofty rapids]

map ftw

map is O(1)

[faint pasture]

Char to int, int to offset in a texture atlas

[narrow kite]

You mean this essentially?

[dark drum]

Pretty much but promote to a variable instead of using the make map node.

[steady night]

anyone have advice on how to fix corrupted chractermesh0 >..>??

[dark drum]

I've only had it happen a long time ago. I had to reparent the BP to something like actor, save/compile and then reparent back to character and fix everything that broke.

Thankfully I didn't have much in it.

[steady night]

aye.. tried that dosent help :/

cant even locate whats causing it.......

[lofty rapids]

save the project

is it corrupt ?

[steady night]

the project ?

[lofty rapids]

the bp

[steady night]

yeah

[lofty rapids]

make a copy see if its fkd

[dark drum]

Check the saves folder to see if there's an auto save of the BP before it broke and replace the corrupted BP file with it. (Via Windows explorer with UE closed)

[steady night]

hm

i mean in the autosaves, problem is i dont know wich bp broke since its a child on my character

mainchar-enemychar-playerchar-unitchar all of them broken

[lofty rapids]

what exactly is broke ?

is it the details not showing up and the mesh disappeared ?

[steady night]

the mesh on units is fucked up

and i remeber if i have this (0) = Corrupted

[lofty rapids]

wdym "fkd up" ?

[narrow kite]

Would it be a good idea to select all meshes in a level and merge them into a single asset before a build? (So it's a single draw call)

[lofty rapids]

so the bp is fine

in the details

where you have mesh

click on the folder

and double click the mesh

is it fkd in there ?

[steady night]

no

[lofty rapids]

do they all use the same skeleton ?

[steady night]

many of them yes, uin this chase 4x bps use the same skeleton

case

[lofty rapids]

and do more use the same skeleton ?

or it only these ?

[steady night]

only these i think..

[lofty rapids]

check the skeleton

the skeletal mesh

[steady night]

its fine

[lofty rapids]

make a copy of the mesh

select the copy instead

[steady night]

hmmmmmm

[lofty rapids]

that might be an issue

[steady night]

should i try and reimport the skeleton ?

[lofty rapids]

back up everything

but sure, try whatever you think will work lol

[steady night]

yeah ill copy the projject and then try

[lofty rapids]

if its missing, i mean importing it makes sense

[steady night]

yeah but it has been working for months then suddenly it just breakes...

typicall ue

[narrow kite]

The broom is busted. It may be more efficient, but it's not the attraction it once was.

I think it should work I'll test it out and verify a little later

[thin panther]

lil late but it won't necessarily be a single draw call You get 1 draw call per material. You'd also lose out on benefits such as occlusion culling, so I'd opt for no. You'd probably have a much better time replacing a lot of common static scenery with ISMs or HISMs imo

actually depends what asset you're merging them into. If you're fusing the geo into one mega mesh than i believe what i said applies, but if you merge it into 1 mega bp, then you get no benefit at all

[maiden wadi]

ISMs and HISMs have the same issue as far as culling. And there's no longer a big need to make things ISM/HISM for draw calls with the way Nanite works.

[thin panther]

Ooh do they? That's good to know. And fair point about nanite, but that depends on if you're using it :P

[maiden wadi]

True. 😄 But yeah. It's no longer one call per primitive per material. It's one call per material on anything using it. So you can have three different static mesh components using three different mesh assets all using the same material and you get a single call for it.

HISMs are actually pointless using Nanite. Only useful if you plan on backward supporting allowing Nanite to turn off.

[thin panther]

That's pretty near. So if your scene was comprised of 1000 assets using a single material, it's a single draw call?

[maiden wadi]

ISMs are still good, and do have a very minor renderthread benefit. But they're largely about memory and CPU management now.

[thin panther]

Understandable

[maiden wadi]

Yes. But as most decent artists would point out, not all draw calls are equal. 😄

[thin panther]

True 🤣

[scenic shuttle]

Hi actually I am using unreal engine 5.7 and I have to bake the light and I first setted the sunlight to static as well as sky light to static but the problem and also allow static lighting in projct settings and when I go build lighting only the build lighting just pops up for a second and anything didn't happened

[narrow kite]

Thanks. I probably won't bother with merging often unless it's for other purposes,

[last peak]

Turn off Lumen

[narrow kite]

Tile Test VS

Left - New | Right - Old

[gentle urchin]

now just get rid of the repetition 😄

[narrow kite]

It's awful without the tiling seems kind of necessary. You mean the noticable repeat pattern?

[gentle urchin]

I 'm personally not a fan of repeptition at all

but dont fret, even massive titles do it

[narrow kite]

The entire dungeon structure is walls like a classic dungeon crawler. I'm not sure you can get away without doing some versions of tiling. Especially when tile sizes are modified to fit. If it was a single larger mesh it may be designed for something different.

[gentle urchin]

https://medium.com/@jasonbooth_86226/stochastic-texturing-3c2e58d76a14

[narrow kite]

Depending on the circumstance they have to. It's taking a small texture and trying to spread it out over a large surface.

The materials in fab are designed to be seamless, but that still doesn't mean they will match the size of your mesh. If I don't use tiling at all with that texture it looks horrible. There is a lot of variance though in that density setting.

[last peak]

Hmm, you can add some variation and noise to textures so they are not that repetetive

With a bit of extra work

A second layer of textures could be a somewhat dirty quick fix

[frosty heron]

texture bombing?

[steady night]

Hey i tried replacing the skeleton with another and copying that one also made a black screen :/

[steady night]

or acctualy i noticed when copying a skeleton its blank for all skeletal meshes

[lofty rapids]

hmm idk it could be corruption which is a bummer

[frosty heron]

Show?

Doing anything on construction script? Wdym by black screen.

[steady night]

https://gyazo.com/11484ce368ec78c50d8b29de4b7902de

[frosty heron]

Try assigning the preview mesh

[steady night]

https://gyazo.com/87466ddf01effbbb96a551da31335051

all empty

so somthings wrong

[steady night]

asd

[frosty heron]

Try copying the skeletal mesh asset instead of the skeleton

[steady night]

yeah that works fine

[narrow pulsar]

How would you guys recommend doing in terms of animations for a dual wield system? What I mean is like, you have an idle animation for arm positioning but what about the grips and stuff for each different weapon, how would that work, would you need a different animation for every single combo or what

[lost wolf]

yeah that is always the case

[frosty heron]

Yea i do go for different combo anim montages and different state for the idles / starfing / walk / run.

[lost wolf]

you can have arm only animations to blend on top of something else if you want to be more fancy but it will be pain

[narrow pulsar]

So if I have 10 weapons and you can switch them around to use in left or right hand i'll need... 10x10 = 100 different idle animations?

[lost wolf]

no

you can't hold the same weapon in 2 hands

and you can sometimes just get away by mirroring an animation

[lofty rapids]

do you want different idles ?

[lofty rapids]

the animation not necessarily include the weapon is the thing

[narrow pulsar]

No it's just the grips on the weapons that change

[lost wolf]

different weapons dont mean differnent animations

you can have weapon types

[lofty rapids]

thats an offset

[lost wolf]

like large, medium, small

[lofty rapids]

oh hmm that might be tough

i would use the same anim, see if fits

you could alter the hand as well, but idk much about that

[lost wolf]

even demon souls doesn't have unique weapon animations per weapon

dark souls

elden ring

[ruby cobalt]

hey here's what I used to have that worked but had a major flaw: my mesh was child of scene component.

[narrow pulsar]

Like for example here's the problem I have:
I have 2H Weapons, I have 1H weapons, I have Shields.

I want to be able to use 1 H Weapons in each hand, I also want to be able to use a shield with ANY 1h weapon, etc... I don't want to have to create an animation for every single potential weapon combo to be used (IE: Shield + rapier, shield + arming sword, etc)

I just want 1 idle pose for all of htem (obviously, 2h idle will be different from duel wield pose though) but the grip themselves change and MAYBE the hand positioning

[ruby cobalt]

and this is my current system, which makes more sense but only works for static mesh

[lost wolf]

you can just play the same idle animation for all of them if you want ?

[lofty rapids]

sounds like you need 3 idles

[narrow pulsar]

grips are different for daggers, rapiers, swords, etc

[lost wolf]

grips? like the hand?

[narrow pulsar]

yea

[lost wolf]

then make that different

by blending animations

[frosty heron]

Well animation not gonna appear out of nowhere. If you want different anim then you need the assets.

[lost wolf]

you can have a hand only animation on top of an idle animation

[lofty rapids]

well the one hand no shield, with shield, and two hand i feel like your stuck with 3 idles

or it will look crap

[narrow pulsar]

yea thats what i was thinking

[lofty rapids]

especially blending the anims with the one idle

[lost wolf]

it wont look good for a dagger and longsword

[lofty rapids]

probably skippy af

[narrow pulsar]

okay yea thats the solution I thought of, kinda like how we usually split lower body from upper body animations can we further split hands only?

[lost wolf]

yes

you can do anything you want

[frosty heron]

Blend by bone

[lost wolf]

but i would recommend keeping it as simple as possible

because ultimate freedom also means you can shoot yourself in the foot pretty easily

animations are already complicated enough

if they don't work you gotta be able to debug them

and that can be a lot of fun

especially with complex blending

[frosty heron]

I dunnoe, my animations r pretty simple. It depends how much complez you want it to be.

[narrow pulsar]

i want it to be as simple as possible, it's just a pain to have to change them all the time for each new combo

[lofty rapids]

i'm considering two hand in my game and i realize i need a whole new set

[lost wolf]

ours were not https://youtu.be/1H-0x4xk2Xk?t=490

[frosty heron]

I see the release date in the video.

[lost wolf]

i quit 😂

if they actually make it i will be very suprised cause while i was there the only thing that was consistent is that deadlines were never done in time

[narrow pulsar]

So what you guys suggest is just keep 1 idle animation per stance (dual wield, unarmed, 2h weapon) and then blend per bone for the hand itself (for the grip)

[frosty heron]

I suggest just understanding how state machine, blend per bone or any of the tool you need works.

[lost wolf]

you have to evaluate the results and decide for yourself what looks acceptable, what needs more work, and how much more work you can afford

[frosty heron]

Anination is tricky but once clicked, you can kinda implement what you imagine.

[lost wolf]

yeah you can take it as far as your imagination allows

[frosty heron]

Like i stop looking at tutorials once i know wat tool to utilize.

[narrow pulsar]

I'll look into it

I assume an IK could also work for the grip

although would need different IK for each weapon

[lost wolf]

if you IK towards the weapon how is the weapon going to know where its supposed to be

you could do a fancy system where you have weapons that move in a certain way and then your character that tries to follow them but that would be more on the janky / experimental then fancy side

you can go more procedural but there is a reason why nobody has managed to do that yet its hard

[ruby cobalt]

@dark drum i believe this is what you suggested?
@faint pasture

[faint pasture]

If this is for weapons and such just make the root component a small sphere or something

then you can just add a static or skeletal mesh and your collider will be the sphere + mesh or sphere + physicsasset

[ruby cobalt]

i want everything to be physics. so i got lights and camera as child of springarm here.
and use static/skeletal for my main subjects.
I'm making physics driven visuals: https://www.instagram.com/p/DTLldD3EbpR/

baronlanteigne

here, memory is not recalled but operated.

bi directional midi and osc communication between #unrealengine #touchdesigner and #abletonlive to create a virtual audio visual playground.

plugin available on GitHub, looking for beta testers.

Likes

15038

Instagram

uisng skeletal meshes for their jiggle bone possibilities it's so much fun

[faint pasture]

You can still have it all be physics

It'll work, I've done it

Base:
SphereCollider

Sword:
SphereCollider
SwordStaticMesh

Flail:
SphereCollider
FlailSkeletalMesh

[ruby cobalt]

so you apply physics on small sphere. set "dumb" child asset to attach to the sphere?

yeah yeah

[faint pasture]

attachment with welding makes the 1 physics particles collider BE the sphere + mesh collision

sphere can be tiny and inside the mesh

it's just there so the base has 1 component to turn physics on and off for

[ruby cobalt]

and it let me use the collisions of the bigger item?

i use mouse to grab items.. it would move the child relative to tiny sphere instead of moving the sphere no?

so i'd do like: on grab, move the sphere instead

that's why i felt a relief when i switch to using the mesh as root

it makes everything simpler later on for my tiny human brian

[faint pasture]

green is root

You can have a static mesh as root, but then you're kinda up a creek when you really want a skeletal

[ruby cobalt]

yeah exactly. that's the situation i'm in now.

[faint pasture]

This is what I'd do. That way an actor with no mesh at all can still sim physics etc

it'll at least have a sphere

[ruby cobalt]

but i love how clean it is.. so the solution i'd have is to import actor component to each of my parent actors (static, skele)

[faint pasture]

Also if you ever wanted an item with no mesh at all (just particle effects or whatever) it'd work

[ruby cobalt]

yeah that's my current setup for camera on spring arm.. i got that same static mesh root.. hidden

i'll give it a shot.. since my current code based on static mesh can be used for the sphere.. and i can try adding another static/skeletal mesh on top

worried that it will break soemthing else later on since the actual physicsControl parent won't be the objects but the controllers

and using actorComponents added to BP_actors wouldn't be a good idea?

[faint pasture]

That will probably break if you don't actually have a mesh set on the mesh component

that's why I said a sphere collider, it's invisible and always has a collision

[ruby cobalt]

ah yeah i had that issue. won't work unless a mesh is set. but I have a pivot mesh i use for that

a gnonom

[faint pasture]

Whatever you wanna do, I'd use a sphere or box tho

[ruby cobalt]

ah it makes a big difference?

that might be a deal breaker.. my thing is mostly aesthetics.. i need the physics behavior to match the actual mesh i see

[faint pasture]

That's totally doable

[ruby cobalt]

i've set up part of it.. but it seems tricker to set up with PhysicsControl plugin

[honest sundial]

Hi Guys, Im trying to make an inspection system from a tutorial i have found on youtube, However i have run into this issue where upon starting the game, The ui for the inspection is there and my inputs are disabled. Im not exactly sure where this is happening or whats causing the problem. could anyone help please?

[lofty rapids]

for one thing i don't know if youw ant to to the disable and enable on the loop body ?

[lofty rapids]

you destory self at the end there, not sure you wanted to do that, is self the player or controller ?

or is the input on the actual object your trying to interact with ?

[pine carbon]

I need some general advice. I'm trying to make an input prompt scroll down from the top of the screen like in DDR or FnF. My first attempt was using a UMG animation on Widget anchors. I was told this was poor practice since it won't cooperate with pausing, and it makes it difficult to have multiple prompts scrolling down since they'd need to either all be preloaded at initialization, or overlaid on top of one another resulting in dozens of widgets instances. So I've been trying to do widget components where I control the translation offset with a lerp on game tick. Which does work and is easier to have multiples of, but it is much more difficult to fine-tune the behavior. I would really like a way to do this that can implement curves so I can refine the motion. Any ideas?

[lofty rapids]

idk DDR or FnF, but you want the input prompt to scroll down where you add input ? and you want multiple ? is there a video of this or something

if you want multiple stacked on top of each other

then the animation way sounds difficult

instead of interp, you could use a timeline and a float curve as well

[pine carbon]

You've never seen ddr?

[lofty rapids]

i don't know what that is

if you want it to curve there is a curve class

[pine carbon]

If you're European I think it's called Dancing Stage. Popular Rhythm arcade game.

[lofty rapids]

actually if you put them in a vertical box it would be easy i think

[pine carbon]

Nono I mean animation curves.

[lofty rapids]

oh

you want these things to float at you

oh ok

dance dance revolution

i see lol

[pine carbon]

The big issue is that I need it to scroll relative to 2 constraints, which is why I used anchors. I don't think there's a way to alter curve keyframes at runtime so the only solution I was able to find was altering the bottom padding and animating on anchors. But altering bottom padding at runtime is also really annoying for reasons I haven't completely diagnosed yet.

[lofty rapids]

so you want the things to float at you ?

if i'm getting this correctly

let me check ddr,

i have not seen in so long lol

[pine carbon]

I've done some fairly low-level graphics programming in the past, and almost everything screen-related I work with uses 0.0-1.0 or -1.0-1.0 screen-relative coordinates. UMG does not support this anywhere I found except anchors.

[lofty rapids]

you can do some maths to get that range

[pine carbon]

Yeah but I can't do it for curves.

[lofty rapids]

so whats up with curves ?

you want the inputs to curve ?

[pine carbon]

One sec.

[lofty rapids]

theres a curve class like i mentioned

[lofty rapids]

honestly i would probably each one gets a timeline

[pine carbon]

These frickers.

[lofty rapids]

ok right your using sequencer

[pine carbon]

These are great. It's very simple right now because I'm testing for viability but I want these features.

Honestly in general if curves could be created at runtime, that would be an incredible feature for Unreal to have.

[lofty rapids]

ya i mean i would do like like spawn each one seperate, running on there own timelines, and adding wahtever curve you need

or using whatever curve you want

set the timeline to 0-1

and lerp based on a curve probably

something like this

but to have them go straight is really easy

you just spawn it, give it a destination

and interp

i'll see what i can wip up that is like ddr and i'll show you

idk why you want curves tbh, it looks straight ?

[pine carbon]

^

[lofty rapids]

you want a curve i see, so what is the curve going to say ?

how it gets to endpoint ?

like in ddr it go straight up

you want to say have the ability to curve and still get to same spot ?

[pine carbon]

It can do this. It is harder to get it to stop somewhere other than the bottom of the screen. Fiddling with the bottom margin can help. The more significant issue is that because of how this works, the only way I have found to get multiple of these is to layer a bunch of them on top of one another. Which isn't great for performance I imagine.

[lofty rapids]

i would probably spawn these as spereate widgets and handle it manually

instead of the animations

[maiden wadi]

This is just another case where you separate UI from Gameplay.

You need data set up for your beats. Your input should try to trigger them within a timeframe set on the beat or whatever.

The UI or even niagara or whatever you use to display this shouldn't do anything but read the data and move some widgets or particles to display this data within a timeframe.

[lofty rapids]

idk what would be more performant just spawning seperate wigets, or widgets inside the main widget

[pine carbon]

There's already a robust system for actual scoring and sound and all that. This is exclusively visual information and will be powered by the underlying scoring system.

[maiden wadi]

Depends on what you need. If you want to get super performant in UI space, don't even spawn widgets. Make one UserWidget and draw brushes in it's Draw function.

[lofty rapids]

how do you already have the scoring down ?

[pine carbon]

This is new to me. Could you elaborate on that?

[pine carbon]

Several weeks of work with the Harmonix plugin and a whole lot of dispatchers.

[lofty rapids]

but where do you score points

like when ?

is it just the core system, but you didn't test it ?

hmm

[pine carbon]

No it's fully implemented.

[maiden wadi]

Make a UserWidget and override it's OnPaint. Then you can draw things in it. Bit easier in C++ IMO

[pine carbon]

I am attempting to spice it up. Right now the single current correct input is simply displayed above the player's heads.

[lofty rapids]

if i ever finishing compiling shaders i'll take a look at this lol

5.7 do be much faster tho

i have a fealing you can just spawn some widget and then interp

[lofty rapids]

i noticed ddr go up, you are going down ?

[pine carbon]

If I can do one I can do the other. But yes.

[lofty rapids]

so are you familiar with blueprints ?

[pine carbon]

Quite.

[lofty rapids]

create a widget that is your arrow

or whatever it is

then create it in bp

promote to variable

[pine carbon]

Already have that, the video I showed earlier is using a slot as a placeholder.

[lofty rapids]

then set its location

then i would go two ways with it

one is timelines per item

the other is interp an array

i would make a function that do that

or event for timeline

hmm, i can see ddr goes up

i'll go down like you did for the sake of making samesy

[pine carbon]

I think I have an approach that'll work if I can figure out how to get the freaking screen space coordinates. The socket location value is correct but Project To Screen returns (0, 0). Presumably because the player controller is invalid for some reason? I don't understand what's up with THAT, though.

There IS a controller attached to that player.

[lofty rapids]

i'm still not sure what you need a curv fur

so what i did was spawn them

then run the animation

but you need them to hit the exact point where is above teh character right ?

https://streamable.com/rkar4r

[pine carbon]

So . . . I found what's wrong I think? What I can't figure out is why. Calling this from a later event does have Player 1 working but I can't find what's changing.

[lofty rapids]

i see its multiplayer hmm

[pine carbon]

Is there something about the first player controller that isn't properly initialized during BeginPlay?

[lofty rapids]

why are you using the done, whats the for players for ?

it looks like a useless loop

pull from for players

and valid them there

or is the array what you need them in ?

[pine carbon]

So that I can do exactly what you see above, use print statements to test things.

[lofty rapids]

i'm confused because of this

[lofty rapids]

well your just using the done

[pine carbon]

In that screenshot.

[lofty rapids]

your not even using the loop

[pine carbon]

I know.

It currently looks like this.

That loop is entirely there for debugging purposes. What I am trying to debug is Player 1's controller being none during setup.

[lofty rapids]

👀

get rid of the first loop first of all

[pine carbon]

I also have a copy of that print loop in an event that fires on every music bar, and in there, player 1 has their controller. Which makes sense considering that I can control them.

[pine carbon]

There are 2 because I was checking if for some reason specifically the first access of player 1 was none. Which is incorrect, it's still none in the second loop.

[lofty rapids]

ya no just loop once

no need for two

[pine carbon]

Dude I know what I'm doing with the loops.

[lofty rapids]

🤷‍♂️

[pine carbon]

There were two for a reason, there will go back to being one if I need it to be one, they will all be removed when everything works. None of this code is in its final position. All of this is testing things to figure out what works.

What I need to figure out is why Player 1, and ONLY Player 1, has a None controller in BeginPlay.

[lofty rapids]

ok good luck

[pine carbon]

I'm sorry I'm not trying to be hostile. I appreciate the attempt to help. But I promise I know what I'm doing with the loops unless they could somehow be causing the None themselves.

[lofty rapids]

no your not being hostile, your just redundant loop that literally have no purpose

but you must have a reason lol

[pine carbon]

^^^

[lofty rapids]

its your code not mine do whatever you want

[pine carbon]

I was eliminating a possibility.

[lost wolf]

have you eliminated all posibilities

[pine carbon]

I don't know what all possibilities are, but generally if I'm bothering to ask here, I've eliminated all possibilities I've thought of myself.

I also put a fat delay in front of it so it doesn't seem to be a time thing.

[gentle urchin]

Being valid at partx of exec but invalid at partx+1 with no modification would be troubling, but can happen if you multithread

[lofty rapids]

still makes zero sense

[gentle urchin]

Which bp users normally dont do

[lost wolf]

there are infinite possibilities

[lofty rapids]

so first thing get rid of all the code, just loop and print the players

one loop

print the names

[pine carbon]

All players are valid, already checked.

[lost wolf]

one people

one destiny

[lofty rapids]

one code

[pine carbon]

A bunch of stuff is run on them before this point in the code is reached to attach other widgets and sync up rhythm events.

[lofty rapids]

so even one is valid ?

they are all valid

?

or one is not valid ?

also your pulling from an array

and getting zero

your not using anything with the loop

[gentle urchin]

Thats part of troubleshootingn

Sometimes you ad crap just to see if it exist

[lost wolf]

delete everything until it works

[gentle urchin]

Reducio

[lost wolf]

its a valid tactic if you really really dont know what to do

[lofty rapids]

well when you show that code

[lost wolf]

then if you have a minimal example people will be able to help you a lot better

[lofty rapids]

and thats all i can go by

i only know what you show

don't show random shit you tried that didn't work

try to show what your actually trying that isn't working

because this code is like non sense

[lost wolf]

it is rather violent spaghetti

server or at the very least this channel name should have some italian references

🤌 blueprint

[pine carbon]

This is the majority of my Unreal debugging process. I usually try to trim that when I screenshot the stuff that's relevant, I apologize for not doing that here.

[lofty rapids]

it looks like your just throwing spaghetti at the wall

hoping it sticks

[pine carbon]

Yes that's correct.

[lost wolf]

he's cooking

if it sticks just right then its done

[pine carbon]

Anyhoo, here's some more info. From the first event tick, the Player 1 controller is valid. But it is not valid in BeginPlay even after 2 attempts and a 1 second delay.

[lost wolf]

so it never becomes valid

but it is connected

is it valid anywhere else

[pine carbon]

I am changing so many things that it would be unproductive to make it clean right now. I promise I do clean up my code when it works.

[lofty rapids]

[lofty rapids]

do whatever you want i'm not a judge lol

[pine carbon]

Hm?

It does become valid in the Event Ticks.

It's just not valid in BeginPlay it seems.

[lost wolf]

sounds pretty normal for a controller

it will be different in editor and packaged builds as well

just always check null on it

[pine carbon]

The part where I'm checking is also the very last portion of BeginPlay, so it can't be that anything else in BeginPlay is suddenly making it valid somehow.

[lofty rapids]

you can... its a bit hacky

but delay and check again

untill its valid

[lost wolf]

i would not assume controllers are connected in beginplay

[pine carbon]

Why would it be valid on the other 3?

[lost wolf]

players can begin playing and then remember they have to connect a controller

[lost wolf]

is this only in editor ?

can you reproduce the same in a packaged build /

[pine carbon]

Yeah, haven't tried this packaged yet.

[lost wolf]

it could be something weird the editor does

[pine carbon]

Although based on what y'all are saying it sounds like this was a fundementally poor approach.

[lost wolf]

or with the way it supports multiple controllers

it's one of those things that typically will break a lot

since people that are working on the engine will never really test specifically if all 4 controllers work in beginplay in editor

[pine carbon]

This is DEFINITELY something we have encountered.

[lost wolf]

if you upgrade versions that could suddenly change since it might just be a bug

gotta make code that handles it properly

[lofty rapids]

is it multiplayer on one machine ? or network ?

[lost wolf]

dont do stuff in begin play, do it in event tick with a bool HasBeenSetup that happens on the first tick that it can find the controller

[pine carbon]

This is what I ultimately need to do. Get the position of the player's head, and convert it to screenspace coordinates. It looks like I need a Player Controller to do that, but if Player Controller access in BeginPlay is inherently unreliable, what would be a better method?

[pine carbon]

Local.

[lost wolf]

i wouldn't rely on player controllers being in the correct position during begin play either

you can have a warm up period where you just let the engine run a bit before doing anything

[lofty rapids]

for one thing like i mentioned your only getting player zero

but are you just testing ?

and nothing is plugged in either, but thats just you showing something ?

if the players in the loop are valid

pull from there

that node works for a reason most likely

it looks like a foreach on your players

but you want player zero specifically ?

[pine carbon]

What I REALLY want is to make it player-agnostic but you can't DO that in Unreal. T'was not developed with the concept of non-splitscreen local multiplayer.

And Player 1 is the one everyone else sees the viewport of.

[lofty rapids]

you probably can actually

but i see what you mean

so you want to get player zero

well idk why your checking the foreach

is that checking the same array ?

that would make sense

#multiplayer might be able to help you more

[pine carbon]

Yeah, it's just a macro.

[lofty rapids]

what happens you delay for 5 seconds

and check if player zero is valid

in the array ?

why not just get controller 0 ?

controller index 0

should be the main window

if 0 is what they all see like you say

right click, type: get controller

like you would in a single player

if you try to get controller: 0 and its not valid thats wild

[pine carbon]

That IS valid.

No idea why that works but you know what, I will take it.

[lofty rapids]

how are you populate your array ?

i think you can just grab the already in build players array

from one of the controllers

or fk i think its game state

one of them lol

might only be for networked idk

[pine carbon]

Array in a parent Game Mode class that I did not make, but which our entire project uses.

This is the first time this has been an issue since the controller works just fine when you're actually . . . playing. I'm very confused, but your suggestion does seem to be working.

[lofty rapids]

so then player zero is valid after some time ?

as far as the array

[pine carbon]

Player zero is always valid following the code that spawns them. It is specifically the player controller for it that isn't.

[lofty rapids]

🤷‍♂️

is the controller valid after some time ?

i would have to look at the code, but if it works then it works lol

[pine carbon]

Yes and no. It is valid after BeginPlay, but not because of time.

Even a 10 second delay in BeginPlay doesn't have the first player's controller return valid.

It will be valid instantly in the Event Tick though.

[narrow kite]

Unreal Engine does not have a single project-wide "default maximum triangle count" setting for all assets
. Triangle counts are typically managed on an asset-by-asset basis using Level of Detail (LOD) settings, the modeling tools within the editor, or via the import/export process with external software.
The default maximum triangle count for individual asset LOD generation is a very high number (e.g., the default for the "Max Triangle Count" property is something like 4,294,967,295), which effectively means the engine uses the full model by default unless you configure LODs.

That means if you drop a single asset into an open world environment you can tank performance and it may be extremely difficult to identify the root cause. I am going through every single mesh in the game because this is actually tanking areas dropping FPS by 50% in many cases. One case was a literal skeleton model that tanked FPS that bad because there is no restriction.

Additionally, the editor started running at max 60 FPS. It was running for hours in the background. Eventually I closed the project and opened it. The results where still bad. I closed it again and closed the Epic launcher. Only after that did it appear to correct the FPS issue.

A better method would be setting a cap vertices count that can be over written. On a per level, or over ride on the mesh for example so you could control on an individual level certain assets that would exceed your personal threshold. The current method is built to inevitably lead to issues which will be difficult to resolve.

This is also more difficult because it's not visible in the asset. You might expect a massive highly detailed structure to cause issues, but it can also come from a small mesh that doesn't look very detailed.

[dark drum]

This feels like a moot point. Triangle budgets have always been a developer responsibility, not something the engine should hard-cap.

Epic shouldn’t be enforcing global tri limits, that’s production discipline. LOD setup, asset validation, and performance budgeting are core parts of game development.

Also, in UE5, Nanite changes the triangle discussion significantly for static meshes. Raw tri count alone isn’t the primary bottleneck anymore. Material complexity, overdraw, shadows, and draw calls are often more impactful.

As for the editor FPS issue, that sounds like a separate resource or configuration issue rather than something caused by the lack of a triangle cap.

I'm curious to know how long you have been using UE for?

[maiden wadi]

Why would you just... arbitrarily cap the vert counts?... Like what would you even do with that? What verts get removed? How do you even begin to justify the amount of stupid complexity that would cost in CPU time to factor?

[narrow kite]

You won't be able to use me for an example of anything. My professional career as a game developer is exactly 0 days. My real career is a different industry. I have used the engine through for maybe 2 years. However, I could have 0 experience with Unreal or even game knowledge and that wouldn't matter.

The basic principle is more simplistic. The generic (software) could prevent this issue, but it does not. It's a just a simple logic flow.

[narrow kite]

You can already do exactly that and Unreal will auto calculate it for you. It's already an engine feature. You just can't force it to update assets or prevent or warn.

You literally do the same thing, it's just on a per asset basis. Just change the max count and update it.

[maiden wadi]

That's nothing to do with the scene though. There is no triangle count cap in a scene.

If you want to cap all of your assets, make a quick editor utility that iterates your own assets and limits them to N triangles.

[narrow kite]

Even a tool if it doesn't exist (I presume it does not) to check assets for a vertices count exceeding (x) amount?

[narrow kite]

That would still go back to the original issue of "who" should be managing it. I could say, go use a tool to downsize all of your textures in bulk. However, I would say that would be more effeciently handled by the engine itself if feasible.

Which it is, and it is already a feature.

[dark drum]

This is pretty foundational material in game art and technical art pipelines. Triangle budgets and LOD strategy have always been developer responsibilities.

[maiden wadi]

I would never downsize my textures in bulk. That's what platform cook settings are for. 😄

[narrow kite]

That's not really the correct ideology on this. Epic's role is also to support game development in many aspects. Including revenue from 3rd parties. It's in their own interest for the developers to succeed. Thus anything they can do to improve the process is a win on both sides.

[dark drum]

@narrow kite it seems like you have little to no understanding of any of the art pipeline.

[maiden wadi]

It is the correct ideology. What you're proposing is authoritarian. You're saying Epic should just force people outright to limit their things. What happens if some movie scene for hollywood needs a stupidly high tri count mesh?

[pine carbon]

Allowing developers to be less responsible with their asset performance is not the solution.

[maiden wadi]

It is the solution. Because as a developer you shouldn't be a fucking moron.

[pine carbon]

The people who already have good pipelines won't need it and the people who need it will only be enabled to use worse practices.

[narrow kite]

Reducing texture in the editor (aside from any production) can save significant storage space, lesser back up storage etc. It's a built in feature also.

[pine carbon]

To be clear, I am agreeing with you.

[dark drum]

I can remember one of my lectures who talked about a prank they played when developing a PS1 game. They stuck a 1 million tri ladybug, really small in a patch of grass.

[pine carbon]

Like . . . a real PS1 game?

[maiden wadi]

Yikes. Poor PS1

[dark drum]

Yea, it was over 15 years ago now. Lol.

[maiden wadi]

Man, we released on the PS4 and that alone was 😬

[pine carbon]

Because those didn't even have depth buffers, I can't imagine what kind of graphics gore that would cause.

[narrow kite]

You realize even if it was a developers mistake or lack of knowledge that still negatively impacts Epic? You know that right?

[pine carbon]

Do you mean the reputation of Unreal games for being poorly optimized?

[dark drum]

No it doesn't. 🤦 That's like blaming a car manufacturer for someone's bad driving.

[cloud token]

Agree

[narrow kite]

Both the reputation and their business model. If that's a case like Metal Gear Epic is probably getting a direct cut of that.

[pine carbon]

I would actually agree with you there, Epic does have a somewhat undeserved reputation for enabling bad games because developers using their FREE engine are not good at optimization.

[pine carbon]

This is true, but Mordor is not wrong that it happens.

[narrow kite]

Ahem. We ALL know nothing is free 😉 You know better

[maiden wadi]

You're GREATLY over estimating Epic's stake in the engine. Unreal is a side gig for them these days. Nothing means anything to Epic outside of Fortnite.

[pine carbon]

I don't think a tri limit would help though.

You can go on YouTube right now and find a lot of people who do not know game dev correctly identifying a trend in poor performance of Unreal games, and then completely misdiagnosing that as an Epic problem.

[pine carbon]

Yes I am paying with my sanity.

[maiden wadi]

Gets even funnier when they lump a Unity or Cryengine game in there. 😂

[narrow kite]

There is another Niag FX pack available for free with 50 effects. It's a little botched on the release, but you can use it now.
-- Requires 5.7
-- You can't search it by name in fab
-- You can't search it by VFX in fab
-- It's under the tutorials

[cloud token]

Just because you have a race car doesn't mean you should drive like one in cities?

I have also never heard of thats the car or manufacturer responsibility that the driver isn't speeding

[pine carbon]

I admittedly am not terribly familiar with Unreal's methods of optimizing performance. I put most of my effort into making scalable code over performant. I should work on that.

[narrow kite]

Learn stat gpu and stat unit

[pine carbon]

I will look into that.

At the end of the day, Unreal is a C++ Engine, and the whole selling point of C++ is that it isn't going to hold your hand.

[narrow kite]

You can use a logic tree to get the right answer on that.
If (Epic makes a % of sales on (X) game)
then (Epic makes less money on failures):

[maiden wadi]

There's nothing wrong with the engine in regards to performance. You are perfectly capable of creating performance friendly games.

Something that the fucktubers gloss over widely is the circumstances.

Raytracing tech is semi new. It's been played with for well over a decade now but engines only started using it in the last few years.

Any card below a GTX 3060 has no hope of running RT tech well.

Shit got expensive, people don't upgrade as much as they used to. (This includes people running consoles not upgrading to new gen but still wanting all the new games with 100% fidelity)

Now you have a lot of people with outdated hardware who complain they can't turn their game up to ultra settings and aren't happy playing on low or medium.

Upscaling DRASTICALLY helps RT tech. Less to trace, but people are stupid about it because they use their settings wrong most times and or delude themselves into thinking there's a fidelity decrease when in most cases there's not.

And at the end of the day, as a developer, you are perfectly capable of clicking a few checkboxes and disabling all RT features in your game and getting old DX11 performance without the benefits of Nanite, Lumen, VTs, and VSMs

[pine carbon]

Yeah basically that.

I do think there's quite a few lazy devs out there who don't put the effort into making lower settings actually look . . . good, but that's once again a user issue. They absolutely COULD.

[cloud token]

A friend was mad because he couldn't run a game using nanite with his gtx 1080

[narrow kite]

It's mostly correct, but with some major flaws in your thinking. Like "just clicking a few boxes". If you search how to optimize Unreal you could spend 100+ hours going over tutorials, various method, and conflicting information. It's not a "tick this" issue.

[pine carbon]

Good lord that's a decade old.

[maiden wadi]

It is if you want to disable RT features. There's literally a few project settings.

[pine carbon]

Congratulations, you have successfully described the experience of learning to use Unreal Engine as a whole.

[cloud token]

Yeah and you need at least 2060 to run nanite

[maiden wadi]

Kinda. The 20xx series were the first with hardware for RT, but they're still REALLY bad at it.

[cloud token]

Supported*

[dark drum]

It's almost like you have to spend time learning how to use something. Crazy.

[maiden wadi]

And you can run RT features on older cards. It's just all done in software and at slower speeds without the hardware to help.

[cloud token]

Yea, but don't expect anything that isn't technically supported

[narrow kite]

You need to expand you're thinking to a more basic logical structure with critical thinking. If an accounting software was released that companies accountants had difficulties using it, you don't just say, all those lazy and incompetent accountants just don't know what they are doing.

[pine carbon]

How does raytracing hardware work btw? I've implemented raytracing on the raw GLSL level before through the actual math approach, and I'm pretty sure any GPU could handle that. It couldn't do it QUICKLY, but it's basically just math when you do it that way. The most you need is a GPU that supports vector operations.

[maiden wadi]

You're simplifying game development by an amazing factor. And also Unreal is not just used for game development.

[dark drum]

How much do you actually know about the art pipeline? Or are you still on the 'just get assets from fab and drop in level stage'?

[pine carbon]

Come to think of it, I have the same question about HDR lighting. That's also something that I see touted as a hardware feature, but I've done that on a code implementation.

[narrow kite]

If Epic allowed the developer to put a threshold on vertices count you can't really claim there is a downside to the dev. If you don't want it set it high or don't use it.

[cloud token]

Profiling

[maiden wadi]

And then people would complain about things in their levels randomly disappearing instead of their FPS being bad.

[dark drum]

Or as a developer you could tell your artists a tri count when they're modelling in blender, Maya or 3DS max. 😉

[pine carbon]

Your claim is that poor performance in Unreal games is detrimental to Epic's reputation, so they should make more performance options to enable worse devs to make performant games. The problem is that you are helping devs who lack knowledge and you aren't really helping them.

You're enabling them to get further mired in issues they have created before they start to become a problem.

The problem will remain and they will not be any more equipped to handle it.

[narrow kite]

Unreal is designed to be used by literally anyone. It's not necessary to make excuses for them. That includes cases where small dev teams turn into something larger. You can't rely on an already established business model and it doesn't matter anyway. It's simply a feature that could exist.

[maiden wadi]

Are you talking about max tris in a scene or on a specific mesh type?

[pine carbon]

And there could also be a feature that deletes half your tris when you import a model. That doesn't mean it should exist.

[narrow kite]

One of the suggestions would be a better way to identify those assets. Epic already has tools to help identify these issues. From some of the ideology I wonder if everyone thinks Epic should even be providing those tools at all 😉

[pine carbon]

I'm very much in favor of Epic improving Unreal's approachability. I have complained many times about how I find their documentation lacking.

[dark drum]

They exist. 🤦 Profiler being one of them. Then there's the asset validator to name a few

[narrow kite]

Epic doesn't even need to provide those tools those terrible devs should just use their own pipeline and not have the issue to begin with!

[pine carbon]

But at the end of the day, it's a big freaking engine. You can do anything you want but you need to figure out how to do it.

That sentiment honestly goes deeper than just Unreal. It's kinda inherently part of programming.

[narrow kite]

You have to be realistic though and under stand basic concepts like building a trap.

[maiden wadi]

@narrow kite Are you talking about max tris in a scene or on a specific mesh type?

[pine carbon]

You can accomplish a great deal with If Statement Hell. Doesn't mean you should.

[narrow kite]

I don't know that the separation really matters. Ultimately like I said if you place a single asset in the open world it can tank the performance and be difficult to resolve.

In fact I think Epic 100% agrees with me......... as they provide tools to help the dev identify and correct this.

[cloud token]

Like driving a Bugatti on the road. You shouldn't floor it just because you can.. it still the drivers responsibility to drive properly

[maiden wadi]

It does matter. Because one is a simple validator. The other is a scene concept that has a lot of factors.

[narrow kite]

It could be part of an asset filter

Hell it might even exist

You all better hope it doesn't exist. That wouldn't look good for your arguements 😉

[pine carbon]

. . . yeah? If you slap a comically large amount of polys in your scene, your game is gonna tank. Not sure why that's a problem.

This is a very strange conversation.

[maiden wadi]

You can already write validation for assets of any kind. We have quite a few in Atre at the moment. Validating that a mesh has less than N triangles would be trivial.

Rewriting the scene to somehow ignore or remove primitives until you have only so many triangles sounds like a disaster.

[narrow kite]

You would be surprised at how little it took to lose 50% of the FPS rate that you couldn't tell by looking at the asset

[pine carbon]

Not entirely sure what you mean.

[narrow kite]

This guy right here caused a 50% fps drop just by himself 😉

[cloud token]

Epics fault

[pine carbon]

If you KNOW that's the issue, you have diagnosed the problem. Get a better model. Or possibly a stronger computer.

[narrow kite]

Epic "could" have had built in options to prevent it 😉

[dark drum]

How many tris does it have?

[narrow kite]

Patty you want to explain things to him.....like how to resolve it with none of those options

[narrow kite]

It has as many as I grant it. I'm not even sure I just limited the amount to a point at which the asset looks fine

[pine carbon]

Did you miss the raytracing discussion? Some computers cannot do some things, end of story. If you're running on a potato, there IS a limit to what you can do no matter what the engine tries.

[dark drum]

Did you make the asset yourself?

[cloud token]

Why? Its the creator decide how detailed it should be. If developer want a super detailed asset then performance will impact..
So developer need to either fix the asset or find something thats suit game performance

[narrow kite]

The issue is probably seen best in Stat Unit. You will see the amount of vertices being drawn. You can simply limit the asset or use various methods to do so.

It's not a PC performance or trace issue

[narrow kite]

I did not

[narrow kite]

The tools already exist from Epic. I can put a 3M vertices asset into a scene and just lower the detail to an exact number. That's all Epic/Unreal doing that. And for a best example without Nanite because that changes things.

That's also not really true in many other aspects. It's the same thing as releasing a 4K texture. If you look at fab for example you can get 4K 2K 1K etc.

[dark drum]

https://tenor.com/view/michael-scott-the-office-mood-me-face-gif-7842240

[cloud token]

But how is it epics responsibility that the developer use appropriate assets for there game?

[dark drum]

Or film production...

[maiden wadi]

I'm getting really confused. I thought the point was that Epic didn't give you the tools to do this?

[cloud token]

No movies should be 1024 texture and lowpoly

[narrow kite]

Epic's role is to provide a framework for developers to create games. Which includes aspects such as tools for them to optimize correctly and trouble shoot issues.

[dark drum]

I think he's just upset he can't just use random assets from FAB and he actually needs to put some thought into what he's using.

[narrow kite]

Even if it was their own devs doing their own projects within the company. It still applies.

[maiden wadi]

Man. We've redone the pine tree in Atre like four times because that thing is a performance eating asshole.

[cloud token]

There is?
Remember you can make cinematic in unreal too

[narrow kite]

That's right, but it doesn't change the statement. Nor the concepts

[pine carbon]

Just checking, but you DO know people make meme assets specifically to tank games, right?

[maiden wadi]

But they do have tools to help you optimize. Renderdoc, Insights, stat displays, overlays.

[dark drum]

I can imagine. Trees can be difficult anyway. Either high tris count or over draw nightmare.

[pine carbon]

There are modelers out there who will publish horrifically optimized models for the sole purpose of destroying the computers of unwary developers. It is possible you are using one.

[lofty rapids]

this should be on mouse move of the entire widget right ?

[cloud token]

I didn't know that. I thought they just wanted it to look as good as possible. Not care of performance

[maiden wadi]

Should be. It might only apply to hit testable parts, but I dont think so. Think it's geometry.

[narrow kite]

I wouldn't count on the "intentional sabotage". I did hear that on the Internet though so it's probably true. In the larger scope outside fab yes many artists are rendering high quality, but that is correct. You can down scale much easier than upscale. It also depends on the purpose. Some people want a photo realistic environment with max quality while other are going low poly.

[cloud token]

Yea

[lofty rapids]

for some reason its not applying to the entire widget only a selective part

https://streamable.com/8907j9

hmm, i need mousemove over the whole widget i suppose i could bind to it ?

but the override i thought would work

[maiden wadi]

Can you test by setting the visibility of that white block to HitTestInvisible?

From the look of it if that is the cause, you can just put a blank image behind the whole thing with DrawAs = None and make it Visible.

Out of curiosity, what is this for exactly?

Might be a different way if you're looking for some sort of mouse location effect?

[lofty rapids]

the white bar is going to be a thing i can drag which will adjust the top and bottom

[lofty rapids]

its an IDE

[lofty rapids]

ok this almost worked, the only problem is i need the stuff in the top and bottom to be clickable and recive mouse

and for some reason they are not triggering mouseover

they are web widgets

so what i figured was, click on the bar move up and down and let go

would resize teh top and bottom

basically this wight bar is going to slide up and down, which will adjust the heights of the top and bottom sections

but the fkn mousemove isn't firing over this top widget for some reason

so the fix suggested worked for the rest, except the two widgets, if i put them on the same zorder they mouse triggers obvi but i can't click the thing and type, if i drop the drawas none image lower it doesn't trigger

🤔

the top and bottom are web widgets

the browser doesn't seem to have a mousemove event i can bind to

as a temporary fix i can make it invisible while i resize but i would like a way to fix that if possible

[lofty rapids]

https://streamable.com/4un3n9

see if i set the web widget to collapsed it mousemove.
when i resize and keep visible it will give a little wiggle room but the mousemove doesn't fire over a web widget ?

its part of a thing where i can script for my games, and also make different programs

[warped juniper]

This is a fun error...
For some reason, this capsule collision I slapped into a socket of a character, default sizes, default scales for everything, looks MASSIVE.

[lofty rapids]

i fixed it with setting the zorder and using the image in the background thank you

[maiden wadi]

Check it's attach parent's scales.

[warped juniper]

Already said it's all at 1.0

[maiden wadi]

You said it's scale was default.

Ah, for everything.

What is it attached to?

Can you scale a socket on a skeletal mesh? Haven't tested that. Could be inheriting the Socket's scale.

[warped juniper]

Hmmm

Gonna check

@maiden wadi Yeah I know what's the issue now lol

[ruby cobalt]

hmm i got a bug but i think i did something dumb.

can you run Get Actor of Class at construction script to update a variable with an actor in the level?

haha they all show None so i ugess you can't

hmm

[warped juniper]

@maiden wadi Okay the issue is that the skeletal mesh I'm using got exported with base scale 100...

So now I need to see how to re export without it

[tall oxide]

Is there a way to "flush all keys" in UE4?

[maiden wadi]

Yes, but I don't think it's blueprint exposed.

[tall oxide]

Yeah Its not, Only in UE5

I was thinking an alternative for that

[livid flare]

Hi any method to make the on component hit event fire on meshes with physics disabled?

[faint pasture]

Movement components

[livid flare]

Sorry can you elaborate

this is what I have, not sure how to use movement component here

[dark drum]

I think it's just a check box under the collision settings for the mesh.

[livid flare]

I have the generate hit and overlap events activated

in collision settings of the mesh

it only fire with physics activated

[crimson briar]

I think the object itself must use some movement to generate hits without physics, it won't get hits from other physics enabled bodies. But I'm never sure if I understand this stuff correctly

[livid flare]

If I use the begin overlap node instead, it works but looks like it won't work with replications

[whole ridge]

If anyone has time I have a question. I'm hoping to get nudged in the right direction.

What I've done: Top down game, WASD movement as well as the mouse click and hold.

I've got it figured out to switch to first person when I tap "C".

The camera movement is fine but the WASD is wonky. It doesn't seem to line up or react perfect with the camera movement.

[dark drum]

Show how you handle the movement.

[whole ridge]

I will pretty soon! I honestly wasn't expecting a response so quickly, lol.

I have it split between my character blueprint mainly for the camera view change, and my controller blueprint for the movement portions of course. Just to give some context.

I even opened a new project for first person to make sure my IMC wasn't set up incorrect and the inputs. I even mimic'd the functions within the event graph for movement and aim.

I know that's not images just trying to remember as much as I can being away from my PC.

[lofty rapids]

The problem I see is you mimic the setting and the camera is different

Depending on how you move you probably want a slight difference

[floral stump]

seems like the collision channels don't blocking each other

you don't really need the physics to be enabled for hit detect

[faint pasture]

What is doing the moving here?

The thing doing the moving is what handles detecting hits. Either that's physics, a movement component, or just your own custom code

[livid flare]

I'm using vehicle channel for collision. I could try creating another channel

[livid flare]

It's the vehicle movement component, how can I use that

[floral stump]

don't need to make new channel, just double click the vehicle channel and enable the collision block for the other object collision

[livid flare]

it's already all set to block, the hits are with actors from same parent class

[floral stump]

is this advanced vehicle system plugin? and you want the vehicle to hit another vehicle?

[faint pasture]

Show your vehicle component layout

[livid flare]

its a vehicle traffic system using chaos physics, I'm also trying with AVS but I got a different issue with the logic there cause the asset was made for chaos

[floral stump]

I think you are using skeletal mesh setup which needs a physics body for interaction, this could be the reason

[floral stump]

try to open the mesh and show the collision of it, if it has no collision, add a collision body to it

[faint pasture]

All of the components, I need to see what the root of this actor is

[livid flare]

Yes that make sense, I'm gonna give a try thanks

[pine carbon]

Which one of these options will change where the left edge of a widget appears on the screen.

[frosty heron]

Dont draw as box

[whole ridge]

The first three are from the Characters BP.
The last one is inside of the Controller BP.

The pressing of "C" works great, camera works great etc. For both how I have the Top Down set up. So far, the First Person camera looks great. My walking is just a little weird.

[pine carbon]

Would you mind elaborating on that?

[frosty heron]

Question not very clear, do you mean the outline around the element?

[pine carbon]

I mean the left edge of the top level of the widget hierarchy.

Or do they just not work like that.

Maybe I should describe my end goal and see if I'm doing it all wrong.

I got these screen space coordinates. I want the widget to be aligned with the x coordinate.

[dark drum]

Can you share how the Aim and Move functions are setup?

[whole ridge]

Sorry, I recorded this to show the issue better.

[low glade]

Hey Im trying to integrate an fps counter into my game so that players can measure performance, I've seen people divide 1 by delta time to get the fps, is this a reliable method or should I try to use the stat fps command (only reason I haven't used that is i can't customize the font and stuff)?

[lofty rapids]

i see your using the same spring arm for both camera modes ?

[whole ridge]

I honestly didn't think twice about it. When I get back home in a few I'll make a new one and see what that does. Camera wise I haven't seen an issue. Unless my issue is masked by a missing spring arm?

[lofty rapids]

whats the issue here its not working right ? i can't tell what direction your supposed to go vs what your doing lol

[lofty rapids]

if you think about this from what i can tell, your rotation is the direction of the springarm

[lofty rapids]

are you not even using these ?

your adding rotation, this is will keep adding up when that event fires ?

maybe try to set instead ?

[whole ridge]

When I move around it doesn't quite match the area I'm pointing to with the mouse, like normal FPS.

Backwards doesn't work correctly. Thinking about it though, that spring arm may be the issue..

[lofty rapids]

is it fkd in both modes ?

[lofty rapids]

whats the spring arm look like as well ? is it on center ? did you move the camera at all ?

consider the spring arm is not the same as camera rotation

if its not aligned properly

[whole ridge]

Only in first person. Top down is perfectly fine.

Yeah the way they are set up are also good far as I could tell.

That's fair too. I'm antsy to get back on to try this out 🥸

[lofty rapids]

do you even use the spring arm in first person ?

[lofty rapids]

is it first person thats teh problem ? is third ok or they are both fkd ?

i see you do you use aim/move for first person

[whole ridge]

Yeah first person is the issue.
The top down works as intended. I have WASD working for top down or the mouse click and hold.

The first person is almost perfect, just the weird movement issue.

[lofty rapids]

ya what i think your doing is moving based on the player facing direction

but think it doesn't change when you rotate the camera

i would use the camera right/left forward/back similar to your third persona view

so when you rotate the camera, the player isn't rotating

and your moving right/left based on where your facing

not the direction of the camera

[gentle urchin]

Usually thats enough to get it going. 1/deltatime = fps aa you say

[fiery swallow]

it doesn't work for generated frames though :<

[gentle urchin]

Correct 😄

[cloud token]

Which should not be implemented in first place anyway 😡

[narrow kite]

If you have 15 Niag effects in a level. Torches, Fountain etc. Should you just let them all load at run time and always run? Or set up something like a collision box so they are only active when the player is within the collision range. On a performance level which one is better and is it signifcant enough to matter?

[floral stump]

generally don't run any particle effect if no player can see it

if you place a particle effect in a streamable level, set it to patially loaded, it will unload automatically if no player is in range

[narrow kite]

That is where I was going with it. However, the collision boxes are rather large due to the view distance. And technically the player could be loading/unloading (turning on/off effects) by moving in and out of a collision box they can't see.

They are not currently streamable levels. Not sure I want to go that route.

[floral stump]

if you are using unreal 5, this is the main feature for what people upgraded from ue4 to ue5

[narrow kite]

I'm getting closer to a steady 120 fps even in the editor. With naturally some dips. I haven't noticed that much with the niag. It's just a matter of optimizing in cases where they might consume a lot more. There could be 40 in a level for example or more resource consuming effect.

[floral stump]

use streamable level, they are awesome

[maiden wadi]

I'm fairly sure it was one of the mid tier reasons to upgrade to 5.

[narrow kite]

I originally started an open world with the partition system, but it led to a a lot of issues. Don't streaming levels really need to separate out every asset as to what's being loaded or not?

[floral stump]

no, you have an option spatially loaded

if this is true actor will be auto load/unload if in range

if you want something to be always loaded, set it to false

[narrow kite]

Sounds like the world partition system basically on a smaller scale

[floral stump]

thats all

yeah and its very improved, put thousands of objects in level with spatially loaded enabled, and let the engine to handle load/unload

I have like +65k objects in my level

and fps reached 900 in high res, shipping build

[maiden wadi]

Well that was a really random detour. ANYHOW. What you're really after is a NiagaraEffectType. It's essentially a scalability profile for your system so that you don't have to manually do other performance impacting things like listen for overlaps from arbitrary sources and code weird systems to manage your particles. All entirely irrelevant to level management.

[narrow kite]

Thanks. I was looking up information on it. That does appear to be a good solution. If I understand it correctly it would be setting up an environmental type/category then having anything in that spawn at a set distance from player.

[maiden wadi]

I think so? I'm not fully following your last part of "then having anything in that spawn at a set distance from player"

Essentially if you have a torch, or a fountain. These things just have a niagara system on them. That system has a NiagaraEffectType applied to it which culls it's render or tick based on the conditions you allow at set scalability levels.

You then do not need to care about spawning or distance to the character. It comes with the torch actor or fountain actor and just automagically works.

[narrow kite]

It looks like there are several options including a render distance. Which I presume in the game naturally means distance from the player essentially as it's controlling the viewable area. Thus in a sense if you move into range of the setting it would spawn.

[maiden wadi]

Not spawn as much as start ticking/rendering, but yes.

[narrow kite]

Yes more technically activating paticle effects.

[whole ridge]

I narrowed it down the more I just played the game. The First Person camera is replicating the left and right limits of the TopDown Camera rotation. I cannot find where this is on my First Person Camera. It's driving me nuts.

[lofty rapids]

in hear in move you are moving the character relative to itself right ?

this is your first person camera movement

switch out the the right and forward

with camera right and forward

[lofty rapids]

" The First Person camera is replicating the left and right limits of the TopDown Camera rotation." <-- are you saying third person events are firing while you are in first person ?

[whole ridge]

I'm following I am just confused. Probably from being tired. Do I need to make a variable and have the reference the first person camera as an object?

[whole ridge]

I believe so, but only the camera

[lofty rapids]

do you see in the move function wehre you move around ?

[maiden wadi]

Should really have separate IMCs for these control sets so they can for one be rebound easier and separately, and two be simplified.

[whole ridge]

I also did that, I have one for the TopDown and one for the FirstPerson

Thank you though. At first when I did this I didn't realize that right away and had to do some research

[maiden wadi]

Ah. Yeah I just went back up through the conversation.

[whole ridge]

I have something super small just not lining up and I think I am overlooking it somewhere. I've debugging and it doesn't give me much to go off of.

[lofty rapids]

i'm pretty sure its because your moving rleative to your character that your camera movement isn't ligning up

when you rotate the camera the character doesn't rotate

atleast i think so thats what it looked liked to me

like you probably notice you can strafe move forward/backward, but you can't rotate

[maiden wadi]

Ah. You mean your FPS camera can't look up? It's stuck looking downwards?