Видео

Thanks, will do!

Glad it was helpful!

it is the best software that i am aware of. Mixer seems pretty good for the most part though as a free alternative. As far as their TOS goes, it's worth double checking yourself (in case you haven't yet) what subtance's TOS says. there's been a lot of bad stuff coming out around adobe's TOS but as far as i understand that's centered around the creative cloud package mostly. It's entirely possible that substance operates under another TOS. (you can just straight up buy it on steam for instance, which isn't possible with other CC applications, so i imagine there might be more differences.) Now, even if the TOS for substance isn't as batshit insane as the rest of CC, of course it's worth considering if you want to give adobe money or not. If you're working professionally, maybe look into mixer if you want to avoid adobe. If you're working just on personal projects... I mean i would never advocate for anything illegal... but i hear sailing the high seas is rather pleasant this time of year...

might be worth your time looking through my unreal basics course playlist to get a decent grasp on fundamentals first :) ruclips.net/p/PLoReGgpfex3x295NSo3hZD6Ylq6hPqdz8

@@thegamedevcave Thanks

very much depends on what youre trying to save on honestly. As so often with optimization, context is key. there usually isn't a single "make it better" button. but in general distance fields are fairy well in gpu performance yeah. You optimization will often come from limiting draw calls, not so much from optimizing a few nodes and functions inside a single material. (within reason)

player character probably best to be persistent

@@thegamedevcave Ok, ok it's my bad, i forgot check "Make Visible After Load" in BP Level !

Thanks 🔥

yeah they can be a bit of a handful to work with, i still think for a lot of things it's very much worth it to do but only for things that are reoccurring on a project wide basis. If one actor just needs to tag another actor as something which nobody else cares about, just using actor tags makes way more sense as i understand (I dont really use them much myself). And when/if using GAS, that will pretty much make your whole workflow centered around gameplay tags (which can be great, but also frustrating at times)

that's what the child blueprints do, it takes all the code from the parent and also runs it/ has acces to it. so you dont see it in the event graph, but it's still running. Try it, make an actor and hook a print string node up to the event tick, then make a child bluerint of it and add nothing to it's event graph. if you put the child blueprint into your level you'll note that it does a print string every frame. That is because the parent blueprint's code is doing that.

might be because the mobility of the actor is set to static, i remember having some people having an issue with that at some point.

I believe there is a node called "simple IK" which lets you put in the first and last bone of the chain instead, i think that should work with longer chains of bones too for something like a tail :)

@@thegamedevcave Thank you! That is what I was looking for. Though I think I might have found a better solution using the distribute rotation. I think that might behave more tail like

@@Jensonian18 that's actually a really good idea !

You're very welcome!

yeah, dont try to edit anything about the data asset itself while data aseets can be changed on runtime, you should really treat them as assets, not just blueprints, you wouldn't try to change data about a static mesh for instance either, same goes for data assets. instead metadata like that should go into the itemslot. i'm working on a followup series that'll include some metadata stuff like durability. For something like durability speficailly, i'd probably just put that as a variable on the slot itself. and disable it when it hold an item doesn't have durability. For other kinds of metadata i'm working on setting up a ssytem that uses custom Objects that can hold any amount of data of functions you want :)

​@@thegamedevcave I'm looking forward to your video about it. For now I'm trying an approach using a blueprint of type object, which also has the usual inheritance set up (item -> equipment -> weapon). The object has an attribute for the data asset and the equipment has an attribute for durability and the inventory struct has the object instead of the data asset. Feels a bit clunky too though, and I'm not sure if I'll get issues with that approach later. Like I already noticed, but both my approach and probably yours too has an issue with metadata for each item of the stack, as it saves metadata for the full stack and not one item. If you have something like a health potion with multiple uses but also stackable, I'm not sure how it would work with either approach .. besides making them non stackable to begin with, or move the one that does not have all uses left to a new slot and make it unstackable :/

@@Lilium___ i can't think of many games where items with metadata are stackable in an inventory system honestly. Probably for this exact reason. an item slot usually is just a reference to an item with a number telling you how many of that item there are. if each item needs to actually be it's own individual thing, each slot needs to turn into an array instead. But i also dont think you'll want to do that because if each item in a stack is unique, how does a user interact with a specific item in that stack? If items have different behaviour, it's probably better for the user too if they don't stack so they can acces "potion that heals 50 health" speratly from " potion that heals 100 health" . combing those within 1 stack will probably lead to game design issues rather than programming issues. which is why i dont think you see this type of unique item stacking that much in games.

@@thegamedevcave Yeah I think that's what it comes down to in the end. I've seen some games that let you stack weapons with full durability, but then again you can only equip one weapon (and not a stack), so the metadata of the stack will never be updated, only one item at a time, when its equipped. Having an array of metadata, one for each item of the stack can solve that, but as you mentioned that can be very clunky for the user. Technically if you have a cake that can be eaten 6 times, you can make it stackable and just subtract one after every 6th use. But good luck implementing that without making the whole system clunky. Thanks for the input, I'll stay with non-stackable items with metadata, otherwise it will just get too complex for almost no gain.

you're 100% right. i kind of did a shit job explaining what i needed here. what i was looking for wasn't just a waiting node from blackboard time (because yeah, that already exists thankfully) but more or less an async wait node. Something that counts up in the background while the rest of the tree is allowed to keep running and then after X amount of time flips a bool (or in this case actually an enum value) so the tree starts to run the other branches, until the timer rolls over again, at which point it'll switch back :)

I like the music. What's wrong with it?

​​@@svendkorsgaard9599 its a bit too common and overused. For allot of people it may seem lazy and that kinda ruins the vibe of the video. But do whatever you want if you happy with the results.😁

Glad it helped

as you can see in the video itself, higher poly count will make the displacement smoother. yeah it is because of hard edges, but adding more geometry, even if just flat faces (while not great to do, definitely don't' subdivide it like 10x of course), does help solve the issue. I'd recommend you try it yourself, try importing a simple low poly cube, add the material and then add simple subdivisions to it and see the difference. I'm pretty sure it's to do with the vertex normal. when the cube's faces onto have 2 tringles each, the vertex normals point straight up, perpendicular to the face, but when there are more tringles on a flat face like that, the vertices on the edges have their normal pointing out diagonally, meaning that the offset is done diagonally which prevents the taring issue. you of course don't HAVE to make specific components to display outlines but if you're going to have a LOT of meshes in your game that all need outlines anyway, you really dont want to give every single one of them their own overlay material, that's just tedious, much easier to just make a component for it once and have it all done automatically.

@@thegamedevcave Subdividing usually also smooths edges in addition to just adding divisions. The smooth edges are the thing that fixes the issue, not the polycount. For the character I was talking about adding the outline to the skeletal mesh in the content browser, so that anywhere you use it, it would have the outlines by default. To me that's super comfortable

that's the thing though, in the example i give in this video i specifically use linear subdivision, so the actual edges stay sharp... for the character, yeah that is true, but still that's for every character in your game, if you have like 100+ characters in total, setting up the component like this (even more so when it comes to the static meshes), will save a bit of tedious busywork. Of course, in no way required to do it that way, and if you're just adding it to 1 character there's no need to bother with it, but otherwise, why make things more time consuming on yourself?

​@@thegamedevcave I'm not talking about the shape of the model, I'm talking about the edge type. You can experiment with this in a modeling software like Maya. It lets you select edges and determine whether they're soft or sharp. When importing a model, UE interprets hard edges as being disconnected from one another. It's interesting, look into it! For assigning a material on multiple assets you can select them in your content browser, right click > Asset actions > Edit selection in property matrix.

does this work in ue 5.4.2

there are only so many ways to do it, this is the most straightforward way, i dont doubt there are dozens of people that made videos showing this method.. i recorded this video like a month ago anyway, so just nice timing that we both uploaded this around the same time i guess?

@@thegamedevcave This is simple but performance cost and also two extra draw call one for overly material and the other one for internal mesh I used the @codelikeme channel video which is in post process

yeah it uses up a lot of extra drawcalls, so it's not exactly ideal in that regard. depends on a lot of factors but this is the same basic technique that has been used since the PS2 era for a lot of games. so altogether, it's not THAT big a deal i'd say. Post processing is generally probably cheaper to do though, but is not as consistent in the thickness out the lines.

@@thegamedevcave Just because it's one more draw call, doesn't mean the method is inferior to postprocess ones, entirely depends on the project and frame budgets, there's no 'one correct way' of doing something usually. I personally really like this method (since this is exactly the visuals one of my projects were looking for), so thank you for the well-made tutorial, don't doubt the majority of viewers are appreciative of all the diverse solutions provided.

to be fair it's not just one one draw call, it's one more draw call PER mesh, which can add up quickly so it is something to be mindful of, but overall it does end up looking better than post processing in my opinion and while those extra draw calls might end up adding up to form some performance impact, usually saving draw calls in other ways when possible is probably worth it do be able to use this method :) There's up and down sides to each method for any desired results, this one is the same, just gotta be aware of those and see if they fit the usecase! I'm glad you found it helpful for yours!

All this chat gpt stuff has warped the meaning of "ai" a lot... 😅