not a doctor, not a lawyer, examples I use may not be fully researched so don't take out of context plz, don't @ me
+5
Options
KoopahTroopahThe koopas, the troopas.Philadelphia, PARegistered Userregular
I love the constant improvement, production, and availability of Unity. It's my favorite engine. Unreal puts up a real good fight too, but damn I love Unity.
I gotta say that mobile demo is impressive!
I've been working around putting more procedural stuff on those little screens for a while now, and I keep running into, well, not lack of computing power, but Unity's atrocious swapping management. I don't see any of the problems I've run in to here (which mostly show as visual judder. I'd like to see the camera in that demo do a 360° spin though, because so far that seems to be the move that gets the judder issues. And also maybe on an iphone XR?
But I want those pretty SHADERS, wants wants...
How far I got with futuristic cityscapes/orbital habitats: EDIT: mind your sound levels. It's loud. https://www.youtube.com/watch?v=KfcuH7MhlSs
I gotta say that mobile demo is impressive!
I've been working around putting more procedural stuff on those little screens for a while now, and I keep running into, well, not lack of computing power, but Unity's atrocious swapping management. I don't see any of the problems I've run in to here (which mostly show as visual judder. I'd like to see the camera in that demo do a 360° spin though, because so far that seems to be the move that gets the judder issues. And also maybe on an iphone XR?
But I want those pretty SHADERS, wants wants...
How far I got with futuristic cityscapes/orbital habitats: EDIT: mind your sound levels. It's loud. https://www.youtube.com/watch?v=KfcuH7MhlSs
Hey, who made your music?
"Simple, real stupidity beats artificial intelligence every time." -Mustrum Ridcully in Terry Pratchett's Hogfather p. 142 (HarperPrism 1996)
To bring it back to game development. Steam actually checks that the game supports what it claims to support, didn't know that. Very glad they do and the marketplace isn't as lawless as I've heard.
It's been awhile since I posted here and know this helps in no way to anyone but I swear to god either my project is snowflaked monster or no one is testing that Unity actually can build Android. The layers of issues to get an APK to even build is god damn amazing, including:
Passwords for keystores not being properly escaped in generated Gradle files, so certain characters in passwords will just break everything
Gradle android plugin is outdated and doesn't work with the recommended NDK version. Also, it can't be upgrade due to the Gradle version being hardcoded to the Unity version.
class path hell leading to multiple dex issues
Whats next?
I am pretty close to turning to sheet caking at this point,
This has cemented my desire to move away from Unity. Even if I am just being an idiot and missing obvious, Unity's entire set up is so contrary to the software development best practices I practice at my job every day. You want reproducible builds? Instead why don't you tune hundreds of little gui dials and snowflake your builds to the point that no one is ever going to be able to get them to work again.
...
This has cemented my desire to move away from Unity. Even if I am just being an idiot and missing obvious, Unity's entire set up is so contrary to the software development best practices I practice at my job every day. You want reproducible builds? Instead why don't you tune hundreds of little gui dials and snowflake your builds to the point that no one is ever going to be able to get them to work again.
...
This has cemented my desire to move away from Unity. Even if I am just being an idiot and missing obvious, Unity's entire set up is so contrary to the software development best practices I practice at my job every day. You want reproducible builds? Instead why don't you tune hundreds of little gui dials and snowflake your builds to the point that no one is ever going to be able to get them to work again.
Obvious q. Have you tried the internal build?
Yeah that has a whole different set of errors :rotate:
Edit: yeah looks like a bad upgrade happened at some point.
I am hoping that Unity's package manger is eventually extended to 3rd party developers. Having to check full set of lib files into a projects source tree becomes difficult to manage.
So earlier today I put my video audio assistant game into beta today after ~15 months of development (I started two years ago, but there's been a lot of downtime).
I can't mention anything specific in the forums 'cause I need to keep my personal life seperate, but I'm pretty happy with it so far given the small scope I was aiming for.
I've been kind of shaky all day from pushing that button. I kind of think I'm going to need a drink once it goes out to the world at large.
Before I got started, I literally hadn't coded anything more complex than a single batch suite I did a few years back (itself my first batch code!), and never even touched the coding language I was using before... and my history with coding has been less than stellar. My education is in English.
The moral of the story though is you absolutely can make a game no matter how new you are and how little you don't know so far. You gotta start small and work your way up and fail fast and all that jazz, but no matter where you start you can get there if you're willing to put the work into it.
That coding literature could use some work though.
Here is some shit I've been doing lately: randomised faces, and more and more work on my HL2-ish eyes (no eye spheres, just a shader). Reverse order here so older shit as you scroll down.
Are you guys bored yet? I am. Well, not enough to ditch this project and work on the actually nearly finished ones, but it's becoming annoying.
Behold, my water:
What you're looking at is, first of all, an offset problem with the meshes. This is pretty easy to fix!
Then, you'll notice that underneath my 'high detail' landscape mesh is now a mesh that is identical, except it doesn't dive down where there's water. This is the water surface! It's pretty clever even if I say so myself, and I was planning to do the same thing for roads.
However, it might strike you that the mesh used is pretty low poly. I mean, it looks like crap.
So now I have to decide which way to go.
-I've tried 'off the shelf' mesh subdivision and it's not working.
-I could try to simply generate a bigger array and generate detailed meshes of that
-Tiling! Which is probably what I'm going for anyway... But the first thing to do when tiling would be to try and ruin my landscape generator by making it more modular. The goal; to have it set up with a lower vertical resolution so it can take tiles with standard, stepwise slopes. This wouldn't end well! You can imagine the variety of tiles needed. 0-0-0-0 would be fine. But how many tiles with 0-1-5-9 slopes would I need?
So earlier today I put my video audio assistant game into beta today after ~15 months of development (I started two years ago, but there's been a lot of downtime).
I can't mention anything specific in the forums 'cause I need to keep my personal life seperate, but I'm pretty happy with it so far given the small scope I was aiming for.
I've been kind of shaky all day from pushing that button. I kind of think I'm going to need a drink once it goes out to the world at large.
Before I got started, I literally hadn't coded anything more complex than a single batch suite I did a few years back (itself my first batch code!), and never even touched the coding language I was using before... and my history with coding has been less than stellar. My education is in English.
The moral of the story though is you absolutely can make a game no matter how new you are and how little you don't know so far. You gotta start small and work your way up and fail fast and all that jazz, but no matter where you start you can get there if you're willing to put the work into it.
That coding literature could use some work though.
/highfive
English background here too with a similar background, thanks for sharing!
Yesterday I finally flogged myself into doing some dev work (I am also trying to reconfigure myself into thinking of game dev as entertainment) and as I was porting the swept collision detection library I wrote a year or two ago to my component framework I realized that an entire set of overrides was not implemented. So this is going to involve some original work as well. Though thankfully the missing pieces are all readily formed out of the basic bits I already did, e.g. determing if a polygon stops a moving circle is just a matter of finding the earliest collision with any of its composite line segments (leaving aside edge cases like initial interpenetration that I also have existing tests for).
My favorite musical instrument is the air-raid siren.
Trying to get as much done to finish up Space Cat Deliveries and creating more jobs/stories ended taking me 3 weeks instead of the planned 3-5 days. I'm still not done but the hardest, which was setting up the dependencies between storylines, is done. I've also hit the milestone of the todo list being shorter than the done list and most items can be done in a single evening.
I was hoping to ship this game out before Epic's MegaJam starting this Thursday but it's too late now.
Ugh, I'm really struggling with my content creation infrastructure in Unity.
I need to change colliders for each keyframe of a (2D) animation. So I put a monobehaviour on my character that receives input and pulls collider data from a ScriptableObject at the same time as it sets the Animator.Play() state.
The problem is visualizing these colliders to edit them. Right now, I'm storing them as Bounds in a ScriptableObject, and I'm using Gizmos.DrawCube to show them in Scene View. But I can't figure out how to make Scene View show the correct sprite keyframe while I'm editing the boxes.
The closest I've gotten is storing an AnimationClip in the collider data ScriptableObject, but I'm not sure how to tell the monobehaviour to look into the SO, find the AnimationClip, make the animator play that clip, then make the animator change to the correct keyframe corresponding to the keyframe collider data that I'm editing.
Does anyone have a better approach? Should I be storing an Animator SetTrigger string instead of a reference to the clip asset itself? How would I go from trigger string -> ??? -> pulling keyframe data?
So for procjam I did get Unreal level streaming working correctly. I created several levels, each with a single terrain object. Each level corresponded to a certain configuration of walls. Then I used level streaming to load and place the terrain chunks.
It worked OK but the terrain takes up a lot of storage. I guess one cool thing is that I could go back and touch up the terrain with foliage and more textures without changing any code. I wanted to add some "special" terrain tiles like a big pit or a bridge but ran out of time.
KoopahTroopahThe koopas, the troopas.Philadelphia, PARegistered Userregular
That's just general good life advice when creating or working on things. If you're struggling wracking your brain against an idea or thinking about something, forget about it and come back to it later with fresh eyes. It helps a ton. Doubly so for designing IMO.
Yeah I haven't done as much game dev as I'd like, but when working on papers I always try to schedule it where I have one week between finishing the paper and starting to revise. Then I basically phase out working on it with taking breaks doing as many different things as possible, so that I'm approaching it with different things on my mind.
You can also do things like "Play with a different control scheme" if that's possible, or "play on a different computer"
There's a pretty hefty terrain/world bundle at the Unity asset store for 50% off... Of course there's no saying how nice those assets will play with the upcoming tech.
I'm attempting to get back into dev and decided to first (continue to) educate myself using free resources starting with the tutorials that are features on the Unity site. After finishing the "Tanks" tutorial I decided to give myself credit by tracking my further progress in a spreadsheet....because I like spreadsheets. In my previous attempts at gamedev I worked through all of the "beginner" tutorials and stopped at Tanks because the first video was an hour and screw that! But now I'm more mature (older and boringer) so I can handle an hour of learning stuff. But I still watch all the videos at 2x speed with captions turned on, really helps burn through the live trainings.
That time includes watching the tutorials including backtracking, experimenting with the techniques in them, debugging anything that is for older unity versions, downloading assets, updating unity, etc.
My favorite so far is the Roguelike tutorial not because its anything special but because the architecture of it is so bizarre (in my mind) for a 2d, turn based, tile based game. I don't think I would ever use something like it but its neat to see really creative solutions.
In that spreadsheet are quick notes, more detailed notes/questions, and a page for topics to explore further.
For me seeing tracking my time really encourages me to do more even if it isn't for any external reward or other validation.
I'm attempting to get back into dev and decided to first (continue to) educate myself using free resources starting with the tutorials that are features on the Unity site. After finishing the "Tanks" tutorial I decided to give myself credit by tracking my further progress in a spreadsheet....because I like spreadsheets
Awesomed for spreadsheet use and also I think tutorials are generally just a good idea.
Though... I give IT trainings. I also create eLearnings. I can't watch tutorials, except it the narrative is outstanding. I think gamedev tutorials are aimed at people who want to get their hands dirty, but they're really boring because you cannot get you hands dirty. You can try and follow (if the unity version matches the tutorial) and pause and copy paste, but there's just something that doesn't mesh.
I'm going to try and start making tutorial like vids in my day job, and so it's something I'm thinking about a bit now and then.
I'm attempting to get back into dev and decided to first (continue to) educate myself using free resources starting with the tutorials that are features on the Unity site. After finishing the "Tanks" tutorial I decided to give myself credit by tracking my further progress in a spreadsheet....because I like spreadsheets
Awesomed for spreadsheet use and also I think tutorials are generally just a good idea.
Though... I give IT trainings. I also create eLearnings. I can't watch tutorials, except it the narrative is outstanding. I think gamedev tutorials are aimed at people who want to get their hands dirty, but they're really boring because you cannot get you hands dirty. You can try and follow (if the unity version matches the tutorial) and pause and copy paste, but there's just something that doesn't mesh.
I'm going to try and start making tutorial like vids in my day job, and so it's something I'm thinking about a bit now and then.
I watch videos the way I read text. I skip past the parts I already know and re-watch a new segment repeatedly as I work with it in my editor on the other monitor. That's why I really appreciate training videos with the transcripts (and timestamps!) posted below the video player.
After watching so many tutorials one thing that they really need to start doing is editing out any typing of code. I can copy and paste from the samples, or if I want to type it myself I can pause the video. There is absolutely zero need to SHOW someone typing code with the very narrow exception of demonstrating IDE techniques like autocomplete or refactoring.
Are there any Unity-heads that can point me to some docs for in-scene editing?
Current State: I handle my sprite animations with the normal Animator graph. My custom ScriptableObject describes my character's attacks, and it stores a reference to the corresponding AnimationClip asset. I made a custom editor to edit hitbox data in Scene View. I also made a component that visualizes hitboxes using Gizmos.DrawCube.
Goal: Edit hitboxes in Scene View by changing the displayed sprite to the same hitbox frame I'm editing in the custom editor.
Problem: I don't know how to change the sprite displayed in Scene View using the AnimationClip stored on my ScriptableObject.
Possible Solution: Change the ScriptableObject to store an Animator stateName string that can be passed into Animator.Play(). But then how would my ScriptableObject retrieve the AnimationClip asset to get data like keyframes using only that string?
Are there any Unity-heads that can point me to some docs for in-scene editing?
Current State: I handle my sprite animations with the normal Animator graph. My custom ScriptableObject describes my character's attacks, and it stores a reference to the corresponding AnimationClip asset. I made a custom editor to edit hitbox data in Scene View. I also made a component that visualizes hitboxes using Gizmos.DrawCube.
Goal: Edit hitboxes in Scene View by changing the displayed sprite to the same hitbox frame I'm editing in the custom editor.
Problem: I don't know how to change the sprite displayed in Scene View using the AnimationClip stored on my ScriptableObject.
Possible Solution: Change the ScriptableObject to store an Animator stateName string that can be passed into Animator.Play(). But then how would my ScriptableObject retrieve the AnimationClip asset to get data like keyframes using only that string?
I got an idea but keep in mind im talking out my ass here since I've never done any of this stuff, also I acknowledge that I could be totally misunderstanding your issue.
I'm pretty sure there are tools to turn 2d art into voxel models. If you did that for every animation frame you could use extract the mesh data to generate mesh colliders for each frame. Then use animation behaviors to communicate which frame you are in to your object controller changing to the appropriate mesh collider in each frame.
Are there any Unity-heads that can point me to some docs for in-scene editing?
Current State: I handle my sprite animations with the normal Animator graph. My custom ScriptableObject describes my character's attacks, and it stores a reference to the corresponding AnimationClip asset. I made a custom editor to edit hitbox data in Scene View. I also made a component that visualizes hitboxes using Gizmos.DrawCube.
Goal: Edit hitboxes in Scene View by changing the displayed sprite to the same hitbox frame I'm editing in the custom editor.
Problem: I don't know how to change the sprite displayed in Scene View using the AnimationClip stored on my ScriptableObject.
Possible Solution: Change the ScriptableObject to store an Animator stateName string that can be passed into Animator.Play(). But then how would my ScriptableObject retrieve the AnimationClip asset to get data like keyframes using only that string?
I got an idea but keep in mind im talking out my ass here since I've never done any of this stuff, also I acknowledge that I could be totally misunderstanding your issue.
I'm pretty sure there are tools to turn 2d art into voxel models. If you did that for every animation frame you could use extract the mesh data to generate mesh colliders for each frame. Then use animation behaviors to communicate which frame you are in to your object controller changing to the appropriate mesh collider in each frame.
That's related but not quite the issue. I already have a system in place to use AnimationEvents which call back to my fighter object, and it seems to work correctly.
The issue is that I can't edit them, which an automated message generator wouldn't solve. For example, lots of games make your hurtbox smaller than your total sprite, and I need a way to visually edit that for each frame.
The laborious way would be to adjust a collider to what might be right, then copy it to my ScriptableObject, then remove the collider, then test. But there's got to be a way to just show a particular frame of animation in scene view, so I can adjust against the actual sprite.
Posts
Twitch: KoopahTroopah - Steam: Koopah
I've been working around putting more procedural stuff on those little screens for a while now, and I keep running into, well, not lack of computing power, but Unity's atrocious swapping management. I don't see any of the problems I've run in to here (which mostly show as visual judder. I'd like to see the camera in that demo do a 360° spin though, because so far that seems to be the move that gets the judder issues. And also maybe on an iphone XR?
But I want those pretty SHADERS, wants wants...
How far I got with futuristic cityscapes/orbital habitats: EDIT: mind your sound levels. It's loud.
Hey, who made your music?
I should have mentioned that in the vid comment: the music is by https://www.bensound.com !
I've been using Kevin MacLeod, as seen in many youtube videos. In fact I had to take a couple of tracks out after hearing them on popular channels.
Here's my attempt at "going viral" (I had a blast doing these).
https://youtube.com/watch?v=PG_mBbYcCT4
To bring it back to game development. Steam actually checks that the game supports what it claims to support, didn't know that. Very glad they do and the marketplace isn't as lawless as I've heard.
Uploaded alpha
Sent baby brother a link due to him mainly being an nba 2kx console gamer and he can give me feedback that is as blunt as i need
http://www.fallout3nexus.com/downloads/file.php?id=16534
I am pretty close to turning to sheet caking at this point,
This has cemented my desire to move away from Unity. Even if I am just being an idiot and missing obvious, Unity's entire set up is so contrary to the software development best practices I practice at my job every day. You want reproducible builds? Instead why don't you tune hundreds of little gui dials and snowflake your builds to the point that no one is ever going to be able to get them to work again.
Obvious q. Have you tried the internal build?
Yeah that has a whole different set of errors :rotate:
I think the issues I am seeing now are due to Google's play plugin, https://github.com/playgameservices/play-games-plugin-for-unity.
Edit: yeah looks like a bad upgrade happened at some point.
I am hoping that Unity's package manger is eventually extended to 3rd party developers. Having to check full set of lib files into a projects source tree becomes difficult to manage.
Need to practice on different computer to see if i set it up properly
http://www.fallout3nexus.com/downloads/file.php?id=16534
I can't mention anything specific in the forums 'cause I need to keep my personal life seperate, but I'm pretty happy with it so far given the small scope I was aiming for.
I've been kind of shaky all day from pushing that button. I kind of think I'm going to need a drink once it goes out to the world at large.
Before I got started, I literally hadn't coded anything more complex than a single batch suite I did a few years back (itself my first batch code!), and never even touched the coding language I was using before... and my history with coding has been less than stellar. My education is in English.
The moral of the story though is you absolutely can make a game no matter how new you are and how little you don't know so far. You gotta start small and work your way up and fail fast and all that jazz, but no matter where you start you can get there if you're willing to put the work into it.
That coding literature could use some work though.
Are you guys bored yet? I am. Well, not enough to ditch this project and work on the actually nearly finished ones, but it's becoming annoying.
Behold, my water:
What you're looking at is, first of all, an offset problem with the meshes. This is pretty easy to fix!
Then, you'll notice that underneath my 'high detail' landscape mesh is now a mesh that is identical, except it doesn't dive down where there's water. This is the water surface! It's pretty clever even if I say so myself, and I was planning to do the same thing for roads.
However, it might strike you that the mesh used is pretty low poly. I mean, it looks like crap.
So now I have to decide which way to go.
-I've tried 'off the shelf' mesh subdivision and it's not working.
-I could try to simply generate a bigger array and generate detailed meshes of that
-Tiling! Which is probably what I'm going for anyway... But the first thing to do when tiling would be to try and ruin my landscape generator by making it more modular. The goal; to have it set up with a lower vertical resolution so it can take tiles with standard, stepwise slopes. This wouldn't end well! You can imagine the variety of tiles needed. 0-0-0-0 would be fine. But how many tiles with 0-1-5-9 slopes would I need?
So it's option B...
/highfive
English background here too with a similar background, thanks for sharing!
I was hoping to ship this game out before Epic's MegaJam starting this Thursday but it's too late now.
The problem is visualizing these colliders to edit them. Right now, I'm storing them as Bounds in a ScriptableObject, and I'm using Gizmos.DrawCube to show them in Scene View. But I can't figure out how to make Scene View show the correct sprite keyframe while I'm editing the boxes.
The closest I've gotten is storing an AnimationClip in the collider data ScriptableObject, but I'm not sure how to tell the monobehaviour to look into the SO, find the AnimationClip, make the animator play that clip, then make the animator change to the correct keyframe corresponding to the keyframe collider data that I'm editing.
Does anyone have a better approach? Should I be storing an Animator SetTrigger string instead of a reference to the clip asset itself? How would I go from trigger string -> ??? -> pulling keyframe data?
http://www.fallout3nexus.com/downloads/file.php?id=16534
I also learned i never uploaded a build to itch
http://www.fallout3nexus.com/downloads/file.php?id=16534
So for procjam I did get Unreal level streaming working correctly. I created several levels, each with a single terrain object. Each level corresponded to a certain configuration of walls. Then I used level streaming to load and place the terrain chunks.
It worked OK but the terrain takes up a lot of storage. I guess one cool thing is that I could go back and touch up the terrain with foliage and more textures without changing any code. I wanted to add some "special" terrain tiles like a big pit or a bridge but ran out of time.
It didn't suck.
Also realized where i want to go with the art syle
http://www.fallout3nexus.com/downloads/file.php?id=16534
Lots of neat assets in there, tons of value for $20.
Twitch: KoopahTroopah - Steam: Koopah
Already got plans for who I'll ask after that
http://www.fallout3nexus.com/downloads/file.php?id=16534
Twitch: KoopahTroopah - Steam: Koopah
You can also do things like "Play with a different control scheme" if that's possible, or "play on a different computer"
So far since I've been tracking over the last two weeks I've gone through 15 hours in the tutorials. https://docs.google.com/spreadsheets/d/1-U1EwQJQbucXois9ZukRYjELU7SAQACjCwOSHQCI-II/edit?usp=sharing
That time includes watching the tutorials including backtracking, experimenting with the techniques in them, debugging anything that is for older unity versions, downloading assets, updating unity, etc.
My favorite so far is the Roguelike tutorial not because its anything special but because the architecture of it is so bizarre (in my mind) for a 2d, turn based, tile based game. I don't think I would ever use something like it but its neat to see really creative solutions.
In that spreadsheet are quick notes, more detailed notes/questions, and a page for topics to explore further.
For me seeing tracking my time really encourages me to do more even if it isn't for any external reward or other validation.
Awesomed for spreadsheet use and also I think tutorials are generally just a good idea.
Though... I give IT trainings. I also create eLearnings. I can't watch tutorials, except it the narrative is outstanding. I think gamedev tutorials are aimed at people who want to get their hands dirty, but they're really boring because you cannot get you hands dirty. You can try and follow (if the unity version matches the tutorial) and pause and copy paste, but there's just something that doesn't mesh.
I'm going to try and start making tutorial like vids in my day job, and so it's something I'm thinking about a bit now and then.
I watch videos the way I read text. I skip past the parts I already know and re-watch a new segment repeatedly as I work with it in my editor on the other monitor. That's why I really appreciate training videos with the transcripts (and timestamps!) posted below the video player.
I got an idea but keep in mind im talking out my ass here since I've never done any of this stuff, also I acknowledge that I could be totally misunderstanding your issue.
I'm pretty sure there are tools to turn 2d art into voxel models. If you did that for every animation frame you could use extract the mesh data to generate mesh colliders for each frame. Then use animation behaviors to communicate which frame you are in to your object controller changing to the appropriate mesh collider in each frame.
I found this tutorial really informative on the subject of hitboxes and hurtboxs https://strangewire.blogspot.com/2018/05/hitboxes-and-hurtboxes-in-unity.html and while I'm not interested in fighting games I can see how those techniques could be used to add more intuitive interactions to any game type
That's related but not quite the issue. I already have a system in place to use AnimationEvents which call back to my fighter object, and it seems to work correctly.
The issue is that I can't edit them, which an automated message generator wouldn't solve. For example, lots of games make your hurtbox smaller than your total sprite, and I need a way to visually edit that for each frame.
The laborious way would be to adjust a collider to what might be right, then copy it to my ScriptableObject, then remove the collider, then test. But there's got to be a way to just show a particular frame of animation in scene view, so I can adjust against the actual sprite.
Twitch: KoopahTroopah - Steam: Koopah