BluesM18A1

BluesM18A1

@claymorelabs.xyz

Game Dev, sometimes posts VGM remixes on Youtube and draws stuff. More info on my work at www.claymorelabs.xyz Man inside a robot body I may be, but I greatly frown upon those who outsource critical thinking and creativity to hallucination algorithms.

131 followers101 following158 posts

Pinned

So here's what my game looks like so far. It's got a grappling hook with very springy physics and will be focused on speed and maybe some combat. I'm calling it 'Claws of Thunder' and hopefully there will be a playable demo some time in a year or two. #gamedev

Claws's development is now at a point where I'm running out of new features to implement but finding bugs that warrant going back to refactor lots of code that was written several years ago when I had little idea what ideal object composition practices were. I can't wait to work on art assets again.

BluesM18A1 reposted

Make it exist first... then leave it exactly as it is because quite frankly other tasks take priority sometimes. It's already good enough and nobody will spot that tiny mistake, and the fact that you created something in the first place is already an achievement

I've not been giving myself enough time to draw purely for fun lately, was always a comm or an asset for some bigger project. Never tried putting serious effort into drawing my #KOF main either so badabingbadaboom #art

I take back what I said about 2d polygon-based level construction. I'm doubling down on it and going full hog with extensions, plugins, even some straight up engine modifications to make this level editing workflow as smooth as possible.

It's a source of shame that my hopes of having a playable demo ready by the end of the year are all but dashed. Alas, most of my gamedev experience is with bringing characters to life since that's like 80% of MVP in jam settings. Not so much environment art/level design experience. This changes now.

(translator's note: MVP means "minimum viable product" in this context but I guess it almost makes sense for it to also mean "most valuable person")

I get really annoyed by people using the word "emulation" when they really mean "piracy". I understand being afraid to more blatantly promote piracy but this has the side effect of geriatric legislators, lawyers & judges not understanding the difference and throwing legal emulation under the bus.

Getting so pissed off at Godot's Polygon2D editor that I'm probably gonna put Claws's development on hiatus AGAIN just to contribute fixes to all my nitpicks and annoyances with the editor. I've got a lot of brushing up on C++ to do.

Headcanon: Kraid isn't one guy, but a species that the space pirates domesticated/genetically modded to be guard dogs of a sort. The design for Kraid used in the first game is the wild undomesticated version that hasn't been bred for size and tougher hide that resists Samus's weaponry.

Since Ridley's returns in the Prime series, Super, Other M, Fusion, etc. have logical explanations like being cyborg'd resurrected remains or cloned, and Kraid has no such explanation for his return in Super and Dread, this is how I imagine it. The Kraid on ZDR was probably captured/stolen by chozo.

Deadlock is TF2 if everyone had Scout's movement, a shitty pistol as your primary weapon, Soldier's rockets but you only get to fire one every 3 seconds, and 3x the health of a fully overhealed Heavy. Also you're playing 5CP with MvM upgrades. (No I really don't like it).

Didn't draw the lower legs bc I couldn't figure out what should be there since your reference didn't specify. Anyways here you are xd

aeri@interwebs:~/bsky$@aeriavelocity.dev

If you're an artist and you like me and you see this you should definitely draw Astrid ahahahahahaaaa

Im ngl, way bigger fan of Mother Brain's design from the first Metroid. Girl lost all her aura when Super decided to give her a single big round eye and that goofy ass body. The old design screams "biomechanical horrors beyond our understanding" a lot more.

Hi yea it just occurred to me I haven't posted jack in a while about my game's development. It's been taking a back seat for a while because someone very close to me needed programming help with their Undertale fangame. Ain't much left I gotta work on before I get back to my own project. Hang tight.

Possibly the worst part of wanting to learn a completely new instrument and buying the cheapest one available is that you have no idea if the instrument is built like shit or your technique is just bad. Is that gonna stop me from trying to learn violin? nope.

Migrating PC to new case. GPU was mere milimeters too long because of this stupid notch protruding out the back. Shaved down the plastic with a hobby knife to get it to fit. Looks like I could really use a rotary cutter.

Here I will be posting a thread about what I learned by periodically profiling the performance of my game on a 12 year old potato. A lot of this thread's info may be Godot Engine-specific, but I hope it's insightful either way.

View full thread

Very heavy yet careful use of VisibleOnScreenEnabler nodes drastically lightened the CPU load, but noticing how there were yet further performance gains by deleting half my level gave me one big question about how I approach level structure in my game:

Do I still want to have very large levels that take 4-8 minutes to complete? Should I instead pivot to having a bunch of smaller levels that take 2-4 minutes? Should I try to split the big levels into chunks that are loaded as needed? How would I make loading as seamless as possible?

Here I will be posting a thread about what I learned by periodically profiling the performance of my game on a 12 year old potato. A lot of this thread's info may be Godot Engine-specific, but I hope it's insightful either way.

View full thread

So with that in mind, I did figure out a couple things that drastically lighten the CPU load even on my main workstation (This is the part that's rather Godot-specific):

Every last physics object, even StaticBody2Ds, takes a certain amount of CPU power to account for that really adds up when you have hundreds of them all over a big level. "Pausing" them when far off-screen is very helpful, but removing them entirely when not in use is even better if you can help it.

Here I will be posting a thread about what I learned by periodically profiling the performance of my game on a 12 year old potato. A lot of this thread's info may be Godot Engine-specific, but I hope it's insightful either way.

View full thread

The second-biggest performance concern is CPU power. I decided very early on to have the physics tickrate set to 240hz with no interpolation instead of the default 60hz with interpolation on. This made the game feel much more precise and smooth, but it means the system has to work extra hard.

Now I leave Vsync turned off for testing that way I can see the utter maximum amount of frames the system could possibly churn out per second with the current load, and what I found is running at 150-200FPS isn't that terrible for the physics, but any further below makes things feel awful.

Here I will be posting a thread about what I learned by periodically profiling the performance of my game on a 12 year old potato. A lot of this thread's info may be Godot Engine-specific, but I hope it's insightful either way.

View full thread

Now even if my game is 2D, it still uses high-res spritework designed for 1080p and 4K. As you might guess, this means VRAM is still a very big consideration for this game's performance. 1GB of VRAM is just a hair short of enough to handle this game, even at a low resolution of 1024x768.

Finding this out made me extremely conscious of the resolution of every image file I put in the game, making sure they are *only* as high-res as they need to be. For example, background elements can afford to look slightly blurry, as that helps you read the foreground elements, so I downscaled em.

Here I will be posting a thread about what I learned by periodically profiling the performance of my game on a 12 year old potato. A lot of this thread's info may be Godot Engine-specific, but I hope it's insightful either way.

EXHIBIT A: The Dell Optiplex 990. I got mine for (I think) $35 used. Out of the box it has an Intel Core i5-2400, and I fitted it with a Radeon R7 200 series GPU as it's one of the oldest and weakest that support Vulkan. Later I will be upgrading it to an R5 430 for reasons I'll explain later.

I promise I'm still working hard on my game, it's just that I got very little to show for it. It's all a bunch of visually boring stuff like *checks notes* just making sure everything works as intended and figuring out what works/what doesn't for level design

This is getting out of hand. Now there are two of them! (Really wanted a PJ so I could have more tone versatility, but that classic P bass will always have a use)

Some #art of friends' #OC s I did for them. I should probably publicly open for commissions later. Would just need to set up a paypal business account is all.

BluesM18A1 reposted

Linux experience so far has been a mix of either "this thing works better and faster and more elegantly than Windows" or "this thing took me an hour of googling to get working" with very little in-between.

There's a whole lot of story I want to tell in the form of a comic, but the main thing holding me back besides my debilitating ADHD is my instinct to test the limits of my skill. Slowly though, I'm figuring out a flow that reconciles the desire to improve with the need to finish the product.

the advice I always give about comics making in a professional context is not expecting to draw everything at your 100% like each panel is a single illustration; it's about getting your 60% to be Good Enough. if you don't have deadlines tho, different story depending on what ur trying to accomplish