Vapourware Codexian Game Development Thread

[zwanzig_zwoelf]

So, what is your game like? Can you tell more about it?

I don't want to spoil too much, but you get to explore a small world with occasional dungeons sprinkled throughout to find out what happened to your father while getting stronger through battles and completing sidequests, some of which reward you with new items/abilities. Combat is handled in real time via bump attacks and focused on finding the fastest way to kill each enemy class.

Have you coded it from scratch or used unity or other engine?

I used Unity but did the code/assets apart from logo/cover art, the latter was produced by Attila Bertold Bozó.

It says saga in the title, why is that?

I came up with multiple game plans while figuring out what kind of game I wanted to do next and decided to do all of them. It allows me to treat each game as a separate 'book' in one (albeit loosely-connected) overarching plot.

 

[Tweed]

There's no real reason to play the original Hydlide except for historical purposes. Super Hydlide perfected those old systems.

 

[Kalarion]

C**l thy t*ts, *st*r*sks *r* p*rf*ctly f*ne.

Call thy tats, asterisks ore perfectly fane?

 

[ERYFKRAD]

C**l thy t*ts, *st*r*sks *r* p*rf*ctly f*ne.

Call thy tats, asterisks ore perfectly fane?

Bless you that's almost good. C**l oughta have two vowels though.

 

[Justinian]

Hi guys, I'm making a Vampire Survivors clone. My intent is to translate the "mood" of a more traditional rpg/roguelike to the genre. The gameplay is still similar to VS: you and your autoattacks vs endlessly spawning hordes of monsters, but in a traditional fantasy setting. You pick one of many fantasy races (6 implemented so far, 12 more planned including 6 monster/race combos like in M&M8) and one of many classes (12 so far, will add 4 more in an update today or tomorrow and plan to have at least 40 by the time the game is out of early access). Like VS your attacks are automatic but some of the attacks also include summons (that duke it out with monsters and are targeted and killed by them) and some attacks are aimed with your mouse.

https://store.steampowered.com/app/2078710/Mage_Lords_of_Brams/

 

[Twiglard]

Of course, all assets are placeholders.

 

[infidel]

A more attainable goal I have is to make enough money to justify going down from 5 to 4 days a week at my regular job. But even that seems extremely hard.

Heh, I've had an even more modest goal of making 100$ per month through my gamedev to fuel (or limit?) my music software purchasing addiction. Parasite is in Early Access and from the initial good bump in sales it now fell down to less that 100$ that you need for Steam to bother sending you the money. It took 7 months for the profits to dry up and it looks like Steam steadily decreased the traffic to my game page roughly by half with each passing month. Now I need to finish the game and get another bump of sales (at least they promised they will do a proper launch visibility round). But it looks to me that if you're like me, a nobody with a hobby, there's no reason not to release your game in EA just for that second bump. The caveat is that your game has to be playable and look like something that is close to completion, I think. You can't just sell a page with nothing.

Recently I've read an article from an indie dev/marketer guy who compiled his research on game profits into gamedev profit tiers. Something like <5k$, 5-10k$, >10k$ and >250k$ for the lifetime value. <5k$ is just shit that you can't live off, 10k$ in a year or two is where you can live but you have to be frugal, but why 250k$? Valve said that when they send traffic to games they look on how much money that game made. And his research has led him to believe that 250k$ is the magical number. This is the time when Valve's algorithm decides, ok, you're a good game and puts you on their good list and visits to the game page becomes infinite and plentiful like manna from heaven, since they start putting you on all the important parts of their main page from time to time.

Now that revelation and my own experience has led me to the sad conclusion that there is literally zero point in updating your games past a very certain point. Valve literally does not want you to! A new game from you will receive a guaranteed bump of traffic and sales in return for 100$ paid for the place on the platform. But an old game that you update will get nothing unless you have enough active players to see the bump in activity that might transform into DLC sales or friend sales. The only thing you'll get is that warm feeling...

 

[The Avatar]

Recently I've read an article from an indie dev/marketer guy who compiled his research on game profits into gamedev profit tiers. Something like <5k$, 5-10k$, >10k$ and >250k$ for the lifetime value. <5k$ is just shit that you can't live off, 10k$ in a year or two is where you can live but you have to be frugal, but why 250k$? Valve said that when they send traffic to games they look on how much money that game made. And his research has led him to believe that 250k$ is the magical number. This is the time when Valve's algorithm decides, ok, you're a good game and puts you on their good list and visits to the game page becomes infinite and plentiful like manna from heaven, since they start putting you on all the important parts of their main page from time to time.

I can see why they do that, given all the shovelware on Steam. The best thing an indie dev can do is to make the best game possible and promote it outside of Steam. If it's a good game, people will buy it.

 

[rusty_shackleford]

paying someone to do PR is worth it if you can afford it
especially if you're not good at shilling

 

[Twiglard]

Did a lot of 'work' in the last few days.

boring

• removed padding from tiles; on the toy editor demo's release build, this accidentally increased the framerate from 2400 fps to 3000 fps (adding pragma pack(push, 1) had an opposite but less pronounced effect).
• finally fixed the wrong texel offset off by 0.5. I recommend a test image (to be treated as a 2x2 tile atlas) for this purpose.
• can snap horizontally and vertically while dragging placeables (floor and walls) by holding shift or ctrl, then can switch direction by changing the modifier key and it continues dragging without snapping back.
• the serializer works and in its basic form can be tested at compile-time (i.e. uses <bit>). storing a single chunk still takes between one and three kilobytes.
• haven't started work on undo/redo, scenery, interactibles like doors, or displaying animations, but I have existing code on making animation atlases and metadata.
• tile atlases are loaded from simple .json metadata and selected through a UI pane (using imgui). can paint a selected tile but also a permutation of the entire atlas (each tile is represented at least once while laying them down).

I could make a video but nothing of it looks impressive in any way. But these things have to work under the hood for anything interesting to happen on screen in the future.

 

[The_Sloth_Sleeps]

Did a lot of 'work' in the last few days.

boring

• removed padding from tiles; on the toy editor demo's release build, this accidentally increased the framerate from 2400 fps to 3000 fps (adding pragma pack(push, 1) had an opposite but less pronounced effect).
• finally fixed the wrong texel offset off by 0.5. I recommend a View attachment 30064 test image (to be treated as a 2x2 tile atlas) for this purpose.
• can snap horizontally and vertically while dragging placeables (floor and walls) by holding shift or ctrl, then can switch direction by changing the modifier key and it continues dragging without snapping back.
• the serializer works and in its basic form can be tested at compile-time (i.e. uses <bit>). storing a single chunk still takes between one and three kilobytes.
• haven't started work on undo/redo, scenery, interactibles like doors, or displaying animations, but I have existing code on making animation atlases and metadata.
• tile atlases are loaded from simple .json metadata and selected through a UI pane (using imgui). can paint a selected tile but also a permutation of the entire atlas (each tile is represented at least once while laying them down).

I could make a video but nothing of it looks impressive in any way. But these things have to work under the hood for anything interesting to happen on screen in the future.

what library?

 

[Twiglard]

what library?

Magnum.

 

[The_Sloth_Sleeps]

what library?

Magnum.

How well do you find this for 2D, 2.5D isometric?

 

[Twiglard]

How well do you find this for 2D, 2.5D isometric?

Rendering straight from flat textures makes it way easier than using pre-rotated sprites like Fallout and Arcanum. But to do this I have to render the floor and walls antialiased. I'm still going to have to make large texture atlases to skip having to bind different textures all the time while rendering tiles. With GL only a single texture can be used per draw command. Once atlases are done properly I can render up to 1K triangles at once or change the chunk size constant.

Magnum has a good level of abstraction over GL with Mesh and MeshView types doing indexed multidraw. Also the GL_DEBUG extension means that it doesn't have to check for errors on each GL call.

I struggled a bit to make the first VBOs, the coordinate system, the right kind of projection and foreshortening. But it wasn't much worse conceptually than rendering it all on the CPU (that needs UV coordinates anyway).

GUI was easy thanks to Magnum's ImGui backend. Sadly the screenshot doesn't show the UI properly. Try this instead.

If you get into it then you can ask questions on the Gitter channel. They're fine with general GL questions as long as you've done your homework.

 

[Nathaniel3W]

Alright, my dudes. Trailer for my next game is up. I need you tell me what sucks so I can change that in the next trailer.

(It looks like that's not working. I'm not sure why. If it doesn't load here, I'd appreciate you watching it direct from YouTube https://www.youtube.com/watch?v=32WcYyvMQ70 )
Also, would appreciate Steam page feedback if you have the time. https://store.steampowered.com/app/2091620/Septaroad_Voyager

It may look very similar to Himeko Sutori, but I had to mess with a lot of the underlying engine to make it work with DX11 instead of DX9. I needed DX11 so I could avoid slowdowns with the dynamic lights when characters attack. Dynamic light effects to simulate clouds and overhead leaves too. Also, I made entirely new meshes to get the texel size consistent and to make them all look more pixelly so they would fit more with the characters. And I reduced the palettes on everything.

This is day one of the trailer and store page being live, so if there are any glaring problems, hopefully I can fix them before the rest of the world sees.

 

[Ninjerk]

Nathaniel3W Nothing jumps out in the trailer. The difference in the eyes on the main piece of art looks a little weird.

 

[The Avatar]

Septaroad Voyager. The title sounds a lot of Octopath Traveler. I'm guessing that's intentional?

 

[Nathaniel3W]

Septaroad Voyager. The title sounds a lot of Octopath Traveler. I'm guessing that's intentional?

Yeah, intentional. Started off as inside joke only to me. I started working on my next project and tried coming up with a codename. I named all of my stuff SRV* for Septaroad Voyager as a play on Octopath Traveler. Then when I wanted to announce the game, I couldn't think of anything better. And here we are.

 

[barker_s]

Aren't you afraid of a lawsuit? It's kinda close. I would've gone at least for Pentatrail Excursionist, maybe even lower.

 

[Twiglard]

Managed to get animations running. Because my pixel shader does isometric projection on its own, it was necessary to invert it for this purpose. Fortunately, this is simple enough with Wolfram Alpha:

If this is the isometric projection with foreshortening,

Then this is the inverse projection (note no Z component):

To be honest I'm more comfortable using either Wolfram Alpha or Mathematica than second-guessing my own algebra even for such basic cases.

For animations, frames are rendered inside Blender with a constant camera (note the magic angle).

The simple atlas-making program I made for this purprose extracts image bounds from an image such as this:


Each animation has a defined Fallout-style ground offset which is gotten through rendering a small cube at origin just once for given camera scale. The first image in the first cardinal direction serves as a reference for how much scaling is to be performed. Finally making the images into an atlas (8-bit indexed color used due to attachment size):



game/blender/door % ~/dev/game/build-msvc/install/bin/[snip]-anim-crop-tool door_close.json && gm convert door_close.png door_close.tga && cp door_close.json door_close.tga ../../anim/

And it manages to render itself, but only after much head-scratching with regard to world and render space coordinate systems.

 

[Nathaniel3W]

Aren't you afraid of a lawsuit? It's kinda close. I would've gone at least for Pentatrail Excursionist, maybe even lower.

I'm not really concerned although reddit is dragging me through the coals over it. I don't think anyone there knows what they're talking about. They keep saying copyright when what really would apply is trademark, but I'm not violating trademark either. I'm not using any part of their name and you can't just trademark every possible combination of numbers and people on a journey.

Twiglard that's cool and all, but what exactly are you doing? Are you prerendering 3D models as pngs?

 

[barker_s]

My comment was mostly meant to be a (low quality) joke, I know you're not in any legal danger in regards to the title. And hey, at least you got a title! 5 years into development and mine is still 'untitled 4x game'.

 

[ERYFKRAD]

untitled 4x game

Just call it Foreign Exchange and prepare for a massive collective groan as people realise the terrible joke.

 

[Nathaniel3W]

ERYFKRAD It took me a second. Thank you for a joke that's so bad I can't tell it to anyone but my kids, and a joke that's too complex for me to tell my kids.

barker_s Sorry. Yeah. I'm probably just on edge. I got the joke though. I actually did go through possible variations in my mind. Don't think I came up with yours though.

What's the setting of your game? I think we need to name it today. Maybe put it to a poll if necessary.

 

[ERYFKRAD]

Thank you for a joke that's so bad I can't tell it to anyone but my kids, and a joke that's too complex for me to tell my kids.

Imagine how I feel, I came up with it.