Não é um jogo. Não é uma coletânea. Não é um software para PC. Não é uma ferramenta. É algo que eu sempre sonhei em criar para o Mega Drive, algo que eu imaginava desde a minha adolescência e que agora está se tornando realidade (graças à IA, lógico). Em breve, espero mostrar este meu projeto, que, para mim, acho difícil que outro me dê tanta alegria quanto este. A criança dentro de mim está pulando de alegria ao ver que está dando certo.
Lá vamos nós... Deixa eu pegar aqui a minha bola de cristal. Hum... O que eu vejo sobre o uso de IA no futuro?
No futuro não irão mais dizer que é "vídeo feito por IA", só vão chamar de vídeo. Farão o mesmo para músicas, imagens, programação etc.
O cara não programa, faz inúmeros prompts (e às vezes um só), e está lá o resultado (jogo, aplicativo etc.). O cara só dirá "eu que fiz", sem detalhe algum, e ninguém mais vai perguntar sobre qual IA usou, e se usou.
Quanto tempo isso pode levar? 1 ano? 5 anos? Só sei que, da mesma forma que a máquina de escrever foi substituída pelos editores de texto nos computadores com impressora, o mesmo vai ocorrer com o uso da IA, porém, desta vez quem está sendo substituído é o talento humano (uau, terrível quando se pensa isso em voz alta).
Você pediu, a coisa está lá. Todos "serão" músicos, programadores, artistas e seja lá o que for. E lembrando, irão virar isso à custa do talento (roubado) dos outros.
Se eu gosto? Não sei, francamente. Mas é para isso que tudo parece estar apontando. O conhecimento humano já está se tornando descartável; tenho até medo de ver o povo "emburrecendo" com o excesso de IA.
Vi amigos que até ontem não usariam IA por nada; agora dizem que estão usando IA como ferramenta (era inevitável, né?), agora estão "programando", e provavelmente amanhã estarão fazendo jogos autorais inteiros em uma semana. E tem algo errado nisso? Não tem, absolutamente. Isso já é a realidade. Eu mesmo não me via usando IA até há algumas semanas.
Bem-vindo ao novo mundo, seja bom ou ruim, seja o céu ou o inferno, seja você usando a sua capacidade ou (roubando) a alheia, seja você empregado ou não, seja seu conhecimento útil ou obsoleto, este está se tornando o novo mundo. Se este mundo vai ser melhor ou não, se vai ter espaço para mim ou para você, também não sei. Mas este novo mundo, como diria Thanos, é INEVITÁVEL.
🔊Sound ON!
Now THE RISE (MSX1) seems corrects: thanks for tips from the developer on the thread, the full SCC sound seems to work. It was a matter of selecting the Yamanooto as the mapper, instead of adding it in the openMSX extension slot.
Anyway, fantastic game, I loved it.🦇
THE RISE for MSX (by Recitta/2026).
Work in progress MSX1 game, really well made and impressive for the hardware.
It's the sequel to the MSXDev’05-winning The Cure.
Designed to run on any 64KB MSX system, the plan is to release it on a custom Yamanooto cartridge.
Recorded out of openMSX version 21.0.
More info:
recitta.com/
** Sega Megadrive - Lionheart port **
Thalion's Lionheart (1993) is one of the best-looking games on the Amiga. 100s of colors on screen with loads of parallax and an atmospheric sound track and just dripping with quality in every aspect.
So what's the reason for the port ? the challenge for one - there's the Amigas Copper co-processor to emulate on MD in trying to match the visuals, the Audio to try and do justice to with an FM flavor to it , try and get some improvements to gameplay also - particularly the control system.
So how does the MD stack up and differ from the Amiga original so far ?
More of the level on screen. The Amiga game plays in a 280×208 window outdoors, cut to 192 vertical lines in the caves. The MD port uses the full 320 × 224 (NTSC H40 resolution) in all sections so 23%-33% more viewable screen thoughout.
60 hz framerate Megadrive vs 50 hz Amiga , 20% faster frame rate. Game play is unaffected - same movement over time , just a shade smoother .
Parallax added in the caves. On the Amiga, cave 2-1 is a 6-bitplane Extra Half-bright screen (32 colours + 32 half-bright copies) with no parallax / single layer. On the MD the half-bright colours come from Shadow/Highlight mode, each tile is drawn shadowed or normal, with the palettes solved to fit. The back wall has been split out of the map onto the otherwise unused plane B for parallax. The large spider boss is two layered sprite palettes, double-buffered in VRAM to retain all the colors and full frame rate.
Its not been easy though - theres no Copper on the MD ( the Copper is a co-processor synchronised to the display that amongst other things changes color as the screen is been drawn ) . No such luxury on the MD so under the hood, an assembly Horizontal-Interrupt (H-int) list rebuilds the Amiga's Copper color changes by writing color ram (CRAM) mid-screen, about 200 color writes a frame, most of them by DMA. The DMA color changes are driving up to 8 color changes per scanline which normally has negative affects for sprites ( DMA steals the VDPs time for building the sprite line cache ) , however with careful timing that issue has been resolved. There is color change going on in the background and foreground at the same time . The water uses per-line scroll effects for large parts of the screen. Color seams are flickered to improve the seaming as per the original game .
The real struggle - the MD's 512-colour palette against the Amiga's 4,096 means the sky gradients get rounded. The H-Int software copper takes nearly a fifth of the MD 68k cpu frame at times, which the Amiga gets for free with the Copper.
@Megadrive4ever8 has been working hard on Stage 1s music and I think hes doing an amazing job ! He's even using channel 3 extended mode to get effectively 2 channels from 1 on channel 3. Dual PCM channels are used for Music also so effectively 6 FM channels and 2 PCM channels devoted to music. The SFX get their own 1-2 PCM channels, and they're the original sounds re-sampled to 14 khz for XGM driver.
Hence Music and sound effects play at the same time on the MD. On the Amiga, Paula's 4 PCM channels are shared, a SFX takes a channel and the music drops that voice while the effect plays.
The XGM1 driver is modified so idle PCM channels don't read ROM to save cpu . The full stage 1 tracks run for 4 mins + in length and is still WIP.
5.) Intros with Scaling maintained . The intro complete with zoom through the door and the walk down the hall are scaled at run time, by the blitter on the Amiga, and by 68000k assembly code on the MD. The MD draws each picture in about 70–80% of the Amiga's time albiet keeps the Amigas timing for accuracy.
6.) 320 x 448 interlaced resolution used for menu screens, the Amiga used high resolution interlaced 640×344 width here which we can't match horizontally but its still looking pretty sharp in 320 x 448 and this mode was seldem used on MD so great to find some use here.
Other difficulties has been the VRAM management. On the Amiga the playfield is a bitmap, the blitter copies 16×16 blocks into it, so there's no limit on how many unique graphics a single screen can use. On the MD every 8×8 tile on screen has to be in VRAM (64kb) . Stage 1-1 (shown) uses 4,745 unique tiles (3,466 foreground + 1,279 background). Once the plane maps, sprite and scroll tables, hero, HUD and enemy frames have taken their share of the VRAM - there are only about 1,500 tiles left for the background / forground. So both planes stream their scrolling tiles to cache whats needed at the time onscreen. The tightest frames are the diagonals. In a single frame you can need a new foreground column and a new row, plus a background row coming in and the water's per-line windows shifting, and all of it has to go through the vblanks (end of frame) DMA tight window.
Levels 1-1, 2-1, 1-2 and large parts of the intro been done and work continues on the rest of the port. The game is a mix of SGDK for cold paths and Assembly for most of the game time frame for speed - inline with my other projects.
Next up is finishing the rest of the stages with some more follow up vids of the next sections, more music to come which @Megadrive4ever8 is looking forward to implementing !
The game is been ported with AI assistance. The sheer workload of it conversion of assets / analysis / debugging would not be humanly possible in any respectable time frame otherwise and I already have several non AI assisted projects to contend with. The engine uses game code from my other projects but has been further developed for specific needs for this game.
#SegaMegaDrive#Genesis#Amiga#SGDK#RetroGaming
Tengen Gauntlet – Development Diary, Week 2
The NES version of Gauntlet, developed by Tengen, is, in my opinion, the most fun version of the game, despite the console's limitations. I know Gauntlet IV is amazing, has a wonderful soundtrack, and is even better than the original arcade version. But the NES version holds a special place in my heart because of the nostalgia attached to it. I spent countless late-night gaming sessions playing it with a friend who, sadly, passed away far too young.
I remember it like it was yesterday: going to my friend's house, opening the brand-new game manual, looking at all the items, enemies, and little details, and dreaming of having a version of this game for the MSX 1 (which was the computer I owned back then).
I tried recreating this game on the Colour Maximite 1 and 2 (there are a few videos showing these attempts on my YouTube channel), but it still wasn't quite what I had in mind.
Anyway, thanks to a lot of help from AI, I'm finally making that dream come true by porting the game to the Mega Drive and fixing a few things I never liked about the original, including the password system.
The game now uses SRAM with 3 save slots, so collecting too much treasure is no longer a problem. In the original version, the password would stop working if you got too greedy, as the manual itself puts it!
I'm also working on fixing slowdown and improving the graphics, using my previous Colour Maximite 2 project as a reference, which was itself based on the graphics from the Atari ST version of Gauntlet II.
Here are a few screenshots. Please keep in mind that the game is still a work in progress, and these images are not representative of how the final game will look.
To celebrate the 50th anniversary of my favorite series, here's a first preview of my dream port: Super Star Wars for the SEGA Mega Drive / Genesis, one of my favorite 16BIT games ever!!!
youtube.com/watch?v=dFGnDxnu…
Project Status – Week 2 / 2026
Here is the status of all projects to date:
Final Fight MD: It has been on hold until now, and I am currently planning to iron out the remaining bugs to wrap up the project and reach version 1.0, hopefully before the end of this year. I plan to use AI strictly to speed up testing, hunt for new bugs, and help fix them, since some are stack overflows that are hard to diagnose. Aside from that, I'll keep all coding in my hands.
Driftin' Rage: All tracks are finished. The uphills and downhills with more realistic slopes were developed using mathematical formulas and H-Int algorithms solved with AI assistance. Some graphical details still need polishing, and visual effects (tornado, night, fog, etc.), as well as new game modes, still need to be programmed. I feel the game has made huge progress, and I'll showcase all these improvements soon. I am already looking for a publisher.
Into The Darkness: Even though I've reached the target performance, thanks to conversion formulas and structures translated into Assembly with AI assistance, I am still not completely satisfied with the low texture quality. I plan to refine this game further before moving forward.
The Legend of Drakonnus: The game map and opening sequence are complete. Now I need to implement NPC dialogue, create the kings and area bosses, and fine-tune a few details in the game's physics.
Mega Commando: The core game is done; I just want to add underground zones and new weapons.
Prehistoric Isle MD: I want to see if I can squeeze out a bit more performance to wrap up the project at version 1.0.
Alone in the Dark: Both the 15MB and 7MB versions are ready; I am just doing a bit more testing.
Konami MSX Collection: I need to improve performance enough before releasing it to the public.
Super C: The port is ready, though I want to make a few graphical improvements before release.
Tengen Gauntlet: The port is ready with extra features (SRAM saving with save slots), though I want to polish some graphical elements.
Zeliard (Master System): The game is ready, and now I need to playtest it from start to finish to catch any remaining bugs.
Que inocência da minha parte achar que seria tão fácil portar Zeliard para o Master System. Os 8 KB de RAM batem à minha porta kkkk, sem contar os tiles que estou tendo que bravamente ajustá-los manualmente (ok, eu gosto de sofrer), porque se deixar pra IA, vira uma salada. A IA ainda queria (me forçar a) usar os gráficos do Tandy 1000, mas eu sabia que dava pra aproximar mais do MCGA, e novamente, eu tinha razão. Apesar da IA estar testando a minha paciência de uns tempos pra cá, já que com certeza "nerfaram" ela, ao menos a música e os efeitos sonoros do Tandy 1000 já vieram prontinhos e couberam como uma luva no SMS. Aguardem mais novidades.
Então você vê pessoas que em um único prompt com o Fable 5.2 fazem jogos FPS realistas para PC com gráficos AAA. Opa, fui testar isso no mundo retrô...
...Então comecei um port de um jogo para Master System e a IA arregou em vários pontos.
A verdade é uma só: pra fazer jogos, as IAs lidam muito bem com sistemas e arquiteturas robustas, elas brilham nisso. Mas quando falamos de sistemas extremamente limitados, ainda cometem erros básicos. Se eu não guiar em cima de algumas especificações, não ajustar inúmeros tiles pra caber na VRAM, ajustar cores e testar o resultado, mesmo as IAs ditas "top de linha", choram miseravelmente.
Sem esse controle de qualidade e capricho, facilmente ganhamos mais um projeto que entra na categoria de "IA slop", e que inacreditavelmente, muita gente nem percebe a atrocidade feita porque jogam uma única fase pra postar vídeos e fica por isso mesmo. E com isso ainda temos um monte de gente batendo palma para jogos, portados ou não, que nem foram testados direito, que nem tem um acabamento adequado, e que muitas vezes, nem serão concluídos.
Não estou falando mal da IA ou questionando as limitações ou simplicidade de jogos antigos (que eu adoro!), estou falando sobre polimento, capricho e qualidade do resultado.
E assim temos todo o mundo retrô sendo invadido por projetos classificados de forma justa como "IA slop".
Brand new Sega Genesis game!! This is Morning Star, a beatem up/hack n slash game inspired in 90's classics as Golden Axe and Dungeons and Dragons. The game will feature 4 playable heroes, 2p mode, magic attacks, magic items, mountable creatures and a lot more! Check out the first demo footage, all characters, enemies, levels and music 100% human made!!
Programming: me
Lead artist: Paulo Italo Medeiros
Artist: Thiago Limaverde
Levels: @amaweks
Music: @RDiggoSilva#indiegame#retrogames#sega#goldenaxe#segagenesis
Konami Antiques MSX Collection por Mega Drive / SEGA Genesis - Goonies gameplay
A tribute to the MSX, my first computer, and to Konami, the company that arguably made the best use of this legendary 8-bit platform.
This collection brings 12 classic MSX games to the SEGA Mega Drive / Genesis, packed into a single 4 MB ROM using a custom conversion tool that recompiles Z80 code into native Motorola 68000 code (using Claude Opus 5.5).
In this video, I showcase the cover artwork for all 12 games, followed by a short gameplay demonstration of "The Goonies".
Included Games:
- Antarctic Adventure
- Comic Bakery
- Hyper Rally
- Hyper Sports 2
- King's Valley
- Knightmare
- Magical Tree
- Monkey Academy
- Mopi Ranger
- Sky Jaguar
- The Goonies
- Yie Ar Kung-Fu 2
This project is still under development, and there's room for further improvements. Some games may experience occasional slowdowns or frame skipping.
A little piece of 8-bit nostalgia brought to the Mega Drive!
PS.: The cover artwork was quite a lot of work. I had to make countless manual adjustments due to the color and resolution limitations. I won't pretend I'm completely satisfied with the results, but I'd rather move forward than spend too much time on details that I suspect only I'll ever notice!
youtu.be/6Fm-Cs1IWM0?si=eWQC…#SGDK#MegaDrive#SEGAGenesis
Passando a madrugada retocando caixinhas de jogos da década de 80 para fazer uma coletânea para o Mega Drive que acho que menos de 10% das pessoas terão interesse em jogar... Tudo bem, estou curtindo muito isso. Me sinto bem nostálgico mexendo com esse tipo de coisa; cada jogo dessa coletânea, em cada teste que fiz, me fez lembrar bons momentos da minha infância.
Sempre soube que o SNES conseguia; não me surpreende...
Estranhando por eu falar isso? Ora, nunca fui fanboy de nada; eu também gosto do SNES! Eu gosto é de consoles e micros de 8 e 16 bits, que venham todos!
** SEGA GENESIS - RUFF-N-TUMBLE WIP PORT **
I love Ruff-n-Tumble on the Amiga, its got it all - Amazing pixel art / aggressive Soundtrack / Fun but tough Gameplay to boot.
One of the negatives is the frame rate however, it targets 25 FPS, this was not uncommon for Amiga games , its understandable with the 5 bitplanes the Amiga needed to shift around ( supporting 32 colors in the game area ) , the more bitplanes the more processing that's needed to move screen/objects around.
So part of me always wanted to rectify the frame rate - and that's how this port came about !
How does the port stack up ?
- 60 FPS solid gameplay . All movements / animations / timings were reworked from 25 FPS of the Amiga to 60 FPS for the MD port , so the port just has a smoother feel to it. This wasn't a given by any stretch and streaming animation was a key issue , the first stage uses up to 1450 tiles in its map tiles alone , so once you get HUD / Video Planes / Sprite List & Tiles for player and player attacks theres nothing left for enemy objects and thats just the first world. So both layers had to shift to streaming scrolling tech i developed for SOTA, so things are only in vram if they are onscreen presently , once that was in tile usage on the 1st world dropped to 850 tiles ( from 1450 ) for the backgrounds, this left some breathing room for Enemies and effects. However it still wasn't enough vram without a strong DMA object streaming system to send the tiles to VRAM , a lot of optimisation went into it so the game could run 60 FPS will full animation and nothing cut . If DMA streaming lagged that means the frame rate has to slow so i didn't want that .
- larger viewable game area on MD : 200 pixels high + 24 pixels high semi transparent HUD vs 192 pixels on Amiga + 24 solid HUD . This was almost by mistake , I thought the Amiga port was using a 256 high pixel screen size so I designed the engine to get close to that on MD with a transparent shadowed hud for contrast so you could see through it a little. The MDs (NTSC) display is 224 pixels high and the Amiga (PAL) is 256 pixels high. The hud is 24 lines high but it overlaps on the MD but with transparency you can still see critical things in the hud area . What i didn't realise at first was the Amiga uses less than its native resolution which was common to save ram / blitter time etc. So hence MD comes out favourably here still as it was a bit overdesigned !
- Despite having 8x less colors in the palette 512 MD vs 4096 Amiga , the MD does amazingly well here . I also had to put both background planes into Lock-Step so both are layered 1:1 , which gives up to 30 colors per tile on the MD rather than 15 so it compares well to the Amiga original which can have > 15 colors ( normal single plane tile limit on MD ) . I think from the side by side comparison you can see there isn't a noticeable quality drop and since it looked amazing in the first place I was very happy with the result !
- PCM SFX are in an all working with similar quality to the Amiga , XGM2 driver is in use so 13.3 khz PCM samples are used , which is more than enough to do a good replication of the Amiga SFX. FM / PSG soundtrack is been worked on currently hence video has no sound at present. I'm confident with the guy doing the music that its going to be a very close match to Amiga. On MD we can use 2 SFX channels where Amiga used 1 SFX channel and would play the last SFX over the top of the current SFX , the MD one moves low volume SFX to one channel and higher volume SFX to another channel and mixes it up a bit .
- Controls , there are some nice options to utilise the extra buttons on the MD controller and that should help the gameplay a bit also.
- Size, did i just throw ROM at this to make it work ? not really . The Amiga has 1024kb of combined chip ram required for the game , the MD has 136 kb of Ram/VRAM so i couldn't compress and decompress everything the amiga could but i got close . Without optimal music yet ( it should drop overall rom size ) we are about 19 mbit on MD vs the 14 mbit version of the Amiga . So whilst there's a bit more rom been used on MD its still well below the typcial 32 mbit rom that's in common use these days and a lot of efforts were made to minimise rom where possible . It started out prior to optimising rom heavily at 31 mbits of assets, i could have left it there but thats not how i do things haha.
- The game is fully complete apart from waiting on Music and a few playthroughs to catch remaining bugs.
- I did use some AI to aid this project , its 20 levels - its a lot of work and its not happening without some help , I used code from my SOTA / Parodius and EBH projects to make the engine , there has been no compromises . I think you can see some love went into this and the end result is what matters / garbage in is garbage out and vice versa.
- Game is written in C/SGDK for some parts ( cold pathes ) but a lot of custom assembly / custom modules are used for hot path also .
- Once music is done it will get a full release !
#Amiga#SegaGenesis#Megadrive#SGDK