36 replies to this topic

[Nintendo Penguin]

Just curious, but how come nobody ever makes new NES games? Is the NES harder to make games for? What would be involved?

Thanks!

[ArcadeAction]

See www.nesdev.com . I am busy on my second project to learn 6502 (NES) Assembly, but it will take a while before I'd be able to write real games.

[batari]

Nintendo Penguin, on Sat Oct 29, 2005 5:23 PM, said:

Just curious, but how come nobody ever makes new NES games? Is the NES harder to make games for? What would be involved?

Thanks!

I've never programmed the NES, but according to others, it's no harder to program than other 6502 machines.

However, compared with other machines, it can be very difficult and/or prohibitively expensive to make carts of your game. And releasing your game on cart seems to be a primary motivating factor for writing the game in the first place.

As I understand, the carts are not just an EPROM on a board. Even the garden-variety NES games require a lockout chip and mapper. The lockout chip can be pulled from other carts, but the mapper is an ASIC that is far from trivial or cheap to design/build for a single game produced in small quantities.

Despite what I said, I understand that you can recycle parts from some early common carts, like Duck Hunt, but nobody ever seems to do this, perhaps because you are limited to the mapper in that particular cart, which is pretty simple. This doesn't seem to be a big deal to me because I still think simple games are fun. I can only speculate that it's because the NES seemed to be more popular with a generation that expected cooler graphics and bigger levels than us 2600 gamers, and maybe they don't want to make simple games, or maybe nobody would buy them if they did...

[Memblers]

There's a lot of people writing NES stuff. It seems that the games that are beyond the simple stuff rarely or never get fully developed though.

I'm making NES carts also, I have 2 different versions of Squeedo available (only has bare PCBs for now, Midines uses these currently), and a simpler board design on the way.

I've also released a multicart, "Gargage Cart" with Munchie Attack, Hot Seat Harry, and Solar Wars. It was limited to 24 carts though (mostly since I was using recycled PCBs). A sequel is being worked on.

[Bruce Tomlin]

For smaller games you don't need a mapper.

If I were doing larger games, I'd probably design my own board and come up with my own mapper on a 16V8 or 20V8 chip, preferably copying an existing mapper's behavior.

But yeah, the CIC chip is the easy part. It (and the extra cart shells) is part of the reason that I'm snagging just about any common NES cart that I find for a dollar. Once I'm ready, I should have a nice stockpile of carts and CICs. But there is still a lot of potential left to be exploited in the 7800 and the Colecovision.

[Memblers]

You probably know already, but be aware when stocking carts for CICs that the most common ones use globtops. Especially SMB/DH, and SMB/DH/Track Meet. I'm still buying those however, since my newer cart will pass it's CIC signals through.

[~llama]

I tried NES programming before I even owned an Atari and couldn't figure it out, but it was because I didn't know 6502 and was just parroting the various tutorials I found. I'd like to return to the docs and actually gain an understanding of how the hardware works but I don't really have time. &%#$ college work.

[TMR]

i have a fully functioning game engine for Reaxion NES but haven't had time to finish it... t'be honest, part of the problem is that i haven't found a decent tool for half the jobs i needed and indeed still need to do so a lot of the data is cobbled together using C64 programs and a lot of the time best guessed...

Really must sit down and draw that titles logo...

[Shawn Sr.]

Mario Open Golf and the version of Mario 2 with all 72 pins on the connector are great ones for larger games.

Grandtheftendo anyone? I cant wait till Brian finishes that. In case you don't know about it is GTA 3 fully ported to the NES, The map is bigger than metroid. Just picture playing GTA 3 with the style of game play of GTA 2 and that is what Grandtheftendo is.

[~llama]

the problem I really ran into with NES programming is that everyone doing real work has their own custom toolchain for dealing with the graphics and maps and in some cases custom assemblers, and none of them want to release their tools until their game is done, so there arent any really amazing public tools for that stuff. There are a few exceptions but its nothing like, say, Game Boy or Commodore where all of the graphics tools are freely available and easy to use.

I didn't want to have to write my own tools to use before I even started working on the game, so I didn't bother with NES once I got a simple sprite engine working.

Once the toolset is mature and publicly available in one place, standardized like DASM is for VCS (and Channel F ) then I think NES programming will be a little harder than it needs to be.

[Bruce Tomlin]

~llama, on Sat Dec 24, 2005 4:28 PM, said:

the problem I really ran into with NES programming is that everyone doing real work has their own custom toolchain for dealing with the graphics and maps and in some cases custom assemblers, and none of them want to release their tools until their game is done, so there arent any really amazing public tools for that stuff. There are a few exceptions but its nothing like, say, Game Boy or Commodore where all of the graphics tools are freely available and easy to use.

I can understand that. There is a very good reason why I wouldn't release certain kinds of custom tools until I was finished with a game, and that's because I wouldn't want the trouble of supporting the damn thing when it's already enough work to make a game.

At least I don't need much of a toolchain for Colecovision... all the sprites and tiles are 1-bit deep, so I can just eyeball them from the Photoshop window and type in the hex codes directly. The 7800 is a total pain in the butt to do this for, and eventually I came up with a crappy perl script to convert graphics using ASCII like # . and X characters into hex codes, along with the macros to separate the scan lines.

When I need a bigger map wider than 32 tiles for my RPG, I'll probably be doing another cheap-ass perl script to create the DB statements for it.

[brpocock]

Bruce Tomlin, on Sat Dec 24, 2005 9:50 PM, said:

The 7800 is a total pain in the butt to do this for, and eventually I came up with a crappy perl script to convert graphics using ASCII like # . and X characters into hex codes, along with the macros to separate the scan lines.

I don't know anything about the 7800's art formats, but point me at some specs and I'll get you a PNG-to-7800 converter. NetPBM and Perl are your friends.

[sack]

I started building a devkit over the summer. Basically the plan was to take a cart, run the ROM lines running to an edge connector on an FPGA board with 512k of memory so i could run it as a mapper emulator. Other idea was that on pressing a switch I could make it bank over the 512k flash with a debugger and get the machine state back.

started a degree since then so not had time to finish it I'm afraid...

[Pixelboy]

Seems like the NES homebrew scene is still in its infancy, and is in no hurry to pick up steam, even though there's a wealth of documentation on the net for the hardware. I'm sure it will grow in popularity once someone comes out with a useful and easy-to-learn dev tool.

We should remember that in the beginning, the Atari 2600 homebrew scene revolved mostly around getting binaries to work on emulators. It took a while before homebrew cartridge production got organized, which helped the scene to grow even more. I think that, while the NES homebrew scene holds significant differences to the Atari (and Coleco) homebrew scene, it needs the same kind of impulse to make it grow. I'm sure it will eventually happen, hopefully sooner rather than later.

[Cybergoth]

I've been doing some work on porting "Park Patrol" to the NES, but nothing ready to show yet.

[Allan]

brpocock, on Wed Jan 18, 2006 4:29 PM, said:

Bruce Tomlin, on Sat Dec 24, 2005 9:50 PM, said:

The 7800 is a total pain in the butt to do this for, and eventually I came up with a crappy perl script to convert graphics using ASCII like # . and X characters into hex codes, along with the macros to separate the scan lines.

I don't know anything about the 7800's art formats, but point me at some specs and I'll get you a PNG-to-7800 converter. NetPBM and Perl are your friends.

Yea, the 7800 graphics info is mixed together with the color info making it difficult to convert pictures in drawing programs into data that the 7800 can convert into graphics. It would really help 7800 development if there was a program to do that.

The way GCC did it was on the St with Neochrome and then (I guess) they just did the calculations by hand. Here's the Robotron Neochrome pic they used as a template. I was using this file to make some Super Pacman graphics.

Allan

Attached Files

•  7800_SUPER_PAC_MAN.zip   3.63K   123 downloads

Edited by Allan, Fri Dec 22, 2006 8:59 PM.

[Bruce Tomlin]

I suppose it is a little surprising that the NES isn't getting much homebrew love these days. But there really are a few issues that are probably stumbling blocks for your average homebrewer:

* Who is selling cartridge boards? Nobody that I know of. You can scavenge some, but you have to find one with the right mapper.
* Fresh boards are a bit more complicated than for other systems, because of the metric pin spacing, the double bus, and the security chip.
* Old toasters have bad cartridge slots, meaning more trouble getting games working.
* They need a double bus, which means at least two ROMs, and the CHR data has to be maintained in a separate file.
* You need a RAM chip to get more VRAM. The standard 2K limits you to fixed tile/sprite games, and is the one point where NES is clearly inferior to Colecovision.
* You need to choose a video mirroring mode.
* You need a mapper to go beyond 32K code, and you have to choose one of the many out there. A few mappers could be implemented with just a GAL chip, but those are probably the less common ones.
* The mapper usually needs to control the VRAM too.

Still, that's no excuse for people not doing simple 32K or smaller games and developing them on an emulator, so I guess it's maybe not that easy to code for after all. I mean, it was a good machine for its time, but are its trade-offs still worth it?

The killers for me are the cartridge boards and the fact that I really don't like writing 6502 code. Now that I've started on Sega Genesis, I think it's much easier to homebrew for than NES, or even anything else. Am I just ahead of the curve, or will the NES end up getting skipped for homebrew?

[Miqorz]

The guy who created Grand Theftendo also made a C compiler which assembles to 6502. If I get some time after I'm in college I'm SO buying the hardware to setup a dev system!

[thegoldenband]

Tangentially, did Grand Theftendo ever get finished, or did a playable ROM get released at least?

[Artlover]

thegoldenband, on Thu Aug 30, 2007 11:53 PM, said:

did Grand Theftendo ever get finished

Yeah, it was finished a while ago, but the guy went back on what he said and never released it, instead screwing around with unneccessary custom mappers till such time as his websites went down (suposidlly because of a C&D from Rockstar or something to that effect).

All things considered, I think the whole thing was just a scam from the getgo.

[yuppicide]

What about Grand Theftendo? That game looks great.

http://www.grandtheftendo.com/

.. damn nothing on the page anymore. Here's some links to pictures of it:

http://www.gebh.net/...mo/20041028.gif
http://www.nilsth.com/pics/gta8bit.gif
http://www.emunova.n...ad449ea843d.gif

Edited by yuppicide, Sat Sep 15, 2007 1:26 PM.

[Bruce Tomlin]

Does anyone have any evidence that GTT was anything other than a hoax? All I've ever seen are some static shots that could have either been made in Photoshop, or screen shots of simple demo code in an emulator.

Then the guy conveniently drops the project after it was supposedly "finished", with no intermediate versions ever released. We haven't ever seen that happen before in the internets, have we?

[Atariboy2600]

Well here some homebrew roms for the Famicom/NES some games are small but fun.

http://rapidshare.co...2/fcfr.rar.html

the roms are from mini cd that came with a Japanese book on Famicom title is "Fami Emulator Kanzen (=Perfect) Guide 2004".

The book comes with a CD. The CD contains some emulators (e.g. VirtuaNES version 0.92, NNNesterJ 0.23, FCE Ultra v0.98.9 and some more), some free roms (no official FC releases of course) a rar file out of the freeroms folder on the CD. It contains some free roms (some of which you might already know) and is 1.16mb.

[dwhyte]

Well... reviving an old thread here, but a lot of dev'ing is going on with BatariBasic... I haven't seen anybody mention the NES Basic compiler... I've just got neshla, nesasm, and nbasic installed on my linux box... I'm thinking of tackling the 2A03 and PPU once my A8 project is done... though by the looks of things I'll need to make some tools before any nes coding begins...

Edited by dwhyte, Sun May 4, 2008 6:30 AM.

[Vigo]

yuppicide, on Sat Sep 15, 2007 9:24 PM, said:

:xWhat about Grand Theftendo? That game looks great.

http://www.grandtheftendo.com/

.. damn nothing on the page anymore. Here's some links to pictures of it:

http://www.gebh.net/...mo/20041028.gif
http://www.nilsth.com/pics/gta8bit.gif
http://www.emunova.n...ad449ea843d.gif

Unless he used the MMC5 Mapper (which I very much doubt), I am 99% sure these Screenshots are fake. Why? Look at the score bar. You have a grey time indicator, next line is an green cash indicator and next line is the orange health status. The NES is not capable of doing that, because colour attributes can only affect 16x16 pixel areas, not 8x8 like in this screenshot. Unless you use MMC5 Mapper. Another example are the blue windows in the right house. the 2 shades of blue collide with the 2 shades of grey. Or the orange parts on the left building colliding with the blueish colours of the rest. In the next pictures, exactly the same, he has drawn the graphics using 4 colours distinct colours per 8x8 tile, which is impossible on a standard NES.

Edited by Vigo, Sun May 4, 2008 10:10 AM.