I've just coded the N106 sound chip for my NSF player but have no idea how it's supposed to sound. I didn't find many NSFs that used the chip. I'd like to find out if I'm getting it mostly right. I recorded it playing 1-50 - Mappy Bonus Stage from 2A03.org. Does it sound generally correct?

http://www.io.com/~greens/nes-emu/1-50-mappy-bonus_stage.mp3

I'd also like recommendations for N106 NSFs to get. I liked some of the Mappy Kids tunes and found Erika to Satoru no Yumebouken pleasant.

Sounded a little high-pitched to me. Are you using the right formula? The NSF format specs have a good freq formula for it... but the doc by Goroh on NESdev has a bad formula.


My personal favorites are:

DDS Megami Tensei II
Sangokushi 2 <-- be careful, there's a copy of this going around with a bad play address
Erika to Satoru no Yumebouken


and some of the homemade ones... I don't rememeber the name... I just have the filename:

nemesis2_ext
ed2_chite <-- definatly get this one
rs2_dun1


I believe they're all covers if that helps you find them.

Yeah.. .now that I listened to the NSF, it's definatly way high pitched.

You could always try playing the NSF in another player to hear how it sounds.

uploaded an example for you:

the nsf:
http://www.angelfire.com/sd2/disch/nemesis2_ext.nsf

how it should sound:
http://www.angelfire.com/sd2/disch/nemesis2_ext.mp3

Thanks for putting up a correct version. It turns out I was playing an octave too high because I had referred to Goroh's Namco 106 doc. Before that I had been playing two octaves higher, and I figured that the first fix I made which lowered the octave must have been it. It sounded pretty funny. So it seems there is no octave doubling feature at all. Is there also no "sound stop" as mentioned in Goroh's doc? (high 3 bits of channel register 4)

The sound reminds me a lot of TMNT II for Game Boy. I really liked some of the tunes for that. I'm pretty sure the Game Boy hardware has a similar wave table, but maybe with only 1 bit of precision per sample. I like the sound of it and the simple operation (i.e. it doesn't have a linear counter from hell). The saw waves make the APU's square waves seem really rough in comparison (the Gradius cover songs are a good example).

I've downloaded quite a number of NSFs and wrote a scanner to show the sound hardware they used and N106 and FDS seem the only common extension hardware. I count only two that use MMC5, and five that use VRC6/7. I suppose that sound hardware might get more use in the future. I have maybe five Namco games that use N106 sound, but 30 or so songs from 2A03.org that use it.

It took me a while to make sense of the frequency formula. I was trying to figure out how the hardware could would work to result in that behavior. I think the frequency lowering effect of increasing the number of active channels might be due to a "round-robin" sharing of a single D/A converter, where each channel gets a tiny slice on it and then the next channel is switched in, at a very high frequency (hundreds of kHz, maybe even over 1MHz). I know that this method was used to allow polyphony on single channel sound hardware (there was a program called Music Construction Set for the Apple II, whose audio hardware is literally a single output bit which the CPU can toggle, which played 4 or 8 voices simultaneously).

"I'm pretty sure the Game Boy hardware has a similar wave table, but maybe with only 1 bit of precision per sample."

The Game Boy doesn't have DMC. Instead, its triangle channel can be reprogrammed with any 32-sample 4-bit waveform, and the channel's output can be shifted right by 0, 1 or 2 bits, for 100%, 50%, or 25% output level. No, the Game Boy doesn't have saw waves as such, but some games (such as Panel Action Bingo) do reprogram the triangle channel with a sawtooth waveform.

The biggest problem with Game Boy square and noise channels is that they don't have software envelope control; though there is a decay rate register, the volume of a square or noise channel can't be changed manually while a note is playing; ADSR and vibrato are impossible.

That said, I don't plan on ever writing an NSF that uses external sound hardware, just because the 72-pin NES doesn't support it.

I got an actual copy of DDS Megami Tensei II in the mail today.

I'm pissed off. I can't get NO sound out of it. All I hear is drums and sound fx!!!! >.<

Are you playing MTII on the normal NES. If you are, then get a Famicom.

I wonder if Famiclones have the potential for custom chips.

I am playing it on an NES.

The only tune that sounds legible is the 2nd battle theme when you've grown older.

I know the Tri-Star for SNES works with extra sound, some other clone systems might also.

NES's sound I/O is on that unused expansion port.

Yes.

-- jdc

Oh btw, just in case y'all never knew what DDSMegaten2 looks like, I found the game's intro after you turn the machine on to be a little rather disturbing. I jumped when I first saw it.

Yo! I was just listening to the NSF of Erika to Satoru no Yumebouken, and I found something shocking about it. The music in the first stage of LJN's Karate Kid video game (the 3 tournament matches before going to Okinawa) is also in Erika too!

I'm very happy to say....

My Namco106 nightmare is over.

I got a Famiclone, and now I can hear that deep deep buzzing of the Megami Tensei II title screen music.