SID audio out pipeline standard?

[m00dawg]

Is the SID audio pipeline generally a standard way of doing things? In other words, would a similar design work on, say, the audio output pins of the 2A03 (NES)? The reason I ask is that I have found that the audio-out of the NES is *noisy* . The regular audio out has a discernible buzz. I thought it was due to a bad power brick but I have tried using batteries too and, while it's better, it's not amazing. I suspect my RF modulator section needs some love (such as new caps) but I can only get to the big cap which I have already replaced. The others are covered by the RF shield and I don't have the tools to desolder that thing. It could be something else too since I have grabbed the audio from one of the last resistors before the RF section and it still had some buzz to it.

I noticed that I tapping the audio pins off the 2A03 was orders of magnitude quieter. The tri/noise/dpcm pin does still have some buzz. I suspect that might be a grounding thing? The square wave pin sounds immaculate. To make sure I don't damage the chip, I was wanting to make a proper audio out right off the 2A03's pins as was as to capture audio in from the cartridges separately. That will not only help for mastering but, I'm hoping, should greatly reduce noise.

The audio in from the carts is kinda of optional. I was thinking about doing that to use the PowerPak for it's emulation of the extra sound chips (such as the VRC6). It seems to sound better than over software emulation. I suspect because it sounds grungy-er and generally more 8-bit.

Anyways, after reading online, I found a brief mention of how to output from the 2A03 that had some similarities the audio out of the SID (from the ucapps.de schematic) so I was hoping I could base my design off that. I did notice that the C5 cap was opposite to how I thought it would be. Could anyone perhaps explain why?

Thanks!

(P.S. if you want to know what the ultimate goal is for this, check out my band's page, victimcache.com. The NES audio there is using FamiTracker plus some mastering trickery. It sounds good but doesn't come close to the character of the real NES).

[m00dawg]

Aha! I found this: http://nesdev.parodius.com/NESAudio.gif (included in post). It has similarities to the SID audio stuff but is generally more complicated. I was able to get some knowledge from forum post. Specifically, the this link (thanks nILS!). But there's still some questions raised.

For one, if I'm having a problem with noise, should I even look at reimplementing the NES audio stuff directly if the current implementation seems noisy? I read that the 74HCU04 is really a hex-inverter but I am not sure what purpose that would serve. One schematic I found online (here) just uses a few resistors, but that seems too plain - don't you need at least one cap for protecting the chip and such?

[nILS]

That schematic seems like a 2-bit DAC/VCO with an output buffer to me, but then again I dunno anything about that kinda stuff. The inverter + RC is basically an oscillator... If you wanna give it a spin here's a simulation of it.

[Hawkeye]

nILS, the simulator is awesome! thx.

[m00dawg]

Hah I forgot about simulator, thanks nILS! Super helpful as always!

By removing the inverter, I noticed that the output level drops from 3.49V to about 1.5V. I did not notice any audible drop between the CPU pins versus the RF out though? I think I read online that using the hex inverter was sort of a cheap way to amplify the signal, but that doesn't make much sense to me :) Either way, I'll keep playing around with the simulator. I forgot about the gem!

There are some audioheads on the NESdev forums so I may give them a shout to see what they think.