![soundhack for windows soundhack for windows](https://static.gameguard.ac/media/gameguard-6f2e0960a2.png)
Using digital technology to simulate discrete analog components is nothing new.
![soundhack for windows soundhack for windows](https://www.soundhack.com/sh-content/uploads/2011/02/Screen-Shot-2015-06-24-at-11.27.30-AM.png)
Clock speed is also a major difference between the FPGA side and the Arm side, but again the FPGA can do things in parallel giving it a massive potential nobody has said OR implied that to the best of my knowledge, but it's quite correct that it ABSOLUTELY _can_ be done in the same manner it was for MAME, as I pointed out over a year ago.
![soundhack for windows soundhack for windows](https://rekkerd.org/img/201107/soundhack_pitchsift.jpg)
#SOUNDHACK FOR WINDOWS CODE#
If the Arm side of the MiSTer is used, then the Mame code could be almost used directly, but I don't know if the Arm side is fast enough, and it is using a computer instead of dedicated FPGA so it is theoretically less efficient. As I mentioned previously, a lower sample rate could be used if the computing power requirements are too high for perfect sound. I suspect that the relatively massive FPGA in the MiSTer could be programmed in a way to specifically handle the Donkey Kong math very efficiently using more of the FPGA to get the job done quicker. I don't know how the Arm compares to a Pi 3, but only a subset of the Mame code would be needed.
#SOUNDHACK FOR WINDOWS PC#
This led me to a short time of using a PC instead of the Pi and eventually I discovered the MiSTer and I haven't really touched Mame since, except for a few games that are not on the MiSTer. I have compiled the newer versions of Mame on my Raspberry Pi 3 (and maybe also the 4 but I don't remember) specifically for the new Donkey Kong sound and when I played the game, it was clear that the Pi was not fast enough. The big question in my mind at this point is how much processing power does it take and how powerful is the FPGA compared to the Arm side of the MiSTer. I agree that using the Mame source code as a resource makes total sense given the amount of effort that went into it. So far no experts have weighed in on that.
#SOUNDHACK FOR WINDOWS DRIVERS#
Info on how analog audio simulation was only added in in MAME 0.224 about a year ago:Įlsewhere in the forums I've been advocating for the use of the new MT32-Pi-Lite official board (which attaches to a Rasp Pi) to be used to basically run MAME's audio drivers and pipe the sound back to the DE10 to be mixed in, similar to how MT32 does it.
![soundhack for windows soundhack for windows](https://images-na.ssl-images-amazon.com/images/I/61PwaF%2BpYrL._AC_SY355_.jpg)
Maybe you remember using Multisim or something back in school where doing a transient analysis of the first 1/10th of a second required 10 or more seconds to solve. MAME itself has what is basically an entire SPICE solver that can simulate the entire audio schematic continuously, which requires a lot of CPU power. You will notice that it starts off with a list of nodes and then the exact components from the schematic. Roland's Aira line use ACB (analog circuit behaviour) which is basically what MAME uses, and what we should be talking about for DK. Everything it can do is based on its giant WAV memory. The Fantom (I had a Fantom G6) is PCM sample based. These make sounds directly like a Jupiter 8, and are mixed in with the digital sounds (from the MB8884 after the DAC). In the bottom right, you'll see an NE556 surrounded by a lot of analog buffers, resistors, transistors, capacitors. Look at the DK schematic, TKG4-14-CPU page 3.