Testing sound of different MSX sound chips and their FPGA implementations

It doesn't look (or rather sound, to be precise) like just filter issue for me. F.e. the high-freq fuzzy sounding of MFR PSG seemed to be similar to same thinig with the MSX-AUDIO emulated in GR8NET if you compare to the real Philips Music Module. And Eugeny said that this probably implementation-specific issue with DAC frequencies resonating with generated sound frequencies and sent me this link: Nyquist–Shannon sampling theorem so it might be much more complicated than just low-pass filtering. Actually, i'm going to do some comparisons to SCC and OPLL implementaions as well, but i need to select the tunes which will use most from those chips, to highlight the possible performance issues. Any suggestions, maybe? :)

Yes, maybe, i will also add sound from Zemmix Neo when i'll get it back soon. I'll do a second round with other chips - SCC and OPLL when i get my Neo handy And will include emulators (need to figure out how to record emu sound directly, without DAC/ADC, i guess we don't need to pass it over real soundcard for emus)

Ok, that makes sense. So i will be routing sound from audio-interface output via mixer and back to the audio input Funny, but should work. Just need not to forget to switch off the input monitoring (which feeds some input back to output) to not make a nasty loop

Ok, i will do both for emulators - with and without passing coundcard+mixer. Actually recording with just soundcard takes less time for me, as i don't need to install some virtual sound devices to my PC. That UCA202 interface is not super-high-end but it is semi-pro, so DAC/ADC should not add much distortion, but i wonder about mixer though.

In my test MFR was connected to Modulon slotexpander, and the expander itsefl has RCA jack soldered instead of pin 49 connection to computer. So it only has been affected by Modulon circuitry, not the computer audio circuitry. The C2 has also pin 49 cut off and the 3.5mm audio jack soldered, so C2 sound was not affected at all (it has jack output on PCB so probably i was not mandatory to cut the pin off, but i did this cut off to reduce possible additional noise coming from that pin from computer btw)

Only Zemmix Neo (i got no OCM and no real korean Zemmix) when i receive it back.

You have probably reversed the carnivore file with the MFR file.

My top is that:

1. YIS503
2. FS-A1WX (filter too strong and frequencies a bit distorted)
3. Carnivore (good filter but frequencies distorted)
4. MFR (bad filter, we can hear the data transfers and frequencies very distorted)