Emulators VS. SID - Proof that Emulators suck

YouTube via xxxogchris. "C64 4 ever!"

Update via Gwenhwyfaer in the comments:
"A page in a similar vein here. But playing some of the examples on this page shows up something interesting - the filters seem to be rather more enthusiastic on the sidplay variants than on the SID variants. Now, connecting this with the fact that a non-trivial number of SIDs out there actually had knackered filters (apparently it's a bit sensitive), it becomes possible to wonder whether a piece of music composed on a hobbled SID will sound equally dreadful on sidplay and on a working SID, and can only be heard as intended on another hobbled SID... and the thing is that the sidplay2 developers certainly have access to known-good hardware SIDs (in the HardSID, for example, which sidplay2 also targets), whereas whoever did these comparisons probably only has their C64, whose SIDs might be hobbled.

So I'd say that whilst the evidence of a difference between the C64 used and sidplay2 is clear, when combined with what's already known about SIDs it doesn't necessarily show what the authors of the page and video believe it shows. Further data is required - anyone have a SIDstation or a HardSID for comparison purposes?

And then of course, there's the fact that sidplay isn't just emulating a SID, it's emulating enough of a C64 to control the SID too; and that side of things might be the source of the differences, rather than any inaccuracy in the SID emulation itself. Again, a HardSID would illustrate that nicely. In which case these serve more to illustrate just how much the overall architecture of the C64, rather than simply the sound-producing bits of the SID, has to do with the sound of the chiptunes in question - surely an interesting result, but not quite what they're claiming.

In short, sure, there's a definite difference there (and yes, the C64 does sound way better), but that doesn't necessarily mean what it's claimed to mean; we aren't provided with enough information, or enough points of comparison, to tell. Yet."

"It also turns out that sidplay2's filter response curve is extensively configurable, presumably for solving just these kinds of problem. I wonder why they haven't at least attempted that kind of configuration? One would think it'd be the obvious place to start... Also, sidplay2 can emulate either the 6581 or 8580 SID, which sound very different; by default, it tries to figure out which SID to use on its own. Perhaps these are simply examples where it guesses wrong, and needs to be told to use the 6581?"