How is that even possible though? The Aquarius runs at about the same speed as a Speccy, doesn't it?

Unsure how and no idea. The few times I managed to get it running last week, it was actually sounding way too slow.
And this is after I removed the few opcodes of the ZX keyboard read routine. It was still sounding slow so I removed the two spare timing NOP's within just to see how much affect that it would have; and it did near to nothing.

But then as I ran it again last night in the same emulator (no changes to the emu), I was incredibly surprised to hear it running at the speed that it is.  And this was without the keyboard, but with the two NOPs in the inner loop.

(For those playing along at home, it's the SquatM engine.)

Identical as far as I'm aware - Both are at 3.5 mhz

Will keep fiddling with the sprinkling of NOPs to get it sounding nicer.
However I am now wondering if it is the emulator that is a bit dicey, and isn't playing nicely.
Unfortunately I don't have an Aquarius computer to try it on.

https://youtu.be/y0M9qiRdWEk

https://youtu.be/y0M9qiRdWEk

I sort of managed to get some of these tunes working / not working simply as binary files, and just needed to load them in as such. VirtualAquarius emulator allows the loading and execution of a binary at whatever address you give it - (stating the obvious : same origin address as when it was assembled)
Those tunes that write back to memory, of course, will fail, since Im trying to run these in the cartridge ROM area. 

CAQ cassette header:
Further looking and fiiddling has finally found the CAQ header, and a somewhat self creation of a BASIC loader for the binary. Just type RUN after CLOADing the CAQ file in.

Found the CAQ stuff within a few sub-directories of James Tamer's VirtualAquarius emulator.
The header is here for future reference :
https://github.com/bushy555/Aquarius_1- … uarius.asm

...however in saying this, Shiru's "LetsGo" is ~15k, and I can't get it working with this header, either in VirtualAquarius or MESS emulators. This CAQ header though seems to work fine in many other programs - perhaps it has an 8kb limit?
I've managed to sort of getting "LetsGo" working in the VirtaulAquarius emu just by making its origin at $4016, and executing it here. I'll attempt to get a video of it when I can get it working next.
(There are BASIC pointers and variables stored at $4000-ish, hence the $4016).
Weirdly though it works maybe 2 out of every ten times I try to load it in, even after killing the emu and restarting it.
Clearly Im still doing something wrong somewhere.  Doesn't matter - I love me fiddling around.

I've managed to do about 15 tunes from about 11 or 12 engines.
https://github.com/bushy555/Aquarius_1-bit_music

Two locations for the Aquarius rom cartridge : $e000 for 8k roms and $C000 for 16k.
I can only get the combined engine + music that is in total 8k to execute in any emulator.
Up to 16l assemble fine of course to $c000, but just wont run.
Typical normal binaries assemble to $4000 - and I can't get these working either.
I want to say that, initially, the Aquarius seems very "finicky" and very particular on its needs. But it will just come down to my lack of knowledge and understanding of this computers setup.

Ive been using an old emulator 'VirtualAquarius' : http://www.oocities.org/emucompboy/
Or, roms for mess/mame can be grabbed here : https://github.com/bushy555/MESS_MAME_ROMS

I'd youtube a few tunes, but, I don't think that its really worth the effort.

The Aquarius computer has both an internal speaker as well as the fancy AY chip.
Internal speaker hangs off port $FC.

Very very small steps, but managed to get ZSPEAK and Beepola's "1943" song within the Huby engine playing through the internal speaker in two of the Aquarius's emulators.
Have spent a few hours trying to get Phaser 1 / Buzzkick / Squats / Quad / WTBeep / musicbox ...with no luck yet.
Biggest issue is/was finding the cryptic cartridge header at $c000 and the m/l BASIC loader (and/or) the header of the Aquarius's emulators tape format being  .CAQ
However, have just found more details, so I might be getting somewhere now.