If it's ultra sound what they hear I can't imagine how much of a high pitch it'd be like and how anyone could consider it appropriate for music xD

Ultrasound to humans, but not for a wide variety of other mammals!* To them, music aimed at a relatively limited frequency range of ~20 Hz-20 KHz might well sound unnaturally muffled. Then again, that may be a matter of taste even for her fellow beastfolk ;)

* (Sources I can find on a quick search vary, but place racoons' upper limit between 65-80 KHz, anywhere from an octave and a half to around two whole octaves higher.)

The real question: what kinda crazy über-tweeters does she have in that stack!?

Oilbirds apparently

Ahaha XD

On the other end. mp3's will sound absolutely horrible to them. Mp3's achieve their compression by removing sound data, aka frequencies, Humans physically are unable to hear.
So ignore the audiophiles who claim they can tell the difference between a raw audio recording and mp3.

(That does depend on the bitrate of the MP3 and the specific audio content - it's totally possible to tell the difference at 128 kbps in e.g. classical music recordings, and hard not to notice below that - but it's been a long, long time since people were that hard-pressed for space with their MP3 collection.

It's also interesting to consider how much high-frequency content music created by humans for humans actually has. As an experiment, I once took 96 KHz remasters of a couple favorite albums, opened them in Audacity, and played them back at half-speed, moving what normally would've been at the edge of my hearing into the audible range.

The results were interesting; drums definitely had some higher frequencies remaining, but the electric guitar was distinctly muffled. I s'pect that has a lot to do with the fact that the drums would've been directly mic'ed, but the guitar was likely recorded from a mic pointed at the speaker of the guitarist's amp, and guitar amps are designed for a specific human-range frequency response in terms of both electronics and cabinet design.

But even the tape used for recording has its own response characteristics - which makes you wonder what a live performance would sound like to a non-human listener, vs. a recording...)

In a story i wrote, i had them hear the sounds of stuff just 'cut off' + noise from the compression. To get an idea of said noise, on a *nix(unix, linux or bsd) system. do a 'cat /dev/random | *sound program of your choice*' That's if i remember my bash correctly..

Ah, now that's an interesting point - digital aliasing noise is more specifically a kind of mirror-image of the original waveform at a higher frequency, but MP3 artifacting is something very different and I'm not sure if anyone has formally described it in terms of how it relates to the source material...

On the point about ultrasonics being included in high samplerate recordings: Ehh...sorta. If you look at the datasheets for some popular audio ADC chips themselves, that support 96kHz and 192kHz, and understand how these things actually work, you'll see that it's not that simple.
https://www.google.com/search?q=cir.....+datasheet+pdf

44.1kHz and 48kHz share a mode, 88.2kHz and 96kHz share a different mode, etc. Within each mode, you just change the clock and that's it. Between modes, you change a clock *divisor* so that the actual ADC part is still running at the same mid-MHz rate and far fewer bits than the final output. That mid-MHz rate, not the final mid-kHz output rate, is what the analog anti-aliasing lowpass is designed for, which makes the analog front end much cheaper! (actually brings it into the realm of reasonability, when you understand analog design)

The converter also adds some ultrasonic noise, on purpose, to guarantee a wiggle in the lowest real bit or two. Then the fast and shallow signal is digitally lowpassed (still in the converter chip) to remove that ultrasonic noise and convert it to more resolution, and to anti-alias the final mid-kHz output rate. (just one digital lowpass does all three jobs simultaneously) Regardless of mode, or clock divisor, the turnover frequency of this digital lowpass is always just a hair above 20kHz. The difference for higher sample rates is a shallower slope, meaning that it rolls off more gradually beyond the turnover frequency, *but it's still rolling off there*. Then the converter chip picks out samples from that internal mid-MHz stream to send out at the final mid-KHz rate, and throws away the rest.

The reason to use higher mid-KHz sample rates is *not* to preserve ultrasonics, unless it's a specialty converter that is not normally used for (human) commercial audio. (humans analyzing bat calls, for example) The reason to use higher sample rates for human-enjoyed audio is to:
A. Satisfy a human market that has no idea what it's talking about (audiophools) and
B. Reduce latency for live processing (concerts, etc.), because the shallower slope of the converters' digital lowpass also introduces less delay.

According to the datasheets at 48kHz, the claimed ~1ms round trip (analog to analog) through a Behringer X32 digital live console comes entirely from the converters at either end of the process. So all of the X32's documented processing must be one sample at a time, all the way through, and delivered to the output converters before the next sample arrives from the input converters. With the speed of sound in normal air, that delay equates to the speakers being about 1 foot or 300mm farther away than they really are.

Yeah, that's not a huge shock :/

I imagine that an anthro world, or even a feral one where every species has the same mental capacities as humans, would still do that, except the 20kHz figure is much higher and everything else follows that. In which case everything really does preserve ultrasonics for humans, with no need for that particular "audiophool" argument. (whoever has the highest threshold might still have it though)

Lossy compression (mp3, etc.) would be much harder, if it even exists at all, since different species would likely notice different things. So what sounds identical to one with a bunch of stuff missing, would absolutely not to another.

Could make for an interesting story, if it weren't limited to a technical audience that directly understands all of that:
- Maybe one species invents it first and calls it good, while most other species strongly disagree...and then someone figures out (or finally takes seriously) that it doesn't preserve everything that *everyone* can hear...
- Maybe there are "budget rigs" that are aimed at specific species, so one of those meant for humans sounds muffled to a raccoon and tinny to an elephant. Meanwhile, its human owner is completely oblivious to its shortcomings...
- Maybe a song meant for bats gets encoded at 48kHz somehow, and half of the melody disappears because it's above the required brickwall lowpass between 20k and 24k...
- Professional live sound design, including the PA and acoustics for such a wide range. Maybe the mixing desk requires multiple operators of different species and hearing ranges, etc...

Lots of things to play with here, if you really want to take it *that* seriously.
And likewise for sight, smell, and possibly other senses.........

Radar ears!

Useful as well as decorative!

So many new experiences to have!

Sure, you may have had your species & gender identity scrambled and gotten shrunk to kid-size on top of working food-service in a strange and run-down part of the city, but at least you've discovered new aspects of music you may or may not even like...!

YEAH!!!!!

Reminds me of back when I could scream like a girl and would frequently go into ultrasonic frequencies. That was fun on the vocal cords...

Li'l-kid screaming is a whole category unto itself... o_O