Could someone go a little more in depth on why there's a 1.65V DC offset on the headphone jack?
In what situation would a DC offset be high quality? The only way I can imagine it helping is to reduce the available headroom and thus make the output quieter. I know very little about electronics, so I might be missing something.
since the headphones will only be connected to the axoloti they don't "care" at what voltage level gnd is. by creating a virtual gnd at 1.65v the amplitude can swing both positive and "negative" (below 1.65v), thus providing a clean signal with maximum headroom. the downside is, that you should not connect the headphone out to an amplifier or anything since you will likely share gnd (and thus shorten 1.65v to 0v)
well, there is no dc-offset for the headphone. you do not apply a dc from gnd to signal. gnd is just not at 0v. so instead of -1.65v 0v 1.65v you have 0v 1.65v 3.3v. the headphone has no concept of 0v being gnd.
My advice is to google "can headphones withstand DC offset?" If that didn't persuade you already from ever use Hp out in anything but "free" headphones or oscilloscopes. think about a headphone plugged to a constant 1.6v dc.. that would be "silence". So, even when you are not sending "sound" out, the membranes of the headphones are still being pushed forward (generating heat in that delicate coil). Luckily. the default configuration for headphone out its pretty attenuated (-54dB). but if you push the output up to 6dB. then you will apply to the membranes of the Hp up to 3v (in the same direction!) and that's way more than most Hp preamps do.. My soundcard headphone out gives at full level 2vpp. that's just 1 volt maximum in the same direction, and that is still less than what hp out gives in "standby".. HP out makes for a nice "almost eurorack" Cv/gate output though..
sorry but you did not get the point at all! as i described earlier there is NO dc voltage present on the headphone coil. NO DC OFFSET! it is just that ground is not at 0v. but the headphone has no way of knowing this, and since "silence" is also at this virtual ground, the coil will not pop inwards or outwards. does that make sense?
Alright. So the reason the headphone out is for headphones only is that if you connect it to a powered speaker, bad things with grounding might happen? Is that what a ground loop (50/60hz buzz) is?
if you connect the headphone out to a system with ground at 0v, these two grounds will get shorted obviously. this is of course very bad, since they are not at the same voltage.
ground loop is here: https://en.wikipedia.org/wiki/Ground_loop_(electricity)
No, there is no Dc in the coil until you plug the hp in the axo Hp out.. the headphones HAVE a way to distinguish between potential and non potential (0v). IF you have an oscilloscope. plug Axo hp out to it and you will understand what I'm talking about..
If you read the articles regarding dc on headphones, it says clearly that it shouldn't be over 20 mV. DC offset and Ground loop are different problems, specially regarding their effect on speakers.
can you post a link to an article? all i see are forum comments.
my understanding is that there is no dc offset, since the ground is virtual and from virtual ground to "silence" there is... well 0.00v!! but maybe i am totally wrong 
so please @johannes can you clarify?
what are you measuring with your oscilloscope? from virtual (hp) ground to signal of hp?
ok, can you post a picture where you use hp ground as zero reference and signal (left or right) as display signal?
have you set your oscilloscope to use the virtual headphone ground? if not, well then the offset is to be expected. again, this has nothing to do with dc offset of the signal, since the gnd reference is shifted. in accordance with this shifted gnd the signal is perfectly at zero offset.
