I've been looking at headphone amps and have noticed some of them (Schiit Jotunheim, Oppo HA-1, etc.) have 4-pin XLR balacned outputs that work with headphones that can be recabled to work with 4-pin XLR. I'm a little confused by how this is supposed to help. I'm assuming the way a
cable like this one works is that there's one pin for the signal and one pin for shielding per channel (2*2 = 4 total). If that's the case, how is it any different from a TRS 1/4 or minijack, which also mixes the signals into a single cable at the last stretch? Or is it because there's extra processing within the dac/amp needed to convert the balanced signal to a TRS output jack which inevitably degrades it?
In my head I'm imagining if your source/dac/amp is fully balanced (internal design and otherwise) all the way through, wouldn't you ideally want to have each channel of your headphones with its own cable? i.e., one XLR connector per channel?
But I noticed with some units like the Oppo HA-1, the XLR balanced outputs are pre-amp, not post-amp - so that's not even an option.
Sorry, if it wasn't super clear this wasn't meant to be a troll post - I've been trying to read up on this and just haven't found any good resources yet! Would duly appreciate someone setting me straight here.
On the "old style balanced headphone" wiring with twin 3XLR only two of the pins are actually connected on each connector, for +/- for each channel. This allows for "balanced" transmission, wherein there isn't a common ground wire. On typical TRS you have +/+/- where each channel gets its own + and they share a ground (this may be three physical wires, or it may be four physical wires with two of them joining at the connector). There's supposed benefits to getting away from the common return like reducing crosstalk or whatnot (ditto for the 3 vs 4 wire thing), but generally speaking what "balanced drive" does is provide greater voltage swing for the load (usually you have the - side running inverted so instead of like 0-to-4V you have like -4-to-4V). If you look at the NuForce HA-200 as an example, in single-ended mode it provides stereo output, but in "balanced" mode you get mono output, with one of the output sections providing an inverted signal, and it increases voltage swing (and therefore power) significantly.
There's no "processing" going on here; 4XLR (or TRRS) is just a smarter way to cable this on a single plug. I don't know why this wasn't done historically, it just wasn't. I think Sony was/is the first to promote TRRS for balanced wiring on some of their mobile headphones and devices (and they've proposed this as an industry standard), I'm not sure where 4XLR came into the mix, there's also other connectors that've been used including some proprietary models for portable amps. If I had to guess, I'd say 3XLR was popular specifically because of products like the HA-200 (but we're talking ten-twenty years ago, when Headroom started doing it) and you were using two stereo amplifiers with a common volume control.
"Fully balanced" is well into the realm of marketing fantasia though. In studio/pro audio land, you want balanced transmission over long distances and/or in noisy environments for common mode noise rejection, where you have separate +/- and shield (all three pins are actually connected), but in home settings this is largely unnecessary (the runs aren't as long, and there's generally a lot less noise). It also doesn't matter how that transmission is achieved "in the box" because the whole goal is that your analog cabling isn't picking up noise (or I should say, is better rejecting the noise its being exposed to). The idea that you must have dual DACs and all that is not related to common mode noise rejection, but it does help to sell "higher end" products. A "balanced" headphone amplifier (as in, it has inverting complementary output stages to provide increased voltage swing) is not "balanced" in the same way pro audio wiring refers to the word (meaning common mode noise rejection). It may accept balanced wiring input, but it doesn't have to - there are balanced amplifiers that accept SE inputs, and SE amplifiers that accept balanced inputs, and neither is "wrong" or "weird" - it really just depends on the device. There are certainly advertisers that would have you believe the entire chain needs to be "fully balanced" in order to achieve some higher level of transcendence (or whatever) but that's largely fluff. A lot of lower impedance/higher sensitivity headphones also don't derive much benefit from "balanced" wiring because they don't need the voltage swing - the big market for that kind of wiring is usually low sensitivity and/or high impedance loads, and more recent planar magnetic headphones that can take gobs and gobs of power.
In balanced XLR there is a differential pair for each channel. Left channel has one + wire and one - wire, and likewise for Right. This is beneficial because the pair of wires can be twisted together so that the noise they each receive is the practically the same. A signal that is present on both parts of a differential pair is called a common mode signal. In a differential amplifier or a speaker, common mode signals cancel each other out, only the differential signal is amplified. The common mode noise is removed and you're left with the desired differential signal, so no shielding is needed.
Yes and no. You're right on what balanced wiring does in the pro audio world, but not in how its working with headphones. There is no "common mode noise rejection" aspect to the headphone wiring, because again it isn't about connecting separate hot, cold, and shield, its about running differential drive to increase voltage swing. It's much more similar to how speakers are wired - its not some twisted pair common mode thing, like you'd see in signal wiring (and you generally don't want twisted-pair wiring for transducer loads because it has a lot of inductance and can be really vile for the amplifier to drive). Shielding is not properly part of this discussion - in studio/stage wiring you still have shielded wiring (in many cases it has 3-4 layers of shielding, and is still balanced transmission) because again the idea is to kill the noise, but for headphone (or speaker) wiring you aren't running shielded wiring anyways, and there's no need for it. Generally speaking I would say "yes please" to any shielded signal wiring, be it SE, balanced, whatever - its never going to hurt you and it isn't very expensive. But "balanced headphones" were not originally done as a means of "getting rid of shielding" nor does balanced signal wiring negate the use or need for shielding.
EDIT
Zeus be praised! The original HeadRoom write-up on balanced headphones is actually still on line:
https://www.headphone.com/pages/balanced-headphones-guide
And as far as I know and have ever heard, they are entirely in claiming they invented this. I'm not sure why they chose the connections they did, but I would venture their choices are what influenced things for a long time. The link has pictures that show exactly what they're trying to achieve as well. Keep in mind: at its core, this is marketing literature, so its ultimately trying to sell you something (in this case something that Headroom no longer produces), so there's bias as a result of that.
EDIT 2
The HiFiMan cable also shows another example of how balanced transmission can be achieved - with each channel on its own TRS plug. This is pretty common in studio gear to use TS for SE mono and TRS for balanced mono. Especially if XLR is too big or clunky.
But I noticed with some units like the Oppo HA-1, the XLR balanced outputs are pre-amp, not post-amp - so that's not even an option. 
