@A J Stiles -- Re: XLR Balanced inputs look the part but are they?
"I'm convinced that balanced lines are now a bigger problem than the problem they were originally intended to solve; which was noise pickup on small signals."
You're of course correct. Balanced lines are best confined for use with dynamic mikes and specialized instrumentation that use sensors which have very low-level outputs. I actively discourage the use balanced circuits; it usually complicates things to no real benefit. (Balanced audio became popular in HiFi as it was hangover from both telephony and the days when valves predominated. Not only were audio circuits balanced but also so were AC valve heater circuits to minimise induced hum (here, a humdinger potentiometer would ground heaters at a minimum hum point). With solid-state, balancing is usually unnecessary as there's less noise pickup: heaters are gone, so is the very high input Z of valves and shielding is easier.)
I also agree with you about the impedance matching. Except for very small noise-limited circuits, RF receiver inputs etc., baseband video over coax and in transmitter-antenna power coupling and such, Z-matching is a waste of time. The once common bridging/terminating switch that chucked a terminating load across the input of audio lines [and which screwed levels up] is, thankfully, mostly gone (as S/N is now normally limited by other factors).
There is, however, one audio application where +4 or +8dBm line levels should be balanced and that's where the lines are long such as in radio and TV stations and concert halls etc. I recall an installation where about 17kms* of audio cable were used and it would not have been possible to run unbalanced line-level audio. Here, only the best shielded balanced cable could be used (Belden Belfoil 8761 albeit now superseded), and the cables terminated in patch panels using gold-plated LEMO connectors where the cable earth looped through the panel rather than to it (for source or destination earthing)—all to stop hum and crosstalk. Through necessity, both microphone and line-level audio (and sometimes unbalanced 75Ω video) would run in the same ducts (unfortunately a common practice, but often unavoidable).
* Of course, each cable was only some hundreds of meters in length.
Maintaining the noise to within a few dB of the theoretical thermal noise for 150/600Ω on the mike lines in such installations is a tall ask and it's essential that not only the terminating equipment's (recorders, audio desks etc.) circuits are well designed but also the lines are well balanced. Thus the quality of the cable construction is of paramount importance, both with respect to overall shielding and the balance (close tolerances on the capacitance between cable and shield etc.)
It's in environments such as this that CD-A750 cassette/CD combo would require properly balanced inputs and outputs. Moreover, I'm strictly of the view that XLR connectors should never be wired as unbalanced. Unbalancing XLRs not only causes confusion but it can introduce crosstalk and noise into an otherwise clean audio network. (Hum loops are a notorious problem when two of the three pins of an XLR connecter are grounded, it's a common problem with breakout boxes).
One of the annoying problems when equipment with electronic rather than transformer balancing is used in such installations is that longitudinal hum and noise often becomes a problem. RF detection (RF noise from mobile phones, radio stations, clicks from electric motor switching etc.) is much more likely to be induced into such circuits than when a transformer is used. Audio operators also exacerbate the problem of crosstalk, hum and noise by using unbalanced breakout boxes and such (which they love and won't part with on threat of death—as sometimes threatened by annoyed techies).
Furthermore, in these complex environments, high-level audio from power amplifier outputs (speaker feeds) are also often balanced to reduce both the crosstalk induced into low level mike circuits and to overcome the I^2 R losses in the long cable runs. Here, the so-called '100V line' feeds are fed by low-to-hi-Z transformers. In installations where hundreds of power amplifiers feed from central control rooms to distant locations, there's little other choice.