Impact of phase inversion by preamp

many (most?) loudspeakers have their two or three drivers wired in opposing polarities

hopefully to correct crossovers phase shift and not arbitrarily.

It is more than the bass that is affected. Percussion has very steep transients, sometimes exceeding 20KHz on tape and LP. On quality phase coherent systems, the effect is clearly audible by some. Not all are sensitive to it.

Expanding on what Al said regarding polarity flip circuits, even in fully balance circuitry, they are seldom in themselves inaudible. This can be demonstrated by reversing loudspeaker polarity and listening to the A:B, which is quite likely C:D.

In a balanced preamp the phase switch simply redirects the inverted phase to the non-inverted circuitry and vice versa. Its completely passive.

Nothing is completely passive. The more revealing the circuitry, the more it exposes component 'flaws'.

Experience with phase flip switches in active and passive balanced hardware often left something to be desired in terms of routing and hence inaudibility.

Making both halves of a balanced circuit identical is EXPENSIVE. Add in tube vagaries and the probability of a sonic delta is fairly certain. Not to all listeners, but to some.

Decades back it was demonstrated there were distinct sounds in a system on the same source:

1.  Normal polarity to speakers
2.  Invert channel A at source* and speakers - same Φ as 1
3.  Invert channel B at source* and speakers - same Φ as 1
4.  Invert both A&B at source* and speakers - same Φ as 1

*Balanced out - polarity swapped with identical but inverted polarity cables.

Theoretically all should sound the same, but they didn't. If one subscribed to OOPS, at least 2 & 3 should sound the same, but they didn't. It was postulated that the loudspeaker is an asymmetric load and depending on the other hardware, affects the sound.

@cleeds 
Nothing is completely passive. We may think of C in terms of uF, but there are L&R factors which pertain. Pull apart a switch and examine the contacts. One set may have a contaminants/wear/geometry that the other does not.

A great recording & electronic design engineer built an incredible sounding recording console that used mercury wetted relays as close to the active circuitry as possible to eliminate passive colorations from running traces [audible] to surface switches [audible] and back.

Mic pre's, EQ, recording consoles in the studio.

In the early 80's. Then state of the art LA recording studio. Participants were well regarded engineers and IIRC, some industry sales reps. Source was hot-rodded ATR-102 ½" master tape feeding Bryston amp stack and time aligned studio monitors. Test was change or no change. Blind. Most got most of the changes.

These statements are by no means universal truths :)

EXACTLY!

My point is to get people to consider what's behind the faceplate. Far too much HiFi is buzzword-bingo.

Quite simply, operating the phase inversion switch is subtle- on multi-channel recordings you can't hear any effect at all.

Then I would say that the speakers may not be well aligned. The point of the mentioned test was to determine the audibility of polarity. We used polarity correct recordings of acoustic instruments, 80's pop with all live players recorded with polarity correct mics, analogue reverb [plate and chamber] and 80's pop with drum machines, synths, digital reverb and effects. The polarity correct channel inversion was a side show to demonstrate that identical seldom exists.

100db CMMR is excellent. I've endured 'differential' circuits that barely made 60 at some frequency and drifted all over the map with temperature. No offense, but I note that neither CMRR nor the bandwidth over which 100db is attained is listed on the MP3 web page. The only mention of CMRR on the Atma-Sphere site is an explanation here http://www.atma-sphere.com/Resources/balanced.php

In short, with any multi-channel recording all bets are off when it comes to polarity. Its a mixed bag, plain and simple.

Nonsense.
I polarity tested and corrected all the microphones.
I verified all the lines and electronics.
I made the recordings.
Some recordings were M-S, X-Y, spaced pair and some multi-mic.
Synthesizers are easily polarity checked with a sawtooth wave.
The tracks were 100% polarity correct.

Added effects do not alter the polarity of a recording any more than an acoustic hall alters the polarity of the instruments on stage. The phase of  an effect may differ from the signal, but the polarity remains unless purposefully inverted.

I first became cognizant of polarity in the mid 60's with the Command Series of recordings on 35mm tape. I forget how or why.

For a long time, LPs were marked as to preferred polarity. They encompassed the full gamut of recorded music from Pop to Opera. Obviously some were marked '??'

IOW with effectively all multi-track recordings you can't hear absolute phase.

I'd agree on many and maybe almost all made from the computer era forward, but there are plenty of multitrack recordings from 1955 onward when the Ampex Sel-Sync 8-track was invented that while their polarity is inverted overall, it is consistent from track to track. 

Now if you pay attention to those effect devices and compensate for their polarity (excepting phase shifters!), then you can make it happen.

When not working in a studio I knew to be polarity correct, SOP when using EQ, Limiter, etc. is to patch the track to a 2nd input, insert the device in the track and bring up the 2nd input. If the level increased, good to go. If not invert. Many mixing desks of the era included a Phase Correlation meter. It's a simple matter to use a mic of known polarity to check all the other mics in use. Of course, misteaks did happen ~<;-)

But you are one in a million when it comes to this sort of thing!

Many recording engineers I knew were every bit as concerned and delivered masters with correct polarity. Master Refs where checked to ensure the cutting house was polarity correct.

I meant that I marked polarity on the jacket, not that the LP itself was marked. The only ones I recall mentioning polarity were some D2D.

I just recalled when The Tubes "The Completion Backward Principle" was mastered, the producer, David Foster, and Fee Waybill came over with a stack of refs from all over town and New York. I'd been busting David for a few years about his 'neglect' of the mastering process. I'd heard masters and sometimes the LPs didn't 'cut it.' For whatever reason, perhaps to show I was full of [?] or the band pressured him or ????, he had multiple refs made on different lacquers from multiple labs.

At the time I think I had KEF 105.2, Yamaha separates, Oracle TT [arm, cartridge forgotten]. I'd modified both the 105's and electronics.

I forget most of the details now, except some discs were inverted polarity relative to others, verified with an oscilloscope. The best sound was a French lacquer. In the end, it was mastered at Sterling Sound. As I recall, the LP rocked!

some discs were inverted polarity relative to others

I only checked polarity between the reference pressings. I wanted to hear each with the same relative polarity.

As far as the mechanics, just plug a sampling scope into a tape output and check first impact. This also let me check max level and to a lesser degree compression if the wave shape was altered.

As mentioned, multitrack recording may have inconsistent polarity between tracks. I didn't record or mix the album, so I had no idea as to absolute polarity.

I have no idea what %.

That some recordings sound better in one polarity flipped at the loudspeakers, no doubt.

Whether an upstream switch rather than speaker invert is better, I can't say. More convenient, definitely. Remote controlled, even better.

Today, I don't bother.

What evidence do you have that Polarity is a 50% proposition for CDs?

@geoffkait What proof is there that it is anything but random?