Crossover issue.

Sounds like something broke.  That's the idea. :)

If it were me I would take out a DATS V2 from Dayton Audio and use it to compare the impedance curves of the drivers and the crossovers to diagnose.

I also recommend asking at DIYaudio where you can actually post pictures of your findings.

btw here is an outline of how I fixed a similar issue recently

https://forum.audiogon.com/discussions/a-loose-connection-can-be-madenning

Crossovers use caps and inductors to control crossover points. If any of those was off it would affect response around the crossover points. Resistors are used to help match relative volume between the different drivers. If a resistor is off it will affect volume output across the whole driver range. Which is what you are hearing. Which points to resistors.  

Find the one(s) that are used on the drivers in question. Remove them. Measure them. Replace with correct matched values. If I'm right it will be that simple.  

Would you happen to have a DMM and a tone generator?

"...and have just been redone by a reputable audio engineer..."

Even good guys make mistakes. Something is wrong with one of the component parts, measure until you find the problem part. .

I would let the reputable audio engineer that worked on the crossovers your problem and let him check out what’s wrong. 

Thanks for all your responses.
I feel like the gentleman who discussed the resistors is probably on track.
That seems to describe it perfectly.
Oldhvymec ...
It is a 4 way so the mid is actually 1500 - 3500
Then tweeter 3500 up.
Thanks again all. 

They call me MISTER shill! lol! Thanks.

Mr. Shill
The resistor in that circuit checked out however a bank of 5 capacitors had only 3 connected... Perhaps broke during shipping.
Since there would be more bandwidth going through there that would probably make it louder too correct?
I soldered the broken wire, but will not have a way to check the SPL until I reassemble.
It does seem like that would do it though, agreed?

+1 stereo5

When you say the bank of capacitors had only three of the five connected, they are in parallel I presume?

That’s an unusual way to get a particular value, two maybe three, but five?
I don’t suppose they are marked with values on them? Voltage and Micro Farads?

It should do much more than simply change loudness of a driver, depending on the circuit, it may have two drivers trying to reproduce some of the same frequencies. Instead of there being a hand off of energies in merging frequencies between drivers, there may be an overlap?

Without a circuit diagram, it’s all a guess.
But I must say if a well made solder joint came apart during shipping, I am very surprised. Does it appear to have been all one piece at some stage?
Got good photos of this?

I used to work building crossovers and speakers cables, USB cables for a living. It was tough to solder solid silver wire to micro USB and even mini USB at times, as the plating on the metals made it harder to get a strong bind, even with good rosin.

With point to point soldering it’s a good method to crimp lead wires over or around each other with minimal contact point, using the solder to mechanically seal it, and also protect from oxidisation.

The @erik_squires post mentioned…

If it were me I would take out a DATS V2 from Dayton Audio and use it to compare the impedance curves of the drivers and the crossovers to diagnose. 

Assuming that they came out good (the same) then that would exonerate the speakers.
Without that I would more the cross over to the other box and visa versa. Then you can noodle out whether the speaker or the XO is bad, but soldering sounds like a clue to pursue.

Well my suggestion was just a start.  DATS is a great tool for diagnosing all sorts of problems.  By looking at the bad impedance curve you can tell if one or another section looks like it is missing, and the LCR measurements can discover bad drivers or components.

Yeah @erik_squires and a less elegant (kindergarten approach) is just to swap left parts between and right…

If one had a scope they could drive both LHS and RHS XOs and plot them against each other.
But you’ve also reminding me that I need to find my DATS.

A microphone and test signal would only tell you if they performed differently. 

DATS will give you a very good idea why.

@erik_squires I did not mean a microphone, I mean to just test the XOs with just a signal and probably some power resistors where the speakers go.

Your idea of the DATS does the speakers, so it is conceptually the same type of thing, just I am aiming it the other way to look solely at the XOs.

Either way we are both splitting the system down the center to figure whether (in your case) the back half is working (speakers), and in my case, whether the XOs are NOT working.

Our approaches are matched (like a Darlington pair).

DATS does it all, end to end, and lets you measure the individual components as well, including the drivers.  Everything else to me seems like trial and error.

Ahhh So DATS the speakers, and also DATS through the XOs?