Are you not aware that those components are in a hybrid IC?, so it's not easily possible to measure voltages inside it - although breaking the casing off 'may' allow you to do so.
I've repaired countless amps using these kinds of hybrids, in almost all cases the hybrid needs replacing.
1: -145mv ok
2: -145mv ok
3: 0v OK
4: -19v << driven by Pin 7 with 100R
5: -100mv *FAULT* must -1.5V pulled down by R18 towards V- with current limit by R11 + R13
6: 2.7v OK Mute OFF
7: -20v OK
8: -21v OK
9: -21v OK
10: 30mv = OUTPUT
11: 21v OK
12: 20v OK
13: 40mv same *FAULT* must -1.5V pulled by R18 down towards V- with current limit by R12 + R14
14: -21v OK
15: -80mv
16: ov
17: -120mv
18: -125mv
Follow the yellow line from Pin 7 to pin 5 and find the faulty connection.
When connected pin 5 will drop to two VBE drops to -1.5V below the output near 0V.
I know for sure that nothing is wrong with the external parts to the stk, and i know for sure
that the stk4132 is the real problem here.
Why?
Because:
The pcb that i used to make the measurements (sony fh-414 unit) is working good with a stk4122.
but do the same distortion as seen on the first picture if I remove the good one, and put in its place the
the bad one (stk 4132).
That was the reason that made me open the IC to try to trace the signals inside of it.
Well if you can trace the fault that causes pin 5 to have the wrong voltage, you can solve the mystery.
With those DC levels there will be huge asymmetry and crossover distortion for large signal swing. I trust you learnt something about transistor voltages in PNP's by now.