Class A SE vs PP

[hellokitty123]

Quick question, I haven't been able to find a straight answer.
A complementary push pull output stage operating in pure class A reduces even harmonics at the expense of increased odd harmonics right?
Why does this happen?

 

[PRR]

SE curves one way.

P-P curves both ways.

 

[rayma]

A complementary push pull output stage operating in pure class A reduces even harmonics
at the expense of increased odd harmonics right? Why does this happen?

The PP circuit is symmetric.

 

[manicjovica]

Even harmonics reduced, odd stays unchanged. Maybe ear senses them more without even to mask...

 

[DF96]

Each half of the circuit generates odd and even harmonics. The odd harmonics are in phase and so add up (as the signal does); the even harmonics are opposite phase and so cancel. You end up with the same proportion of odd harmonics as SE, and much reduced even harmonics.

 

[john_ellis]

The answers in the previous posts are right as far as they go but ... to make the PP cancel the output transistors need to match over the current range. NPN and PNP's never quite match but modern devices are better than the old ones. So you might get better performance in Class A using quasi-complementary outputs, which might seem somewhat surprising.
Gain variation with current is typical of the older devices, but John Linsley Hood's simple 10W Class A is a good example of a quasi-complementary PP amplifier that suppresses even harmonic distortion.