I was just playing with the idea to demonstrate the effect on a square wave of different frequency responses.
So I gave it a try.
Looking at 2 amplifiers:
Benchmark AHB2
Frequency response, from Amir's review
View attachment 306150
If I see correctly, that's behaving like a 1st order low pass filter at around 200kHz.
Let's say it's a Butterworth 1st order filter.
What would a 10kHz square wave look like ?
View attachment 306151
Cool.
Now, let's try with the NAD M23, or C298, or whatever amp using the Purifi module.
NAD M23
Frequency response, as per Amir's measurements above
View attachment 306152
That's a steeper response.
Purifi's datasheet says -3dB at 60kHz and -6dB at 75kHz.
So let's try with a Butterworth 2nd order filter at 60kHz
Square wave at 10kHz would look like
View attachment 306153
That's different, indeed.
Does it matter ?
Now let's introduce another parameter:
Human hearing
What happens after our ear's own low pass filter is applied ?
That's a very steep low-pass filter, since some may hear at, say, 16kHz, but at 32kHz, for sure, we don't hear anything.
Meaning 0. Nada. -100dB or so at least.
Let's try and apply such a filter on top of the amp filter.
Here is what we'd get as a result.
Benchmark AHB2
View attachment 306154
NAD M23
View attachment 306155
In both cases, we get a pure 10kHz sine wave.
Of course, since 3rd harmonic and above are completely inaudible to any human.
(And the theory -and practice- tells us that a square wave may be reproduced by an infinite sum of sine waves at odd multiples of the fundamental frequency, with level decreasing as 1/order)
So are those differences in 10kHz square wave any relevant ?
Not at all.
By the way, don't say a Purifi amp can't do any square wave correctly.
Here is one (predicted) for NAD M23 :
View attachment 306156
Well, granted, that one is at 1kHz.
So at 60kHz, we already have 30 harmonics (including the fundamental), which allows a pretty accurate reproduction of the square wave.
OK, I had fun
Now let me grab a beer...