Could Soft Limit function introduce harmonic content?

[Tyrone]

I've been analyzing some sine waves I recorded into pro tools through a DIGI002 and an Apogee AD-16X running on two separate pro tools systems, I sent the sine waves into each converter at the highest possible level with clipping but when I examined the frequency content of the sine waves and was quite surprised to find in that in the lower frequency sine waves (around 100Hz-300Hz) there were harmonics produced in greater number and amplitude than on the 002. I was pretty confused seeing as obviously the AD16-X is a far better converter

could this be a product of the soft limit function on the apogee?

[Wavebourn]

Originally Posted by Tyrone

could this be a product of the soft limit function on the apogee?

Yes.

[Tyrone]

cheers!
How does it this happen? also in comparison to the 002 the sine waves appear visually far more accurate through the apogee even with the presence of these additional harmonics.

[Wavebourn]

Originally Posted by Tyrone

cheers!
How does it this happen? also in comparison to the 002 the sine waves appear visually far more accurate through the apogee even with the presence of these additional harmonics.

Hello!
Multiply any function by any non-linear one and you get harmonics, by definition.

[Tyrone]

Once again thanks!
I've attached a couple JPEG of sine waves at 100Hz it's pretty crazy difference!

[max cooper]

Wouldn't you do the comparison to the 002 with the soft limit "off"?

Or is that with it "off"? (although I don't understand why you'd ask if soft limit could cause the harmonics in "off" so I'll assume it's "on".)

[Tyrone]

I'm doing some analysis of conversion for my final year university project and I wanted to see the effects of the soft limit function which is why it was left it on

[max cooper]

Fine. But you do understand it's apples and oranges, right?

That's like "Why is camera "A" taking sharper photos than camera "B" with radial blur?"

[Tyrone]

yeah of course, you're absolutely right because obviously the soft limit is another variable factor,

I should have done test with the soft limit off to be a direct comparison, if only I actually owned an AD16-X to go an do the test again, DOH! but I'll identify the inaccuracy in my report,

I could probably argue the fact that it's a real world example because on the rare occasion I get to track anything with apogees I always have soft limit turned on

[zmix]

Originally Posted by Wavebourn

Hello!
Multiply any function by any non-linear one and you get harmonics, by definition.

Actually, the "Soft Limit" is simply clipping the waveform, probably using sonething like a Zener diode.

It's an analog circuit.

Looking at the spectrum produced, you can see that the distortion byproducts are all odd harmonics. This indicates symmetrical clipping.

[tINY]

If you look at a voltage vs time (transient) waveform instead of the FFT, it should be really obvious what is happening.....



-tINY

[zmix]

Originally Posted by tINY

If you look at a voltage vs time (transient) waveform instead of the FFT, it should be really obvious what is happening.....
-tINY

It seems pretty obvious what is happening by examining the frequency domain.
Why do you suggest looking at the time domain response?

[Dissdnt_penguin]

Originally Posted by zmix

Actually, the "Soft Limit" is simply clipping the waveform, probably using sonething like a Zener diode.

It's an analog circuit.

I think Wavebourn meant it in the abstract way. Even if it's caused by an analog circuit, distortion could be characterized by a transfer function if you're in frequency domain, and multiplying that by your signal function in frequency domain you can get the distorted signal's function, right?

Originally Posted by zmix

Looking at the spectrum produced, you can see that the distortion byproducts are all odd harmonics. This indicates symmetrical clipping.

I think it's pretty adventurous to guess that out of a logarithmic scale with such a low resolution, but even then I don't understand how odd harmonics indicate symmetrical clipping.

I know it's just a matter of calculating it, but I'm trying to understand it intuitively.
Odd harmonics generate a square wave, right?
So are you assuming that a symmetrically clipped sine wave resembles a square wave?
I guess it makes sense for a very nastily clipped sine wave, but am I correct?

Thanks,
DP

[Wavebourn]

Symmetrical clipping may be indicated by suppression of even harmonics starting from 2'nd when odd harmonics present. In case of perfect meander all even harmonics starting from 2'nd are absent.