Nord User Forum

Nice work ricard and Mr_-G. Interesting exercise. The amplitude part of the spectrum won't show a difference between sawtooth and inverted sawtooth, but the phase part of the spectrum will. Found an interesting analysis here with some further links:
https://sound.stackexchange.com/questio ... orward-saw
I can't distinguish any difference in the examples Mr_-G posted, nor on the higher freq examples posted at the link above. Thinking in terms of air pressure fluctuation, I guess the waveform polarity shouldn't really have any audible effect. But perhaps electro-mechanical effects of an amp/speaker system could play a role?

(edit: Sorry Mr_-G, just reread your post, and speaker effects were exactly what you were focusing on)

Thanks Mr G, that was a serious demo. Actually, I can hear the difference between the two segments of smooth.wav, but it is not as noticable as when the two segments are back-to-back. I'm listening using headphones, as I recall, when using speakers the effect is less pronounced (probably depends on your speaker setup too).

As for the our hears are not sensitive to phase thing, in fact our ears are very sensitive to phase, when it comes to the relative phase difference between our ears, which is part of our stereophonic hearing. What is usually said is that our ears are not sensitive to absolute phase. Not really relevant for this discussion, as in this case it is the absolute phase which we are looking at.

I'm in deep waters here, but I think the effect, like both of you are implying, has to do with the speaker, my theory is that for one of the sawtooths the transfer function of the speaker cone to the surrounding air drops off at low frequencies, similar to the way a speaker behaves if not mounted in an enclosure - the pressure waves from the front and back of the speaker cancel out at low frequencies, causing a loss of bass response. I think the effect is the most noticeable with an artificial wave such as a sawtooth which is highly asymmetric, most real-world signals do not have that degree of asymmetry.

Thanks all, very interesting (even if I sweared myself not to enter such technical debates anymore )

The speakers factor is really relevant, I had never thought about it. Regarding sawtooth waves (i.e. the upwards-then-falling, I guess the amplitude reaches its maximum whatever quality, transient sensitivity of the membrane. Whereas with ramp waves, when the minimu amplitude is reached, the instant increase can be kind of eroded if the speakers is not reactive enough. This can be very subtle when you think of time contants in stake here.

...

But... actually I'm not sure it relevant... because sound is due to positive amplitudes as well as negative amplitudes. So the phenomenon I've just written happens with both waveforms, either in the positive amplitude domain or in the negative amplitude domain. Sorry, just gathering my physics university memories

Maybe it is not the same to push and pull the speaker membrane in and out of the rest position forwards and backwards. There could be some asymmetry in the elasticity of the material, who knows!
In any case, I would assume that such an effect would be speaker-dependent.

Thanks to all for interesting ideas!

Necroposting to ask the obvious question that for some reason nobody asked: would it be possible to replace the output opamps with phase-inverting ones (or add phase-inverting unity-gain amps after the stock amps) to bring the NL2 in line with the NL1?