Ok. Just to be 100% clear: "Harmonic content" as I understand it means, which frequencies are present in a sound, and at what relative amplitudes. Nothing more or less. This is most effectively visualized by a spectrum analyzer. "Timbre" is ultimately a bit more subjective. Not a technical measurement. But the harmonic content as defined above is definitely the thing that shapes timbre most directly.
Now, the fact that the waveform I showed above isn't exactly the same one as was used in that video is irrelevant. A soft clipped sine is just a periodic wave with some harmonics in it, and so is the waveform I used for those screenshots. The phenomenon observed is the same. And indeed, his results are the same as mine! Turning on the HPF offsets the relative phases of the harmonics, which changes the *shape* of the wave, but *does not change* its harmonic contents.
If he had shown a spectrum analyzer in that video as well, I'm fully confident it would've wiggled a little at the moment he turned the filter on, but afterwards showed the exact same frequencies at the same relative amplitudes, same as in my test. Our ears are spectrum analyzers too, and listening to that video I hear no difference when the filter comes on, nor do I expect to.
Now, the harmonic contents do not change, I assure you. But does the timbre change? Well, most people would say no, as the conventional understanding is that we can't hear phase shifts. In practice that doesn't always hold water though. It is true for the pure sine wave case. "I can hear that it's happening half a cycle later." isn't an experience anyone has... But for more complex waves we do sometimes hear a difference. The most plausible explanation being that our listening systems (and our ears) are not fully linear, so changes in total amplitude can matter. That waveform peaking a bit higher may cause your headphones/speakers/ears/room to react a bit differently to it.
When I did the screenshots above, I did notice such a phenomenon actually! The filtered wave sounded subtly different than the unfiltered one, even though the harmonic contents are identical. Only at fairly loud volume on headphones, which agrees with the above theory. Another thing you can test is to listen and compare a sawtooth wave with and without the polarity switch off. Most headphones can sound very noticeably different when the jump in the waveform is outwards compared to inwards.
So that's the grain of truth in what you're saying: filters introduce phase shift, and unless you're dealing with just a simple sine wave, phase shifts *can* impact perceived timbre. And more pragmatically, if your filter happens to increase amplitude, you've lost a bit of headroom, which you might've needed if you're worried about loudness.
But here's the thing: These highpass filters are going on complex signals, not super steady period waves like these. On realistic signals, the phase interactions are complex enough that you're not often going to see large amplitude changes. And if you do, amplitude is just as likely to decrease as to increase. And the aforementioned timbral changes from phase shifts range from subtle to inaudible, so you might not get any of that either. And again, if you do it's about as likely to sound better as worse!
I'm sure you are right though that lots of folks mistakenly think these HPFs have no side effects. And for sure, I don't advice throwing HPFs on tracks for no reason! If there's nothing down there, no need to filter.
But I think I can fairly confidently say I understand the repercussions. And I would still advice folks to filter bass out of every track that doesn't need it*. Any cumulative side effects of the filters are will be far less problematic than the muddy low end.
*And not just bass! Same principle applies in every part of the spectrum really: removing unneeded frequencies in one track makes room for other elements in that range.
Now, the fact that the waveform I showed above isn't exactly the same one as was used in that video is irrelevant. A soft clipped sine is just a periodic wave with some harmonics in it, and so is the waveform I used for those screenshots. The phenomenon observed is the same. And indeed, his results are the same as mine! Turning on the HPF offsets the relative phases of the harmonics, which changes the *shape* of the wave, but *does not change* its harmonic contents.
If he had shown a spectrum analyzer in that video as well, I'm fully confident it would've wiggled a little at the moment he turned the filter on, but afterwards showed the exact same frequencies at the same relative amplitudes, same as in my test. Our ears are spectrum analyzers too, and listening to that video I hear no difference when the filter comes on, nor do I expect to.
Now, the harmonic contents do not change, I assure you. But does the timbre change? Well, most people would say no, as the conventional understanding is that we can't hear phase shifts. In practice that doesn't always hold water though. It is true for the pure sine wave case. "I can hear that it's happening half a cycle later." isn't an experience anyone has... But for more complex waves we do sometimes hear a difference. The most plausible explanation being that our listening systems (and our ears) are not fully linear, so changes in total amplitude can matter. That waveform peaking a bit higher may cause your headphones/speakers/ears/room to react a bit differently to it.
When I did the screenshots above, I did notice such a phenomenon actually! The filtered wave sounded subtly different than the unfiltered one, even though the harmonic contents are identical. Only at fairly loud volume on headphones, which agrees with the above theory. Another thing you can test is to listen and compare a sawtooth wave with and without the polarity switch off. Most headphones can sound very noticeably different when the jump in the waveform is outwards compared to inwards.
So that's the grain of truth in what you're saying: filters introduce phase shift, and unless you're dealing with just a simple sine wave, phase shifts *can* impact perceived timbre. And more pragmatically, if your filter happens to increase amplitude, you've lost a bit of headroom, which you might've needed if you're worried about loudness.
But here's the thing: These highpass filters are going on complex signals, not super steady period waves like these. On realistic signals, the phase interactions are complex enough that you're not often going to see large amplitude changes. And if you do, amplitude is just as likely to decrease as to increase. And the aforementioned timbral changes from phase shifts range from subtle to inaudible, so you might not get any of that either. And again, if you do it's about as likely to sound better as worse!
I'm sure you are right though that lots of folks mistakenly think these HPFs have no side effects. And for sure, I don't advice throwing HPFs on tracks for no reason! If there's nothing down there, no need to filter.
But I think I can fairly confidently say I understand the repercussions. And I would still advice folks to filter bass out of every track that doesn't need it*. Any cumulative side effects of the filters are will be far less problematic than the muddy low end.
*And not just bass! Same principle applies in every part of the spectrum really: removing unneeded frequencies in one track makes room for other elements in that range.
Statistics: Posted by Andreya_Autumn — Sun Jan 26, 2025 1:06 am