Where EQing absolutely makes sense is before clipping. If you clip a bass-heavy drum like a kick drum the clipped low frequency start to have a "farty-cloudy" sound, losing punch and clarity. The trick is to have an EQ before clipping turning down a bass and another to turn back up the bass afterwards (Dan Worrall has the emphasis-deemphasis video on the matter).
One interesting point here is that the original waveform is recoverable only in the transfer function of the waveshaper has a positive or negative slope throughout the entire function, with possibly one inflection point with a zero derivative. You need the Y's to uniquely map back to the X's, and that can only happen if there is only one X value to map back to. This rules out the hard clipping curves with flat sections at the top or bottom, and also many of the "crossover distortion" curves. I don't think your premise is entirely accurate. Imagine you start with a 1kHz sine tone, push it through a tanh() waveshaper, and get a series of odd order harmonics... 3kHz, 5KHz, 7KHz, etc. You can then make an EQ move in the spectra above your original 1kHz tone and affect the now existent harmonics. The extreme example may be to use a higher order highpass filter around 2KHz... for the thought experiment, we can even simplify this to be an ideal lowpass filter that completely eliminates the original 1kHz tone. We are still left with the harmonics created from our waveshaper. One more note. The real complexity of the magnetic saturation seems to come from the hysteresis. The output of a hysteresis model is state-dependent. I doubt it is inversible, although I'd have to take a closer look at the equations to say that definitively.
I never thought of or used the term saturation for harmonic enhancement so going back to your previous video to hear that was new to me. My understanding of harmonic enhancement was to take a frequency and make a duplicate of it at double and triple, or 1/2 etc. of the original, I thought. That's what the term means in RF harmonics--2nd, 3rd harmonic, etc. Good learning nonetheless.
Saturation is one way of producing harmonics, which as you say are whole-number multiples of the original frequency. Of course it gets more complicated than that with intermodulation distortion, and analogue tape which adds non-harmonic frequencies due to scrape flutter. DM
Hello DM, Yes, saturation is not the same as boosting frequencies. A spectrum analyzer will bear this out in no time flat. I've developed several saturation plugins, the spectrum analyzer is indispensable. And as you know, tubes, transistors and FETs saturate as well, with different harmonics being produced, for a slightly different sound. Best regards, Bill P.
Your explanation is very complete as to Plug-ins in a DAW. Saturation as I know it is the limit of analog tape to hold more magnetism. The result is generally a loss of high frequencies and a soft, somewhat pleasing emphasis of the lows and compression of the dynamic range. This is of course, distortion, but not all distortion sounds bad.
I remember when digital was becoming popular in the 1980s as a recording medium. A lot of people tried to convince me that drums and cymbals sounded better on analog and that digital always made cymbals sound harsh, strident or brash. I knew it was because the analog tape was being saturated and making the hf content "warmer". They were actually objecting to the accuracy of the digital recording.
Actually, the transfer function and waveform clipping are not separate things. The transfer function maps an input intensity value to an output intensity value. If the transfer function is bounded, then there is clipping. In any case, a nonlinear function will alter the waveform.
Fascinating…though I will continue to use the term, more often than not, incorrectly. Haha It’s just such a pleasing word to me. Excellent video, as always! Thank you!
Some “saturators” change the overall frequency response and some do not. For example, plugins modeling specific hardware can have a low frequency boost, high end roll off: lots of things over just clipping and/or adding harmonic content. I imagine for those that don’t change the frequency response EQ placement is less important. Those that do, it’s probably more important.
So one can assume that “saturation“ is a marketing thing most of the times? Should they be named “Harmonic enhancers” instead of “saturators”? Can real saturation even be “modeled” for the emulation plug ins?(strips etc)
I mean no disrespect and I love watching your UA-cam channel… but my understanding in the audio industry and especially guitar, electric guitar, kind of music, amplification world and effects pedals world is that distortion usually implies TWO things happening, both a compression and a distortion. Now the word saturation is supposed to also imply a lesser degree of “distortion”, that is less than the word “distortion” by itself implies. So anyway, I’m not sure if you’re the saying the same thing, but this is my understanding as a previous electric guitar player of years ago (and a current practicing electrical engineer EE now).
Taking into account your subsequent comment, when an electric guitar distorts it also compresses the sound. In fact, since saturation affects peaks, any kind will cause distortion. I'd have to ransack my electronic theory books but I imagine that perhaps distortion caused by insufficient standing current in a Class A/B amp would only affect the signal close to zero so there wouldn't be compression. I'm sure you know more than me on that. But yes, a distorted electric guitar has less dynamic range and is therefore compressed. DM
It's a tool to make the mix more pleasing to an ear, more dense, richer and warmer sounding. When used correctly and not too much of course. Too clean audio sounds boring and not dense enough (you hear inidividual harmonics which is very distracting to an ear and it's very boring sounding at the same time). Too much saturation sounds like the audio is dirty and distorted and it should be avoided also. Somewhere in the middle sounds the best
![If you can't hear this then you're not an audiophile [See description for link to followup video]](/img/n.gif)
Where EQing absolutely makes sense is before clipping. If you clip a bass-heavy drum like a kick drum the clipped low frequency start to have a "farty-cloudy" sound, losing punch and clarity. The trick is to have an EQ before clipping turning down a bass and another to turn back up the bass afterwards (Dan Worrall has the emphasis-deemphasis video on the matter).
One interesting point here is that the original waveform is recoverable only in the transfer function of the waveshaper has a positive or negative slope throughout the entire function, with possibly one inflection point with a zero derivative. You need the Y's to uniquely map back to the X's, and that can only happen if there is only one X value to map back to. This rules out the hard clipping curves with flat sections at the top or bottom, and also many of the "crossover distortion" curves.
I don't think your premise is entirely accurate. Imagine you start with a 1kHz sine tone, push it through a tanh() waveshaper, and get a series of odd order harmonics... 3kHz, 5KHz, 7KHz, etc. You can then make an EQ move in the spectra above your original 1kHz tone and affect the now existent harmonics. The extreme example may be to use a higher order highpass filter around 2KHz... for the thought experiment, we can even simplify this to be an ideal lowpass filter that completely eliminates the original 1kHz tone. We are still left with the harmonics created from our waveshaper.
One more note. The real complexity of the magnetic saturation seems to come from the hysteresis. The output of a hysteresis model is state-dependent. I doubt it is inversible, although I'd have to take a closer look at the equations to say that definitively.
the colours on the video are so.. vibrant.. pleasant?? theyre really good! keep it up!
So true. Saturation is not only about altering the frequencies, but also the dynamics(transients).
I never thought of or used the term saturation for harmonic enhancement so going back to your previous video to hear that was new to me. My understanding of harmonic enhancement was to take a frequency and make a duplicate of it at double and triple, or 1/2 etc. of the original, I thought. That's what the term means in RF harmonics--2nd, 3rd harmonic, etc. Good learning nonetheless.
Saturation is one way of producing harmonics, which as you say are whole-number multiples of the original frequency. Of course it gets more complicated than that with intermodulation distortion, and analogue tape which adds non-harmonic frequencies due to scrape flutter. DM
Hello DM,
Yes, saturation is not the same as boosting frequencies.
A spectrum analyzer will bear this out in no time flat.
I've developed several saturation plugins, the spectrum analyzer is indispensable.
And as you know, tubes, transistors and FETs saturate as well, with different harmonics being produced, for a slightly different sound.
Best regards,
Bill P.
Your explanation is very complete as to Plug-ins in a DAW.
Saturation as I know it is the limit of analog tape to hold more magnetism. The result is generally a loss of high frequencies and a soft, somewhat pleasing emphasis of the lows and compression of the dynamic range. This is of course, distortion, but not all distortion sounds bad.
Not all distortion sounds bad, I agree. What would Jimi Hendrix be without the distortion of his Marshall amps? DM
I remember when digital was becoming popular in the 1980s as a recording medium. A lot of people tried to convince me that drums and cymbals sounded better on analog and that digital always made cymbals sound harsh, strident or brash. I knew it was because the analog tape was being saturated and making the hf content "warmer". They were actually objecting to the accuracy of the digital recording.
This is true. We've all gotten used to the sound of analogue drums. We like it and prefer it to accuracy. DM
very excellent ---- so well done
Are transformers a part of "tube sound"? In addition to saturation, there's residual core magnetization. There will be hysteresis.
Actually, the transfer function and waveform clipping are not separate things. The transfer function maps an input intensity value to an output intensity value. If the transfer function is bounded, then there is clipping. In any case, a nonlinear function will alter the waveform.
Fascinating…though I will continue to use the term, more often than not, incorrectly. Haha
It’s just such a pleasing word to me. Excellent video, as always! Thank you!
Thanks for the video, also you sound like Paul McCartney
The more interesting question to me would be why do we like it and find the sound pleasing it all
Thank you Mr
I think I understood hahaha for a minute there I thought the vid was made from the bath tub once you held up the sponge hahaha
The mixing bowl from my kitchen did a great job.
watched again ----- wonderful again
Some “saturators” change the overall frequency response and some do not. For example, plugins modeling specific hardware can have a low frequency boost, high end roll off: lots of things over just clipping and/or adding harmonic content. I imagine for those that don’t change the frequency response EQ placement is less important. Those that do, it’s probably more important.
if it creates "harmonic enhancements" does that mean it creates a chord? Why not just call it a chord?
Very cool way of explaining it
So one can assume that “saturation“ is a marketing thing most of the times? Should they be named “Harmonic enhancers” instead of “saturators”? Can real saturation even be “modeled” for the emulation plug ins?(strips etc)
great video! well said
Love the science. Very interesting.
I found gem finally ❣️
awesome! every video!
Shoot, I meant said it wrong. I meant to say that saturation implies both things.
fuking amazing. im on board saturation is too loose
i'm back..
I mean no disrespect and I love watching your UA-cam channel… but my understanding in the audio industry and especially guitar, electric guitar, kind of music, amplification world and effects pedals world is that distortion usually implies TWO things happening, both a compression and a distortion. Now the word saturation is supposed to also imply a lesser degree of “distortion”, that is less than the word “distortion” by itself implies. So anyway, I’m not sure if you’re the saying the same thing, but this is my understanding as a previous electric guitar player of years ago (and a current practicing electrical engineer EE now).
Taking into account your subsequent comment, when an electric guitar distorts it also compresses the sound. In fact, since saturation affects peaks, any kind will cause distortion. I'd have to ransack my electronic theory books but I imagine that perhaps distortion caused by insufficient standing current in a Class A/B amp would only affect the signal close to zero so there wouldn't be compression. I'm sure you know more than me on that. But yes, a distorted electric guitar has less dynamic range and is therefore compressed. DM
What is saturation? A fancy word for sound clearness destruction.
Yes
It's a tool to make the mix more pleasing to an ear, more dense, richer and warmer sounding. When used correctly and not too much of course. Too clean audio sounds boring and not dense enough (you hear inidividual harmonics which is very distracting to an ear and it's very boring sounding at the same time). Too much saturation sounds like the audio is dirty and distorted and it should be avoided also. Somewhere in the middle sounds the best