WTF is a Non Linearity?

Video Statistics and Information

Video
Captions Word Cloud
Reddit Comments
Captions
[Music] hi and welcome back i sometimes wonder what it's like to try to learn audio engineering these days i mean i did it the hard way right i got stuck in and learnt on the job from my mistakes there was no youtube in those days no tutorials to guide me in the right direction sometimes as a house engineer i would work with a touring engineer that really knew their stuff and was happy to share some of that knowledge but mostly i just had to figure it out for myself but then sometimes i wonder if nowadays we have the opposite problem the internet is awash with information and advice some of which is questionable at best and some of which is just plain wrong how do you tell the difference a video recently popped up on my timeline which illustrates this quite well it talks about the difference between linear and non-linear plugins and almost everything it says on the subject is wrong i'm not going to name the creator as it's a smaller channel and it would feel like bullying but i feel like the world clearly needs a what is a non-linearity video so here it is because the term has a very specific meaning when it comes to plugin design let's see what does wikipedia say in mathematics and science a non-linear system is a system in which the change of the output is not proportional to the chains of the inputs okay maybe that's not so helpful on its own so let me explain here's pro-q3 it's a linear eq of course but that's not because the frequency response is perfectly flat when no bands are active nor is it because this bell filter is nicely symmetrical or because it's precisely the frequency i've dialed in none of that is relevant at all and incidentally it's also nothing to do with linear phase mode at the bottom confusing as this may be pro q3 remains linear in the mathematical sense whichever processing mode you choose it's linear because the input level doesn't matter if i drop the input level by 12 db the output level drops by 12 db but remains otherwise identical likewise if i boost the input level by 20 db the output level increases by 20 db on whatever signal level i feed it this eq cut always remains exactly the same analog gear is usually designed to be as linear as possible when fed with signals at its intended operating level but of course no analog gear has infinite headroom if you boost up the input levels efficiently sooner or later the signal will start to distort at this point the equipment becomes profoundly non-linear adding another 12 db of gain at the input won't give you a 12 db increase in output level anymore because the signal peaks will get clipped and distorted incidentally a side note for my fellow guitar players gain just means volume increasing the gain in a preamp means multiplying the voltages to make a bigger signal that's it gain only means distortion when it's pushing the signal into a non-linearity of some kind and while we're at it vibrato means pitch modulation and tremolo means volume modulation just because leo fender got confused 70 years ago i don't see why we all need to still be getting it wrong today sorry i needed to get that off my chest anyway the reason pro q3 is a linear eq is because there's no distortion at all whatever the input signal level you'd have to boost the level by hundreds of decibels before you run into any issues which means practically speaking we can consider it totally clean and distortion free in fact the only way to stop pro q3 from being totally linear is to enable the dynamic features because dynamics processing such as compression and expansion is also inherently non-linear you can think of compression and limiting as an attempt to create a non-linearity that's free of distortion but of course fast attack and release times still result in distortion and added harmonics and that's half the fun because here's the thing distortion sounds good now if you're a rock guitarist that's not going to be controversial at all guitarists have thoroughly embraced distortion to the point that they think that's what the word gain means and it probably wouldn't be hard to persuade them that a tiny drop of the same poison will also work magic on the drums or the vocals other types of musicians are more cautious don't tell a classical cellist that you're distorting their sound whatever you do but if you call it something with fewer negative connotations like saturation or analog magic they'll hear it as extra richness more resonance more detail more clarity and those added harmonics will tickle their ears just like yours or mine harmonics aren't making instruments sound interesting after all we don't just listen to arrangements of pure sine waves as a general rule and it shouldn't surprise us that adding more harmonics can make things sound more interesting in analog land distortion is basically free the expensive stuff often does it subtly and gently when run at normal operating levels which gives the kind of gentle enhancement that we might perceive as more hi-fi rather than less but you always have the option to drive it harder if you want to and when you do the extra added harmonics that result stick to the same musical harmonic series that our instruments create in the first place in digital land distortion is not free it has to be explicitly coded in all its potential dynamic complexity and furthermore it really needs to be oversampled otherwise you risk some of those harmonics you're adding folding back down is aliasing which doesn't stick to the harmonic series anymore so we used to judge analog gear by how clean and free of distortion it was now we judge digital software by how well it can do analog style distortion and again that shouldn't be surprising a little bit of distortion is usually a good thing in the past the struggle was not to end up with too much these days the risk is that unless you explicitly add a little bit you might end up with too little anyway let's talk about measuring distortion using plug-in doctor here's the first graph you're presented with when you launch the software and it shows the frequency response of the plug-in in question or the phase response if you switch it at the top but notice what this first tab is labeled at the top linear analysis this type of analysis is designed for a linear system with no distortion or compression if you try to measure a non-linearity with this method you're just going to break the test now in the case of a plug-in that's almost linear like this lindell channel strip with the dynamics bypassed and the preamp turned down we can safely disregard the added harmonics as they're low enough in level not to mess up the reading significantly and this graph is likely to be a pretty accurate representation of the frequency response through the eq and filter module but if i add more distortion these graph shapes we're getting just become nonsense [Music] we're not getting any useful information about the behavior of the plugin anymore we've just broken the test [Music] but the doctor does provide some useful ways to measure non-linearities let's start with the dynamics tab this is intended to show the transfer curve of dynamics processors such as compressors and limiters but it can also show the way that analog style distortion affects your dynamic range [Music] or the harmonic analysis tab which is more intended for measuring distortion here we have 100 hertz sine wave test signal at a conservative level of -20 dbfs running through pro q3 to begin with notice the output is just a sine wave still with no extra harmonics on top [Music] now here's the lindell channel strip with the input gain all the way down even with its cleanest setting we're getting a significant harmonic series added now look what happens when i boost up the gain [Music] incidentally don't pay any attention to the noise down here at -150 dbfs this is way too low in level to ever be audible at all but that third harmonic though that's not insignificant even with the gain all the way down there's a limitation with this kind of test however in that we're only presenting it with a single frequency if i shift the test signal to a different frequency we might see a very different pattern of harmonics and measuring the harmonics at just a single frequency doesn't tell us much about the way it might sound so let's try switching to the hammerstein tab hammerstein hammerstein i don't know to be honest and continuing to be honest i hadn't heard of this type of test until plugin doctor introduced it to me here's what seems to be going on as far as i can tell the key at the top right tells us which color corresponds to which harmonic and we can see which harmonics are added to which frequency ranges and by how much probably the best way to illustrate this is using spectre from waves factory a very useful plugin that allows you to add non-linearities much as you dial in eq starting flat we see just the first harmonic tooling along at unity no change whatsoever now let's add a non-linear band 200 hertz and now we see a prominent third harmonic centered at 600 hertz and a fifth harmonic at one kilohertz [Music] if i switch the band to the asymmetric tube mode now we've got lots of even harmonics as well that's two four and six times the fundamental frequency [Music] contrast this with a linear 200hz eq boost from pro q3 we get a bump in the fundamental but no extra harmonics at all here's fabfilter saturn running its warm transformer style notice the third and fifth harmonics are fairly significant down in the low frequencies but reduce in level as the frequencies increase this stays very much the same as i increase the drive until all of a sudden the higher harmonics kick in with a vengeance this is presumably the transformer core saturating fully and starting to clip the signal brutally instead of just thickening it subtly contrast that with its warm tube style with the drive set low and now with the drive cranked [Music] asymmetric tube style saturation gives us lots of the sweet sounding even harmonics at two four and six times the fundamental frequency here's warm tape used subtly and now abused disgracefully this one's good for giggle broken tube [Music] disclaimer plug-in doctor is quite slow and sluggish running this analysis at least on my aging processor and dare i say it also a little buggy at the time of writing it often stops responding to changes in the plugin and i have to reload so that's why i've probably pieced this segment together from static screenshots rather than attempt to capture video as usual and it's also why i'm not as confident as usual that all my readings are accurate nevertheless if you want to analyze a non-linearity this is a better method than a test called linear analysis which is kind of like trying to weigh something using a tape measure okay that's all for now thanks for watching [Music] you
Info
Channel: Dan Worrall
Views: 87,036
Rating: undefined out of 5
Keywords: production, mixing, DAW, plug-in
Id: mytc7i0jm34
Channel Id: undefined
Length: 12min 3sec (723 seconds)
Published: Thu Oct 07 2021
Related Videos
Note
Please note that this website is currently a work in progress! Lots of interesting data and statistics to come.