Quantize or No Quantize (Groove Theory pt. 2)

Welcome yall to the second installment of the groove theory series where we explore ways to get more groove and bop from your tracks. If you haven’t checked it out already, you can read the first installment of the series here. Full disclaimer, this article may not be as useful to the advanced producers out there, but if you are a beginner, or maybe even intermediate, this article is worth a read. In the world of modern music making, you’ve most likely crossed paths with a process a process called quantization. You’ve also probably pondered the age old question, “to quantize or not to quantize?”. There is a lot of stigma surrounding this process. Some people think you should avoid quantization at all costs, while others utilize it to help add a certain character to to their groove. Others view it as a necessary step to help express themselves when their current musicianship doesn’t allow for it. No matter your preference or situation, quantization is a tool that when utilized with intention and control, can render powerful results. With my productions I prefer a hybrid approach and use quantization to not only achieve movement I couldn’t obtain otherwise, but also to speed up my creative process to strike while my inspiration is hot. I’ll be going over a few little tips and tricks that I like to use and hopefully offer some insight for your future beat makings. So lets jump in to the second topic of our groove theory series, quantization!

Quantization and a breif history

So lets define quantization for the uninitiated. I’m going to rip this definition right from good ol’ wikipedia because I think it does a pretty good job of summing up the process (and I may or may not have written this article and forgot to save it and am crying while I rewrite this whole damn thing):

quantization is the process of transforming performed musical notes, which may have some imprecision due to expressive performance, to an underlying musical representation that eliminates this imprecision.

Pretty solid, however I would probably amend this with “… which may have some imprecision due to expressive performance or inability to play…”. So quantization is a basic process. It’s more or less an algorithm that detects mistimed notes, dependent on a few variables, and corrects those timing errors. Lets look at a basic example. A majority of modern day music is made in a 4/4 time signature. This means that every four (that’s the first number) 1/4 notes (that’s the second number) you get one measure or bar. AKA every beat of a bar in 4/4 is a 1/4 note (that’s a lot of fours). Now lets say we are an inexperienced producer/musician and we are making a song that consists of one hi hat every beat (probably not the most interesting song but whatever) of a bar. Not too hard of a task, but your musicianship has you hitting notes early and late, and while your friends wont tell you because they love you, it sounds a little sloppy and not in a good way.

So whats the move? Quantize. So you hit that quantize button and BOOM, presto! Your notes are perfectly in time an each of the bar you hear your note. Flawless victory. But wait… now things feel less human, a lot more rigid and a whole lotta sterile. This is due to the nature of quantization and one of the polarizing factors of the process. This effect becomes even more amplified when considering someone who has great musicianship. Lets say you had the fortune of having Quest Love stop by to re record those hi hat parts (why you would waste that opportunity on 4 hi hat hits I don’t know but lets rock with it for the example). You record him and things are just sounding right. Everything is in time and looking lovely, and you didn’t even need to quantize. Looking at the notes on your computer screen you can see that they are pretty much spot on. So why does Quest’s drums sound so much more groovy than yours? Well if you zoom in closely you will see that there are indeed slight deviations in timing between Quest’s performance and the grid lines on the beat. These small deviations are what the wikipedias definition was describing as “expressive performance”. This expressive performance is what makes those four hi hat notes sound more human. The interesting thing is that If we were to quantize Quest’s hits they would sound just as sterile as our hits that we quantized. So why even use quantization if it’s only going to rob our music of expression. Part of the answer to that question lies in the history of quantization Lets take a quick little trip down memory lane:

While quantization is certainly a tool of modern music , the process existed even in the days of tape recording. Back then this was a far more destructive process where engineers had to use razorblades to cut fractions of a second of tape out of a reel and splice the remaining pieces back together. This would remove the air time between the missed notes, making for a tighter sounding performance. Traditionally bands would record until they hit everything perfectly but occasionally the engineer had to dive in and really cut into things. Eventually, technology worked it’s way into the industry in the form of drum machines and step sequencers. Units like the 808 allowed the user to input triggers on a step sequencer which would play a specified sound at the specific step you dictated. This was FULL quantization. If you set a sound to trigger at the fourth beat of a bar, it will hit on time EVERYTIME the machine hits that part of it’s pattern. There was no shuffle, swing or groove setting on the 808, so we are talking very rigid in-time hits. Eventually units like the 909 came along which offered shuffle variation to your groove. This additional parameter allowed you to add a little bit of bounce and human swing to your productions. So long overly rigid beats and hello swang.

The SP1200 allowed users to record in their own samples/. It also came with a swing setting.

The SP1200 allowed users to record in their own samples/. It also came with a swing setting.

A few years later came the SP1200, which had the same type of step sequencer, however this machine allowed for the user to add in their own samples. No longer were you forced to the on board sounds like the 808. You could now load in a break from a record or your favorite sample and have the machine trigger that. 1200’s also came equipped with a special kind of shuffle and variable quantization control. The user now had more options for their sounds. As the years passed, more hardware because available, adding to flexibility and groove capabilities. Eventually these iconic drum and samplers became coveted because of their signature sound, which was in partt due to the shuffle and quantization feel programmed into the machines algorithms. So much so that you can find a lot of current DAWs offering “1200”and “MPC”shuffle and quantize settings in as an option. These pieces of hardware became staples in the hip hop sound as we know it, and since hip hop ended up influencing damn near every genre, it’s safe to assume that these became blueprint (obvious or not) to a majority of modern day music.

MPC 3000. A machine made iconic by J Dilla and other legendary producers of the era.

MPC 3000. A machine made iconic by J Dilla and other legendary producers of the era.

While quantization can definitely rob your music of it’s emotion, it does also have the ability to add a groove and movement that is actually desired. Once again, the important thing to remember is that it’s all in how you use it. Be deliberate and you may find the beauty in quantization.

What Would Dilla Do:

Dilla

I think this is the biggest argument for why people are so anti-quantization. We all (whether you conscience of it or not) know and love Dilla. His emotion would always sing through all of his productions, and his bop and groove became iconic. Part of this was because he almost NEVER quantized. He believed in playing your machine like an instrument, and keeping as much of the human element as possible. So it’s settled. No quantization because we want to sound like Dilla and that’s that. see ya later… But hold up hold up. It’s important to also acknowledge that Dilla was also extremely talented (understatement of the century), not only as a producer, but also as a drummer. He had the ability to play things in and out of time as he pleased, so not quantizing made a lot of sense. We aren’t all so musically inclined however. In addition to this, not all of us have a keyboard/drumpad/drumachine/etc, so maybe drawing things in with a mouse or using your computer keyboard is the only option. While playing things out until it sounds right is always an option, it may not be the most forgiving to your inspiration and time, and when you want to create and not spend an hour laying a basic drum part down, it can be just as effective to use some quantization to help speed things up. With all of that said I want to reiterate that there is no right or wrong way, only the way that makes sense with your work flow. So if you got the Dilla skill then maybe you can skip this article and quantization altogether, but if you don’t , or are maybe just curious, lets dive into some tips and tricks to help you harness the power of quantization.

Offset:

Offset

This is probably my favorite trick and the best part is, it’s so damn simple. Not only with this give you some wild swing, but you can accomplish this with clicking in your notes with a mouse, or just by playing it out. The idea is that you fully quantize notes, and then slide the notes to the right, or left. This will give the impression of a sloppy swingy groove. Easy enough right? So lets use it in an example.

Here is a basic drum pattern with hi hats coming in every 1/4 note (or beat since we are in 4/4). I played these hi hats on my keyboard, and if you zoom in you can see timing discrepancies. The first hat is late and others are early/also late. You can listen to this example below.

Here is a basic drum pattern with hi hats coming in every 1/4 note (or beat since we are in 4/4). I played these hi hats on my keyboard, and if you zoom in you can see timing discrepancies. The first hat is late and others are early/also late. You can listen to this example below.

Here we have a basic drum beat with a kick and a snare and some newly added hi hats. In this example I just played them in. They are a little sloppy, so I’m going to fully quantize all of them.If we play this back we are now left with a pretty rigid beat. Doesn’t sound terrible, but definitely lacking some bop. So how do we inject feeling back into this part? By offsetting these currently quantized tracks. But here is the important part, make sure you turn off snap to grid. Now just ever so slightly (or aggressively if you want to go that route) nudge it to the right.

First we fully quantized our hi hats. This but the hits directly on beat, which had the side effect of robbing our performance from it’s natural groove. So to remedy this, we nudged the now fully quantized hats over to the right a few clicks. We accomplished this by disabling the snap to grid function which allowed us to slide over the notes freely. The result is some nice swing.

First we fully quantized our hi hats. This but the hits directly on beat, which had the side effect of robbing our performance from it’s natural groove. So to remedy this, we nudged the now fully quantized hats over to the right a few clicks. We accomplished this by disabling the snap to grid function which allowed us to slide over the notes freely. The result is some nice swing.

Sounding much better after that offset. We can make this bop even more by messing with the velocities.

Hear that now? Even more swing. This is all from full quantization and some nudging. While you could probably play this part out, this took all of one and a half minutes to accomplish. And yes, that kick and snare are RIGHT ON GRID. So that swing is pretty much entirely from offsetting those hats. If you want to take this a step further, you can nudge certain notes farther than others, or offset the snare or one of the kicks (making sure that you are snapping to grid of course).


Going Full Quantize:


This next tip is fairly obvious, but is often overlooked by novice producers. Try and avoid the full quantize. Every modern DAW has quantization settings, and they all offer quantization strength control. Unless you are going for the offset method, or want rigid beats, this will 100% rob your groove. Instead, try quantizing your notes at 25% strength. See how things are sounding. If they are still a little too sloppy (and not in a good way) try quantizing at a higher strength. I find 50% to be a sweet spot, but don’t be afraid to use a strength that fits the vibe. Here’s a quick example of this in practice. Here is the part we played in the last example.

Now lets quantize this by 100%.

Super rigid. Lets undo that and now hit it with 25%.

This is what 25% quantization looks like. The performance is still off the grid and more of the expression is preserved

This is what 25% quantization looks like. The performance is still off the grid and more of the expression is preserved

Still a little messy, but not as robotic as the 100%. Lets try it at 50%.

This is what 50% looks like. Still preserving some of the expression of the original performance, but keeping things tighter than the 25%

This is what 50% looks like. Still preserving some of the expression of the original performance, but keeping things tighter than the 25%

Yeah that’s sounding much better. It took some experimenting, but once again, the overall process was most likely faster than playing it over and over again. This is also partyl a taste thing. There are times when 100% may be what you are looking for, and then there may be occasionas where you really want things sloppy so 25% will work. I would always encourage experimenting with lower strength first and working your way up.

Resolution

A standard feature on most quantization settings is resolution size. This usually defaults to 1/16th.

A standard feature on most quantization settings is resolution size. This usually defaults to 1/16th.

Another setting in quantization that is often misunderstood is the resolution of the quantization. Typically, the default quantization resolution for most DAW’s is 1/16ths. So what exactly does that mean and why does it matter? Well that number is what dictates how the program will judge where to put your notes. Remember when we were talking about 4/4 time signatures and Quest coming through to lay his spectacular 1/4 notes hi hat bar? If we quantized his hits with 1/16th resolution our software would bring his notes as close to the nearest grid marker. Since we are in 1/16th resolution, that gives us 16 different points in a bar that it could try and place a note on. In the case of Quest, because he is so on point, if we quantize with these settings, it will nudge his notes to the 1st, second, third, and fourth beats (or 1/4 notes). So this actually works ok. Our quantization would do that same job if we decreased the resolution to 1/8th or even 1/4th. What would happen if we were on 1/16 resolution with a less ideal performance where notes were considerably off.

As I mentioned before, the quantization algorithm will place notes on the closest grid mark and since we have a a resolution of 1/16ths it’s going to break that grid down to 16 points. If we had a performance where our hits were closer to the 2nd 1/16th our computer would now thinks that it should quantize our first note to the 2nd 16th instead of the first. So we are now left with an obviously out of time and late hit. This is all because we chose the wrong resolution. So the remedy here is to choose your resolutions appropriately. If you are hitting a a note on every 1/4 hit then maybe stick with that 1/4 resolution. This works the other way too. If you had a hi hat pattern that hits every 1/8th note and you quantize at 1/4 resolution, your computer isn’t going to have enough reference points to place your notes appropriately. This would result in your computer actually layering up your notes ad eventually deleting others. So in that scenario it probably makes sense to use 1/8th or maybe even 1/16th resolution… catching the pattern? Basically you want to analyze what you are playing and adjust your quantization accordingly. Now in all honesty I rarely think about this when I’m actually producing. I would say 9 times out of ten 1/16th resolution works perfectly fine, but if you are having weird issues like lost notes, or things aren’t quite locking in, you should probably check your resolution. Once again this is a pretty basic concept, but worth going over.

The 3:2 ratio method

Alright so we went over some basic concepts so for the more advanced users reading this, I applaud you for getting this far. Hopefully you have still learned something… BUT I do have one finally tip/trick piece of theory that I will wrap up this article with that I think we can all learn from. I actually just became familiar with this technique, and while it’s something that a lot of producers do without thinking, this is a theoretical way of approaching it All of the previous examples we have listened to throughout this article were in traditional straight 4/4 timing with a standard 1/16 grid. But there is such a thing as a triplet grid. A triplet grid more or less splits up your grid into divisions of three instead of an even two. So grid with a resolution of 1/8 would have 8 available points to add a note to, while 1/8 triplet would have 12. A 1/4 would have four and a 1/4 triplet would have 6 and so on and so forth. If we lay a hi hat out on each open space on the grid and play it back, there is an obvious difference between the standard grid and the triplet grid. The triplet grid almost has a waltz to it. You can get some pretty wild grooves with this, but it gets crazier. If you dive even further you can actually break your grid down to what’s called, quintuplet grid. With a quintuplet grid, there are five notes every beat where there would traditionally only be 4 for standard and 6 for triplet(every 1/4 note on 1/6th resolution in grid size). 

That’s a lot of math, so digest that for a second and lets get into how this is useful. Lets show an example. First off there is not quintuplet grid setting like there is triplet. So we are going to have to make it ourselves. In Ableton I can time stretch midi notes (very useful btw) and since we know quintuplet swing has five notes every 1/4 note (in 1/16th resolution) we can draw in 6 notes and time stretch the midi notes so we have five within that 1/4 note spacing. Why add the sixth note you may ask? Well it gives use a handle to time stretch these notes. Check the gif below to get a visual of what I’m talking about.

The grid is 1/16ths. In order to get into the quintuplet swing we are going to have to time stretch some MIDI. There are other ways of accomplishing this, but time stretching is the easiest. In order to do that we must first draw in six notes. The idea is to take five of those notes and squeeze them into that first 1/4 note spacing (1.2 marker as shown on the grid). We have a sixth note added because that will be our time stretching handle.

The grid is 1/16ths. In order to get into the quintuplet swing we are going to have to time stretch some MIDI. There are other ways of accomplishing this, but time stretching is the easiest. In order to do that we must first draw in six notes. The idea is to take five of those notes and squeeze them into that first 1/4 note spacing (1.2 marker as shown on the grid). We have a sixth note added because that will be our time stretching handle.

Now that we have out six notes drawn in we can time stretch this midi to fit five notes in our 1/4 notes spacing. In ableton you just hightlight the midi and a stretch handle will appear on top. Click on the handle and drag until the five notes are sitting nicely in the spacing.

Now that we have out six notes drawn in we can time stretch this midi to fit five notes in our 1/4 notes spacing. In ableton you just hightlight the midi and a stretch handle will appear on top. Click on the handle and drag until the five notes are sitting nicely in the spacing.

I’m not exactly sure why Ableton needs this extra note, but it does so I’ll appease. So we time stretched the notes and now we have five notes sitting perfectly within our 1/4 note spacing.

Now lets implement this method called the 3:2 ratio rule. We want our first note to run the duration of three spaces in quintuplet grid, and the next one to span two. This is what that looks like.

We started with five notes but if we convert this via the 3:2 rule we are left with two notes. The first notes spaning the duration of three and the second spaning the duration of two.

We started with five notes but if we convert this via the 3:2 rule we are left with two notes. The first notes spaning the duration of three and the second spaning the duration of two.

Things are starting to resemble some swing now. Now bear with me, and lets take this every further. If we want to make this have some real bop, lets add a little variation to this drum kit by adding an additive shaker. Lets have this shaker hit on the one, but instead of having this element follow the 3:2 rule, lets break that and reverse is and have the first hit span two grid points and the next one span three. So a 2:3 rule for the shaker in other words.

By reversing the 3:2 rule for the shaker layer we have created a staggered percussive element. When these two patterns are played together with the kick and snare you get a realyl wonky rhythm.

By reversing the 3:2 rule for the shaker layer we have created a staggered percussive element. When these two patterns are played together with the kick and snare you get a realyl wonky rhythm.

Now it’s super swingy. But WAIT there are a few more steps that will really bring this home…. So here we go. If we zoom in really close to that second shaker we will see that it falls a little before that 1/8 grid marker.

Notice how the second shaker falls just before the 1.1.3 mark.

Notice how the second shaker falls just before the 1.1.3 mark.

Lets take all of the shakers and move them to the left so that the second shaker note falls right on that point.

Lets now slide our shaker notes over to the right so our second shaker now falls directly on the 1.1.3 mark.

Lets now slide our shaker notes over to the right so our second shaker now falls directly on the 1.1.3 mark.

Now lets listen.

Getting sloppy! Now one more step to really give this some swing. Lets mute that first hi hat and replace it with the now late shaker, but instead of leaving that first hit as a shaker, lets make it trigger the hi hat.  Lets also delete that second shaker. You should have something that looks like this now. If you aren’t an Ableton person, the greyed out notes are disabled.

Screen Shot 2019-02-13 at 12.38.15 PM.png

Lets hit play and hear what it sounds like….

Wow. Hear that groove??? Probably sounds familiar doesn’t. Well that’s because it has that Dilla bop because A LOT of his production style was rooted in quintuplet swing. Whether he intentionally knew it or not, that bop came from that swing styling. The above audio example runs through our edited hi hat variation along with the additional shaker part so you can get an idea for what it would sound like had we left the shakers in without changing them to hi hats. We were able to achieve this groove just by drawing in some notes, time stretching, and moving things around all exactly on grid. That’s some pretty powerful stuff if I do say so myself and all using 100% quantization.

Conclusion:

Look I get it. We all want that human sound, but sometimes your musicianship can’t cut it, (which is OK!) or things need to move quickly, in which case quantization can be a great tool. There is a reason it still exists in pretty much every DAW… because it is useful. As with all tools in any creative field, you should learn how to use it with intent, vision and forethought. The more you approach it as a paint brush instead of a crutch, the more you will get out of it. Quantization isn’t for everyone either. Some people don’t need it and others just don’t want it, but some of us do and hopefully with these tips and tricks you will be able to employ quantization with confidence in your next beat. 

Thanks for tuning into the first article of Groove Theory! Two more topics to go yall! Hope you guys enjoyed this post. If you dig it let me know! If you hated it let me know! If you have any ideas or things you want me to cover in a future blog post, let me know!  

I'm on twitterinstagram and facebook, and you can also reach me via email.

Thanks for tuning in!