How important are a few milliseconds? When it comes to sequencers, they can mean the difference between a perfect groove and a sloppy mess. We talk to drum machine pioneer Roger Linn to discover the secrets of perfect musical timing and find out why the swing of his MPC sampler series remains so highly rated.
While researching our recent Passing Notes feature on DAW and drum machine swing, we found ourselves returning to the same few questions over and over again. Do dedicated hardware drum machines and samplers offer better timing and swing than software-based alternatives? Does swing do exactly the same thing in all DAWs and sequencers? What else contributes to the groove of a beat or loop?
Perhaps above all else, one of the main questions we found ourselves asking was why the earlier models in Akai’s MPC series of sampling workstations are still considered to be among the best around when it comes to groove and timing.
The idea of ‘MPC swing’ is so ingrained in electronic music folklore that it’s considered almost heretical to question it – especially among large chunks of the hip hop production community. But can it really be that hardware sold a quarter of a century ago is somehow better than the latest DAWs and drum software?
Eventually it struck us that there was only one person who’d be able to give us definitive answers to all our queries: Roger Linn, inventor of swing and quantisation, designer of the ground-breaking Linn sample-based drum machines, the Akai MPC60 and MPC3000 and the DSI Tempest.
Over a series of emails with Roger we discussed sequencer swing and drum machine timing, starting with the question of the MPC’s magical groove and moving on to the importance of tight timing and whether it’s better to play beats in real time than to program them with quantisation.
only one person could give us definitive answers to all our queries: the inventor of swing, quantisation, the LinnDrum, the MPC and the DSI Tempest.
Roger’s insight into sequencing, swing and timing helps shed light on some of the most commonly debated intricacies of electronic music production. Some of his answers may also come as a surprise to anyone who believes there’s something unique about the groove of the MPC.
Attack Magazine: The main idea which prompted this discussion was a conversation we had in our office about the magic of the 80s MPCs. 25 years on from the release of the MPC60, it’s still held in incredibly high regard and there’s a mystique attached to the timing. ‘MPC swing’ is the watchword for tight, funky timing. Can we start by talking about the MPCs – particularly the MPC60 and MPC3000 – and why they’re so highly revered? Are they actually measurably better in some way than other sequencers or is it a bit of an urban myth which has snowballed since the 80s?
Roger Linn: There are a few factors that have contributed to natural, human-feeling grooves in my drum machines. In order of importance:
Swing – applied to quantized 16th-note beats – is a big part of it. My implementation of swing has always been very simple: I merely delay the second 16th note within each 8th note. In other words, I delay all the even-numbered 16th notes within the beat (2, 4, 6, 8, etc.) In my products I describe the swing amount in terms of the ratio of time duration between the first and second 16th notes within each 8th note. For example, 50% is no swing, meaning that both 16th notes within each 8th note are given equal timing. And 66% means perfect triplet swing, meaning that the first 16th note of each pair gets 2/3 of the time, and the second 16th note gets 1/3, so the second 16th note falls on a perfect 8th note triplet. The fun comes in the in-between settings. For example, a 90 BPM swing groove will feel looser at 62% than at a perfect swing setting of 66%. And for straight 16th-note beats (no swing), a swing setting of 54% will loosen up the feel without it sounding like swing. Between 50% and around 70% are lots of wonderful little settings that, for a particular beat and tempo, can change a rigid beat into something that makes people move. And unlike the MPCs, my new Tempest drum machine makes it very easy to find the right swing setting because you can adjust the swing knob in real time while the beat plays. I first introduced swing – as well as recording quantization – in my 1979 drum machine, the LM-1 Drum Computer.
2. Natural dynamic response on drum pads.
When a real drummer plays a great-sounding groove containing 16th-note hi-hats, he varies the loudness of each hi-hat hit in a way that he has developed over years of practice. If he were to play each note at exactly the same volume, he’d sound like a bad drum machine beat. So to create a natural-sounding 16th note hi-hat part on a drum machine, you need drum pads with accurate and natural-feeling dynamic response (something I introduced on my Linn9000 drum machine in 1984) and you need good drumming skills. If so, then when you program in a hi-hat part with good dynamics from note to note, it plays back as you intended.
3. Pressure-sensitive note repeat (also called roll).
Good pad dynamic response is important, but not everyone has the drumming skills to program a dynamic, natural-sounding 16th note hi-hat part into a drum machine in real time. This is why I introduced the pressure-sensitive ‘note repeat’ feature (sometimes called ‘roll’) in the Linn9000 and have had it in every drum machine since. The note repeat feature allows you to program repeating notes – such as 16th note hi-hats – merely by varying your finger pressure on the hi-hat pad in real time as the beat or metronome plays. At the moment of each successive 16th note, a new hi-hat note is recorded into the beat, and your finger pressure at that moment is used as the velocity level for the new note. This requires far less skill to create a natural-sounding groove. In fact, simply by varying your finger pressure from light to medium continuously and somewhat randomly, you’ll end up with a very natural-sounding hi-hat part.
4. The playback timing should be very accurate.
In my drum machines, I wrote the software in such a way that the notes play exactly at the correct timing location. And for the included drum sounds, I insured that the beginnings of the samples were closely trimmed to minimize any delay at the start. I’ve heard lots of theories over the years about other timing tricks, like introducing random timing variations into the notes of the beat, or delaying the snare on 2 and 4, but I’ve never found these to do much good. In fact, I’d suggest that if the note dynamics and swing are right, then the groove works best when the notes are played at exactly the perfect time slots.
5. You need good beats.
All the swing, dynamics and other tricks won’t do you any good unless you come up with a good beat in the first place. On all my drum machines, I’ve tried to provide factory sounds and beats that sound great and provide excellent examples of what can be done. Even on my AdrenaLinn III Guitar Processor, which includes a simple and very limited drum machine, its 190 included factory beats have very good grooves. In making these beats, I listened to hundreds of great recordings over the years and tried my best to dissect the drummers’ feels and recreate them as drum machine beats. I suggest this same approach in learning how to create great drum machine grooves.
6. I try not to let the technology get in the way of music-making.
On all my drum machines, I tried to design them to be easy to use so you could take it out of the box, turn it on and start making beats. I’ve seen some products that make you read manuals and set up all kinds of complex and unnecessary settings before you can start. By the time you’ve done this, your inspiration is usually gone and you’ve set up a mental association between the product and suffering. I think perhaps one reason people have made so many great beats on my products is that when they have the idea, they can record it into the machine right away.
That’s about it. Not much magic, just good solid engineering and creative beat-making. If some people think my products have a special magic, I take that as a high compliment. I think some people just happened to create a magical recording years ago on one of my drum machines, and in their minds my products will always be associated with great grooves. Years later if they decide to buy a new drum machine, they may see my name on the panel and think: “I can research this thoroughly or I can just buy Roger’s product because I trust that if he did it before, he’ll do it again.” Sometimes customers tell me this and it makes me feel very good.
…and once more Attack Magazine shows why it’s the best dance music resource on the www. Outstanding.
Wow. Fantastic interview, and hats off to the main man of swing for sharing his knowledge and passion. A true gent of our world.
i’m with edward, you guys are doing a phenomenal job. great read.
Yeah fantastic interview, asked the exact questions I wanted answered..
heres what he means with Pressure-sensitive note repeat (also called roll).
Roger is very considered and accurate thinker but couldn’t help but think there was a mismatch going on here regarding computer timing. Roger was specifically talking about software drum machines but I think the interviewer was hinting at timing being generated out of the computer – with regards to sequencing out board gear , especially through a USB midi device as there remains problems there. Anyway it would have been good to have clarified this point.
Roll exist in Ableton Push, so its not true that the controllersdont have that option.
You can´t be serious?! Why have you deleted my post here???
B-BOY TECH REPORT: What brought on the split between Roger Linn and Akai?
ROGER LINN: Akai went out of business and the assets were purchase by Numark, headed up by a very unscrupulous fellow named Jack O’Donnell. Once he bought Akai, he immediately stopped my royalty payments, refused to take my calls and had his lawyer send me threatening letters. I checked around and learned that he has a reputation of being a real bastard, so given that challenging him would have been long and expensive, I let it go.
B-BOY TECH REPORT: What are your thoughts on the more recent lines of MPCs (the stand alone boxes and the new hybrid controllers)?
ROGER LINN: From what I’ve seen, Akai seems to be making slight changes to my old 1986 designs for the original MPC, basically rearranging the deck chairs on the Titanic.
While lasse comment is far off topic. I couldn’t agree more with what roger says there..
Great interview btw
> lasse comment is far off topic
Are you kiddin´? No, it´s not! By comparing his machine to the Titanic, he basically says that the MPC concept is dead! Wow! Mr. Scarth should have asked him why he thinks that. A missed opportunity.
I don’t think his point was that the MPC is dead by using that Titanic analogy LOL! You entirely missed what he meant: ie- the titanic is a big ship – correct? Yes. The MPC is a very complicated system– a BIG system shall we say… Arranging the deck chairs on the Titanic is essentially pointless if you’re going to suggest that you’ve made “improvements” on the technology of ocean liners. In the same way what he meant was that what AKAI has done since the “bastard” took over is a bunch of hyped new features that amount to very little as far as the main system and the main features which are adored: thus, its like “rearranging the chairs on the Titanic” it make no difference to the actual internal structure or the performance as a whole of the machine.
Hahahaha: I love I though! ARE you KIDDING?he basically said the MPCs are about to hit a huge Iceberg, anyone near an MPC get a life jacket!!!!!
Anyway that strengthens my point, if your song sucks (is gonna hit an iceberg) what difference would It bloody make that someone had rearranged some deck chairs? It won’t save your song (or ship) that’s for sure….
BS! You think you’re in the know? Think again, CraftLove!
“to rearrange the deck chairs on the Titanic… (idiomatic) To do something pointless or insignificant THAT WILL SOON BE OVERTAKEN BY EVENTS, or that contributes nothing to the solution of a current problem”
Something (buggy MPC Software) that will soon be overtaken by events (suicide of Akai´s one and only maintenance programmer Pete Goodliffe?) Just jokin´. By the way, brandnew v1.4 of the MPC Software ist still unstable as hell….
“For some reason my 1.4 crashes after I close out the program on almost every attempt.. Even when I don’t have any plugins open.”
“the program randomly crashes sometimes when i’m in the sample edit mode, switching through to chop..”
“same here ! (…) Not Working:
MIDI Clock sending
MUTE/SOLO Hardware / GUI not synced.
(confusing… if you like to use the mouse and the hardware to solo or mute tracks)
WINDOW/FULL SCREEN not activate-able by double-click
“When I try to save a project where I’ve made a beat using 3 or 4 vstplugins, when I save the project, and load it back up again, for some reason it’s as if my last 10 or 15 saves were not saved, so I basically lost the beat, or most of it. So when it opens up its my project after I’ve already saved, and load again its like my last actions werent saved, I’m missing the last two plug ins I inserted on 2 tracks and 2 sequences are missing. Also the sequence was reverted back to an old sequence from when I had saved the project hours earlier. Save function is acting a little funny for me. Also, I pressed undo once and it stepped back so far that I lost a whole pattern on one track, and the last track I was working on. The vst instrument changed presets and also the pattern changed to something I had deleted like 5 minutes prior. So undo is definitely still a mess. And saving a project, sometimes it saves properly, but it stops saving after a while even if it says it saved. It will load back up and bring you back to a prior state way way before the last actual time you saved in that project.”
“MPC STUDIO Freezes when tweaking FOCUSRITE SCARLETT EQ PLUGIN while hitting the pad to hear the sample’s changes.
Program crashes when I close. This is happening after I used the MPC 809 VST.
I shut down the program, kept the MPC Studio ON.
Restarted the MPC Software.
The MPC Software doesn’t detect the MPC Hardware.
Hardware could not be detected.
“Here’s what ALWAYS happens..
When I open my last saved project, I have to open it TWICE in order for it to load all the samples. Double-clicking on my project file “my project 1.xpj” should open up my project with all the samples loaded, but it doesn’t load anything but just the plain software. I have to double-click a 2nd time – pretty much opening the project 2 times..
1. double-click on “my project 1.xpj”
2. it opens MPC software
3. nothing loads.. just an empty template!
I have to double-click on “my project 1.xpj” again for it to finally load my saved project.
This happens both:
A. when I open my last project from my desktop
B. when the MPC Software already opened.
Q-Link 1ST KNOB DOES NOT WORK . BUT THE OTHER 3 DO.
Gee, Roger´s implementation of MIDI has worsened!
Akai MPC 3000 Internal Sync:
Maximum variation between any two consecutive Sixteenth Note intervals:
3 Samples [0.06ms]
DSI/RLD Tempest Internal Sync:
Maximum variation between any two consecutive Sixteenth Note intervals:
16 Samples [0.33ms]
Akai MPC 3000 External Sample Accurate Midi Clock Sync:
Maximum variation between any two consecutive Sixteenth Note intervals:
13 Samples [0.27ms]
DSI/RLD Tempest External Sample Accurate Midi Clock Sync:
Maximum variation between any two consecutive Sixteenth Note intervals:
33 Samples [0.69ms]
P.S.: “to rearrange the deck chairs on the Titanic… (idiomatic) To do something pointless or insignificant THAT WILL SOON BE OVERTAKEN BY EVENTS, or that contributes nothing to the solution of a current problem”
“my experience within 1.4 can be summed as: one step forward and two step backward. like i said on other threads nice impruvments feature wise, but new bugs, much worse plug-in compatibility (osx and au) and too many crashes which before (1.3.1) i never had. Thats a shame, im done really, put the ren on the second hand market. Just updated to Live 9 and komplete 9 and gonna take out from the closet my ol mpd 32. when i fancy hw im gonna use the jjos 2500. happy for the user that dont have this problems and enjoy their device, me im loosing too much time and tracks need to be done. peace.”
—-andreat668 @ MPC Forums
There is actually a bit of “secret sauce” (albeit unintentional) to the early machines (LM-1 and Linn Drum) that he didn’t mention. Some of the samples (mainly the snare) are delayed by up to about 10 ms due to lack of proper trimming. And this is affected by tuning as well. Check the EPROM dumps for yourself.
And regarding randomizing and so on of timing, what most people think of as “randomization” is a uniformly distributed random offset added to each drum hit. This is identical to poor MIDI timing (jitter) and is exactly what you don’t want. It sounds horrible even in very small amounts. A study of timing accuracy in actual musicians (too lazy to look up the source right now) concluded that it’s actually something like 1/f noise. So the timing of each instrument will drift a little bit relative to the others, but it’s correlated over time, so it drifts slowly ahead or behind the beat within a certain window. If you want to artificially add “feel” to a quantized pattern, try that.
Great interview..got so much information wich has never been written/documented before..I have always tought that there is nothin special about MPC timing..it’s all about the workflow..
I don’t know how it works about LM-1 or Lin Drum.
It’s my insteresting in musical electronic like a
keyboard who can play a song automatically by
rhythm pattern programmable.
I assume the interviewer was talking about MIDI output timing from software to hardware.
This is indeed a problem which still exists…. There are of course solutions for clock sync to DAW (Innerclock, ACME, Expert Sleepers). However, I’ve yet to discover a secure method of MIDI note output from a DAW to a drum machine or synth.
As a recording engineer & producer with roots going back to the 80’s, I can assure you that many classic drum machine grooves involved running various tracks of a drum machine through delays or other effects that had an impact on timing (including compression that modified the attack timing). The most obvious one being knocking the snare back by x milliseconds. Extremely common also was getting the 16ths – or even 8ths – from the hat or perc by using a delay instead of actually programming the notes. There is a certain vibe to this that you can immediately recognize once you’ve done it a couple of times. Using delay units opens up the possibility of 1) extremely fine adjustments to swing, etc 2) realtime variation in timing, 3) tone changes resulting from the sound of the effects devices (Space Echo distortion, etc etc). Combine this with the ability to automate these (or punch in on tape originally) and there are vast possibilities for making unique sounding records. By the mid-80’s these techniques were standard with anyone who was serious. So there is more to some of those sick grooves than it might appear. Having said that, some classic 80’s grooves are indeed just a drum machine printed straight to tape….
Wow Ted – great info and insights – thank you!
Nice to have some of the mystique surrounding the MPC swing debunked.
“In newer, faster computers and especially Macs running newer drum machine software, it [MIDI Click Jitter] doesn’t seem to be an issue.”
I’m sure Roger has some great MIDI equipment in his studio, but for most, life is not so simple. Jitter (distinct from latency) is is a HUGE issue, and I would argue that this is one of the main reasons that a lot of producers prefer to stay “In the box” rather than work with MIDI instruments.
With good equipment and a little know-how, you can get good MIDI timing out of a computer, but your average just-starting-out producer with a USB interface and potentially problematic software (ahem, ableton), faces a world of pain getting tight MIDI timing.
Without getting too technical, ableton with a generic USB-MIDI interface… expect >9ms RMS jitter. That’s like having an uncontrollable shuffle of +/-7%, and sounds very sloppy. MOTU and Innerclock both make good correction systems. These are not snake-oil. In some real-world testing, for example, with a MOTU interface and DIgital Performer I saw just about 1ms jitter. Not quite as good as they claim, but certainly listenable.
There are a lot of MIDI jitter myths, and a lot of people who just don’t need rock-solid MIDI timing don’t notice a problem at all. This, combined with the fact that software, drivers, interface hardware, and the USB bus can all be contributing factors can make solving jitter a real headache. Syncing drum machines to software like ableton is one of the prime cases where this comes up, and I would have expected a little more from the #1 drum machine guy on this problem.
Ted for PM >>
I own an MPC 60 MK 1 & MK 2 and have owned MPC 1000 & 3000’s… The 3000 too is great, but when I first tried my hands on MPC 1000, I was really disappointed…. at first I assumed i was just missing Roger Linn’s fine signature(!!!) but later I realised one important component was missing from the MPC 1000 that all the REAL MPC’s (the 60’s and 3000) have…
Odd enough, that very same component is at the heart of every machine that Hip-Hop has ever coveted – the AKAI MPC 60 – 3000, the EMU SP 12 – 1200, the TECHNICS SL 1200 – 1210 –
and that component is simply a (resonating) QUARTZ CRYSTAL CLOCK!!!!! The clock provides a real world clock, and in fact a resonating clock source. A PHYSICAL RESONATING LINK BETWEEN THE VITUAL STUDIO & THE REAL WORLD (if you wish to see it that way) ….
But ANYONE with a superior sense of timing will immediately notice the addition of a quartz crystal reference clock to any virtual studio or turntable… and it completely eliminates jitter, as there is none on a proper MPC…
Nail on the head Brendan Clarke
Bring it on roger, basic sampling and sequencing the bread and butter of what the classic mpcs have.
A means to back up and store data on a format that will be long lasting.
And the mases shall buy.
Do it Roger , we are behind you.
can anyone tell me (without opening a can of worms) why midi accuracy is less of a problem on a Mac than on a Win-PC? Is he talking about midi-recording which involves latency or about “in the box” sequencing in a virtual drum machine like Geist, without an external controller?
Thanks for sharing the interview Attack,…most revealing. Ted, a poster above, is absolutely correct. Adding delays into the equation change the quantisation game dramatically, subtly sent or sent via automation. Also, they include delay lines as well, but these audio buffet delay effects can add some special -subtle- sauce to hat and snare sequences.
Also found it interesting what poster SN says above about the lack of careful trimming of the original samples. You would think that would be a critical aspect when considering timing tests, however just a few ms never killed the drummer 😉 Instead, I can see where it might result in some interesting subtle groove and transient variation. That actually just gave me an idea. Thanks SN 😉
MIDI clock is 24 PPQN with a “ceiling” for baud-rate, which means if you send a ton of MIDI information down the same stream/cable it can max out the baud-rate which creates jitter/inaccuracies (and sometimes on older synths too much information can cause hanging notes, etc.*) (*MIDI Filters could fix that)
It’s always been a good habit to send out Sync signals and clocks on a seperate port so ONLY that goes through a cable to, say a drum machine which is to be synced.
The note on the Atari there (along with other older/ancient computers with built-in MIDI ports, got a Yamaha C1 here) can be looked at this way (given this a lot of thought, as I’m researching PPQN):
A) The Atari has the MIDI port hard-wired to the motherboard, so it’s “second nature” for the computer to send out information through that port.
(like a printer port, it’s a steady stream of information with no trouble at all,
if the driver is set up correctly)
B) This is the big-whopper! PPQN values of sequencer software. Nowadays you have up to 4096 PPQN coming from a software sequencer, that’s a SHIT-TON of pulses per micro-second. The older hardware had MUCH lower PPQN rates; 24/48/96/192 which makes it “easier” for the CPU to output/handle that data = steadier stream of information = better timing.
This leaves the question; how would a modern day computer with hardwired MIDI ports work with a 48/96 PPQN sequencer? It would definitely handle the events in a much easier fashion because it’s less information/stress on the computer/CPU.
USB drivers on the other hand, I can’t comment on.
PCI ports (Express/34 on some Macbooks) are supposed to be more steady, because that port is hardwired to the motherboard. But there aren’t a lot of MIDI interfaces out there for that standard.
The computers got faster/better, but the PPQN values rose alongside the “exponential” growth of computer hardware. So it’s no wonder that we still have some MIDI latency and jitter.
Imagine it in the opposite direction, how would an Atari handle a 4096 PPQN sequencer? It probably wouldn’t! hehe
Nice information about the Quartz crystals, because I’m researching which clock/timer/CPU to use for a hardware sequencer.
The comment Roger Linn makes about modern computer midi timing not being an issue is flat out wrong. Atari’s and Amigas have midi out/in timing that is an order of magnitude better than current computers. Macs are no better than windows and actually worse depending on the osx version. The issue in simple terms has to do with buffers and the timing window for handling the midi data. The sequencer software being used makes a difference too with some having awful midi timing(ableton Live) but the limiting factor is the os which permits at best 1-2ms timing swings between successive note ons on a regular basis. This is measurable. Larger ppqs give potentially greater resolution but don’t affect this basic fact making the gains nil. You can crank up the tempo of an atari or amiga software with lower ppq to give resolution equal to or better than midi and it will remain rock solid. None of this should be confused with sample accurate timing when rendering internally in a daw which is real. All this only pertains to recording and playing back over midi. And this has nothing to do with latency. This is entirely about the ability to play notes where they should be RELATIVE to the other notes in the sequence which is more properly referred to as midi jitter or instability.
I really wish Roger would release a new 4×4 pad digital drum machine a modern MPC if you will. I will never sell my old MPCs anyway but ideally Id like a new machine with lots of real time manipulation options and effects. If anyone can deliver I believe Roger can.
Roger Linn babbles nonsense
Groove has course what to do with musical micro timing in 16th notes pattern
not only clock on 4th note
otherwise no band in the world would grooving
and what band is playing sample conform
random is not the thing
musical micro timing in repetition
and the balance of micro timings of various instruments
that’s all, and different every time
There is a feeling action
Groove is a feeling
mpcs has timing variations
you can test this in audio
Roger linn is a gentleman. Realy.
I am a profesional bass and trumpet player 16 years now and im telling you its not magic just read books about jazz timing and articulation its all in there.
Funny but all my life i thought swing was invented by black people in africa 100s or 1000s years before slavedrivers took them to america. But now i know it was Roger who invented it lol. Dont get me wrong i love electonic music in free time i am working on electonic jazz fusion longplay CD, but if you electonic musicans will not start to learn music history and teory this music will not get far. Sory for my bad english its not my first Language.
i think the most people with problems don’t realize that a good pc is needed to (no dell and hp shit) configr your own choose each component wisely and (look how well the drivers and stuff works) in case an mac and a win is closley the same thery have a kind of diffrent working things (like processing things after and simultan but this is just not the important point the harware under it is still the same but on windows you have to look and “clean” it after a time like you would clean your room and and and there so much things to look at your software can be as perfect as hell(not the case) but it won’t work maybe your motherboard battery lacks on V’s (those saving issues down there could be explained with that) and this list goes on and on so… it happned at least myself i was so in the “things” that i haven’t looked to the hardware and have loosed a cuple of hours searching it on my devices or settings but at least it was just the case that fastboot (bios setting so the pc don’t need always to load all data at new like a ram) so something wasn’t saved perfectly wich cased my system instable after i changed this everything was ok:)
so have a nice day and keep this in mind 😉
(im sorry im not english i hope you can exuse 🙂 )
so 75 % swing is the maximum value of swig in an MPC ?
Does that mean the following is correct :
MPC: 50% equals to 0%: NI Maschine
MPC 75% equals to 100%: NI Maschien Swing
Or does the Maschine have 100% Swing as well ?