'Machine-Gunning' :: My Explanation

[Pogo]

I've seen a lot of people on these forums try to explain the 'machine-gunning' effect using various concepts, and I have to say, I don't agree with any of them.

Regardless of a sample's attack, or the sensitivity threshold of a pad, I believe you will continue to encounter this problem until there is a feature implemented into your sound module that allows a random sample to be triggered on each strike of corresponding strength.

In real life, no two strikes of a drum sound the same. Every strike produces a different timbre, texture, tone, etc. Volume is merely one of the many aspects that give an instrument's sound its dynamic range and versatility.

And now here we have the electronic drum. If the player repeatedly strikes the drum, maintaining a constant strength of impact, the module will trigger the exact same sample with each strike.

This is what I believe causes the 'machine-gunning' effect. It has nothing to do with maximum polyphony or the attack strength of the original sample, it's the simple fact you're hearing an identical waveform over and over. Your ear drums are receiving a vibration in which your brain automatically detects a linear pattern. The effect becomes more and more predominant as the frequency of the striking increases.

This is a hurdle I am yet to see jumped over in the field of electronic drums, and what frustrates me is that such an achievement would be far from technologically challenging. I believe this problem could be granulised if not eliminated by recording not one but several samples of a drum being struck at a certain strength, from which the module could randomly trigger on each strike the player makes of corresponding strength.

[SuperPuss]

I absolutely agree with you about what causes machine gunning, Pogo.

One of the reasons I really love the Roland system is that it does not use samples for the majority of the kick, snare, toms and cymbals. Most of these sounds are created using modelling technology Roland refer to as COSM (Composite Object Sound Modelling).
When coupled with Roland's positional detection and interval control and with the pad settings adjusted correctly, it is nearly impossible to get the exact same waveform repeating.
That, to me, is the biggest drawcard for the TD12 and TD20 modules.

Some sampling programs use multi sample technology to achieve a more authentic result than traditional single and layered sampling technology. These have multiple samples at each velocity that are randomly switched between hits to ensure that the human brain does not detect 'machine gun' effects.

[drumslinger50]

This is very interesting to me. You both have a better technical grasp of an old problem. I have not encountered this with my TD-6V but have heard it on other modules. In the 1990's I did a lot of duo work using a drum machine. To compensate and be more life like I would run the drum machine out put through a Chorus or a Phase Shifter on a wide but minimum setting. It made every beat a bit different.

[Colquhoun]

If the player repeatedly strikes the drum, maintaining a constant strength of impact, the module will trigger the exact same sample with each strike.

If the player repeatedly strikes the drum, maintaining a constant strength of impact,.....he or she is playing a very boring lick.
And thus, playing with more dynamics reduces the chance of machine-gunning.

But I agree, this should happen less with more modern technology.

[Chris Jude]

I would suggest that Pogo's explanation is pretty consistent with how it is commonly described these days, and that the phenomenon has long since been eliminated in most of the higher end edrumming software exactly as he suggests.

[Pogo]

Most of these sounds are created using modelling technology Roland refer to as COSM (Composite Object Sound Modelling).

Yet I have heard many demonstrations of the TD-20, some even by Omar Hakim for Roland officially, in which I still detect this same naggingly artificial effect. This leads me to believe the COSM technology was designed only to simulate the dynamic range in volume of a sound source, and not the timbre, tone and texture as I mentioned in my original post. If so, I'm disappointed. If Roland have restricted themselves to this level for a reason other than to minimise file size, I'd love to hear it.

This effect is the most influential artificiality I have heard throughout the V-Drums series so far. I really hope it's something that's tended to in the future.

[SuperPuss]

You're probably most likely to have noticed it on the toms which, on TD12 and TD20, don't have the positional detection feature on the heads. Strangely, the earlier TD10 module DOES have this on the tom heads. I have no idea why Roland dropped it on the newer modules.
The snare does not suffer from machine gunning (at least mine doesn't) because it DOES have position and interval control on the snare which changes both volume and timbre.

There's no mention of it in the FAQ's but I'm hoping the TDW20 expansion board brings posi detection back to the tom heads.

[groovy7]

I so agree with Pogo...I have a td 20 kit and the machine gunning is terrible given what I paid for it. First of all, the kits that come pre-programmed in the module are mostly not usable. Those programmed kits all demonstrate the evil machine gunning effect to perfection. It is awful, live or on recordings. I use my kit for personal practice too and it is a dream come true in most respects, but the machine gunning has got to go. I know, I know, some of you will say I shouldn't use fills that make it so apparent. What about snare drum rolls?....those are supposed to be a drummers mainstay, right? Ok, no snare drum rolls. Makes no sense to me to have to change my playing to try to somehow hide the drums machine guns. Roland please fix this, it is awful. Especially after a 5 grand investment. Postional sensing....never heard of it....doesn't work. It's just a volume scam, and hopefully the edge/rim trigger will mix with the center trigger and maybe "fool" away those machine gun blues. It Just Doesn't Work. One word...Randomizer...If something like a "randomizer" is put into use in a module, it could be the best sounding module on the planet.

[vdrummer20]

I think some of this is related the fact that on an acoustic drum, as you hit the head and it starts to vibrate, the second and subsequent hits will result in slightly different harmonics and the like since the vibrating head has different acoustical properties then the head at rest. I also think that the harder you hit the head, the more deformation occurs which again will change the sound.

This has similar overtones of the positional sensing post from a few months ago (ducking).

Having some sort of ramdomness added in terms of the harmonics of the sound as well as the volume would hide this. Perhaps that is why it is minimized on the snare with interval control and POS.

[vjboc]

I have noticed if you set the sensitivity to where it does not peak at all , You dont get that machine gun effect. Once the module peaks and you do a single stroke roll, You hit the ceiling of sound thus causes machine gun like sounds. Try setting the sensitivity to where your hardest hit is slightly before peaking. It does make a difference. I have the td-8

[SuperPuss]

Groovy, it sounds to me like you don't have your drum heads tensioned properly. If they are too loose, you won't get position detection and, as VJ mentioned, if your sensitivity is too high, you will get machine gunning.

I have just uploaded a new clip where I do snare rolls in the middle and some flams and other rolls and, I don't know about you but I don't hear machine gunning.
HERE'S the clip.

[Joe_K]

I would suggest that Pogo's explanation is pretty consistent with how it is commonly described these days, and that the phenomenon has long since been eliminated in most of the higher end edrumming software exactly as he suggests.

My thoughts exactly. It has been common knowledge for a long time that machine gunning is due to the exact sample being heard over and over at the same velocity (especially if the sensitivity is so high on the pad that it is easy to reach velocity 127 without hitting too hard ... because leveling out at max velocity is a sure fire way to experience machine gunning.).

And software solutions (dfh Superior, BFD, EZdrummer, Addictive Drums) have shown the way to much more natural sounding dynamics. I don't just notice the same sample being played anymore in my older recordings when I hear a couple fast strokes near the same velocity, I hear them now on the snare even just pounding out the rhythm (don't even get me started on the ride ... ugh!). Too bad not everyone has the time or desire to create a software based system to experience this for themselves (or then do their own mixing and processing) ... but its getting easier all the time!

[groovy7]

Vdrummer...deformation of the acoustic head, then striking the drum head again while in motion from the initial note; thats exactly why an acoustic drum can sound so natural. That was a perfect description of what happens you gave. I think (wish) machine gunning could be addressed in a module design or expansion card. And yes, I have tuned into some of the talk about the never-seeming-to-arrive td 20 expansion card. Will they put somethig like that (randomizer) on it? I guess thats anyone's guess.

vjboc...I will try the trigger sensitivity idea you mention. Without having tried it yet, I can't say much, but I imagine it should decrease the attack part of the sound, because I have noticed that a peaked trigger signal lays into the attack, which possibly accentuates the machine gun thing. ...It's an idea I will surely try...thank you. And are all those my space tracks with the td7 module? Sounds good.

Superpuss... I saw the cy-12 fix you did. I like that a lot. I have a cy-15, will it work the same for that cymbal? There are no dynamics on the bell, just as with the cy-12. And the tension on my heads are factory, but I suppose that doesn't necessarily mean they are correctly tensioned. I will use the recommended tensioning in the manual, and make sure they are correct. Checked out your video...very nice playing. Center hi hat...I like that. I still believe that crossed arms for hi hat playing was a limitation of equipment, and not a deliberate "technique," and in turn, it limits drummers. I still cross my arms though...

Maybe machine gunning is a misnomer in a way. Imagine having the "randomizer" thing in place and even snare back-beats which are plenty far apart wouldn't sound so identical. And here's a few peculiarities about some of this: Remember in school when the drum instructor taught us to play evenly and smooth? This was to try to avoid sounding un-musical on our crazy, unpredictable acoustic drums of course....lol, and I believe in playing evenly and smooth on any type of drum. And another contrast is, I am perfectly conditioned to hearing kick drums sound repetitive, as the beater always hits the same spot, and is often compressed by recording engineers to sound pretty much the same on each note. I don't mind my td 20 sounding this way with the kick drum. But I will probably never be able to truly enjoy hearing toms and snare drums sounding like that, or hi hats or rides or cowbells. They do it too.

[SuperPuss]

Groovy, cheers for the thumbs up on my video. I find the center hihat to be far more natural but I have only been playing for 12 months and the majority of that time with that configuration so it is natural to me.
I do play acoustic kit in a conventional (arms crossed) configuration with a band I jam with on weekends and that is not a problem. Playing on center hat position most of the time seems to have not affected me on the A kit
Yes, the cy12 fix will work on the 15 too. Also, don't be afraid to go tighter than Roland recommended tensioning. The Roland heads are almost unbreakable in that regard and will respond much better when tighter. That will also make them feel less springy and more natural. You'll probably have to lower your trigger sensitivity as you tighten too which will also help with dynamics.

[vjboc]

Nothing compares to hopping on a nice set of acoustic drums when it comes to A's and E drums. I would not go by the manual when setting up an e kit. Every drummer has his/her own style, I believe you can spend a tremendice amount of time tweaking settings to make it feel close to an A kit. But it will never be an A kit, Just like an electric piano will never be an A piano.

I also think another big problem is people compare the 2 two kits as if they are sitting in a room listening to them side by side. I would think the best way to compare the 2 is take an e kit and an A kit that is miked ,Gated ,Compressed, EQed add the necessary effects (reverb, Delay, etc) plug them into a P.A. or recorder. Then compare the 2 sounds coming out of the speakers. You would be surprised how close they sound.

It is just like going into a local recording studio. Alot of times bands that have recorded in the same studio often have a simular drum sounds, but yet each band had adifferent make of drums and cymbals. Because most of the time the engineer has a dialed in setting for the drums and the type of gear the studio has that will give any drum that certain simular tone. Even pro bands will look for certain engineers when they are looking for a specific drum sound.

I would work on each drum setting until you get the most playability out of it. Preferrably starting with the snare. Because I think the snare is the heart of a kit.

Good Luck