Just got a Machine Drum - My initial thoughts/mini review

Elektron Gear Machinedrum
#1

I received my Machine Drum SPS-1 MKII today, played around with it for some hours, then wrote up this “mini review”.
I also have an Analog Rytm, so I naturally drew some comparisons.

Here are my initial thoughts, pros and cons:

Pros:
[ul]
[li]Sends out MIDI when patterns are replayed => can use the machine as a scratchpad, record events into PC sequencer, then finetune (AR cannot do that)[/li]
[li]Mutes also work with external MIDI data (AR cannot do that, mutes only affect the internal sequencer)[/li]
[li]Sound synthesis seems rather simple at first but comes alive nicely when building patterns[/li]
[li]16 tracks on the MD instead of 12 (actually 8) on the AR[/li]
[li]16 LFOs, mappable to any track (vs 1 LFO per track on the AR)[/li]
[li]Sounds are not hard-wired to specific tracks (unlike AR)[/li]
[li]Tracks are not hard-wired to specific individual outputs (unlike AR)[/li]
[li]EQ per track (AR has none)[/li]
[li]Simultaneous low + highpass filters per track (AR has one multimode filter per track, i.e. you have to decide which type you want)[/li]
[li]Has MIDI local off => can preprocess parameters before they reach the MD sound engine (AR cannot do that… or can it now?)[/li]
[li]Patterns can be triggered by MIDI note-ons without having to wait for the current pattern to finish playing (AR cannot do that)[/li]
[li]Did not have to look into the manual once. the only thing I had to look up a few hours in was how to do trigless locks (same experience with AR, I found the UI of both machines to be very intuitive)[/li]
[li]Full SysEx support. It’s even printed in the manual, like it should be (AR has sysex, too, but it’s undocumented)[/li]
[li]Can sequence other gear (“MID” machine) (AR cannot do that)[/li]
[/ul]

Cons:

[ul]
[li]No chromatic mode[/li]
[li]No microtiming[/li]
[li]No non-destructive performance controls like on AR (can use CCs to edit all parameters, though)[/li]
[li]Retrig should be available for all tracks - IMHO - but is only available for the sample playback machine(s) for some reason[/li]
[li]No trigless locks on regular tracks (have to use separate controller track)[/li]
[li]No sample replay (yeah, there’s the UW version but I’m not paying 400 bucks extra for a different version of the firmware, sorry)[/li]
[li]No amp envelopes (just a simple decay and the ability to use LFOs as additional “ramp” style envelopes)[/li]
[li]Closely related to that: one shot sounds only, no note duration (usually not needed for percussion, anyway)[/li]
[li]Why is there just one builtin waveform per ROM instrument (E12 bd/sd/hh/…) ? I really was expecting a sample select parameter here (not an advertised feature but still…)[/li]
[/ul]

It does not make sense to compare the actual sound of the two machines. They both sound good but totally different so I’m still going to keep my AR (that was the plan all along).

Maybe one day Elektron addresses some of the above-mentioned issues (mutes/midi out).
Just for clarification: When I say “midi out” I don’t mean that the AR should get actual MIDI sequence tracks like the MD
(it would be nice, though). I just want it to send out MIDI events when it is replaying the regular tracks so they can be recorded and refined in an external sequencer.

So, these are just my initial thoughts a few hours after I received my MD (#117xx, according to the serial nr, made in November 2015).

I have to say that I really like this box. Even the red screen.

Should I ever grow tired of the sound, I can see me continue using it as a scratchpad HW sequencer that can (also) trigger other synths.

That’s not to say that I am expecting to grow tired of it anytime soon. There may be just a few dozen synthesis parameters per track but since there are quite a lot of tracks which can all be EQd individually, layering makes this a rather deep drum synth, after all.
Last but not least, since I’m just a few hours in, I’ve only scratched the tip of the iceberg, I guess.

Incidentally, I could not have picked a better time to get one since Elektron just lowered the price to 699e on this very day.

So, Thank you, Elektron!

~A happy customer.

MD sampling query
#2

Some corrections here:

yeah, there’s the UW version but I’m not paying 400 bucks extra for a different version of the firmware, sorry

I’d like to retract that statement since it is wrong.

The UW version costs 899e, i.e. 200e more than the regular one and also includes the turbo midi interface and a carry bag. That’s reasonable.
It’s the UW+ version that comes with turbo midi and the +drive that costs 400e more.

Regarding this:

Simultaneous low + highpass filters per track (AR has one multimode filter per track, i.e. you have to decide which type you want)

This is a bit wrong. Actually the MD has a bandpass filter per track. Same as the AR. The big difference is that on the MD you can set the bandpass width, on the AR you cannot choose different lo/hi frequencies, just a center frequency.

One thing I’d like to add to the “pros” list is this:
[ul]
[li]Can link track to other track so that both are played if one is triggered via internal sequencer or external MIDI[/li]
[/ul]

I forgot to mention this yesterday but I consider this a very important feature for layered sounds. There’s no other way to prevent MIDI trigger latencies, otherwise.

Another thing for the “cons” list:
[ul]
[li]Cannot delay a track sound (i.e. insert silence at the beginning)[/li]
[/ul]

I tried to work around this by sending the sound into a delay but it turned out that you cannot have a full-wet delay sound.
Turning down the “vol” of the track also silences the delayed sound (a very bad design decision, if I may say so).
Then I tried to assign the track to an unused individual output and also send it to the delay but that does not work, either (it even says so in the manual). This trick works on the AR, btw.

Speaking of the AR, again: Elektron, if you read this, please consider adding a delay parameter to each track.
There’s still one empty knob-slot on the trig page, right beside “len”.
Track-linking like one the MD would be very much appreciated, too.

Regarding the track delay on the MD : Does anyone know if there is any way to achieve this ? I don’t need it for microtiming, I just want a (layered) track to have a constant time offset.

Could this be done on an MD UW ? Maybe by sending the second (layered) track sound to a RAM record machine which is triggered one step before the track sound so that the sample start parameter can be set to include a bit of silence ? Would that work ?

I have another question regarding the RAM record/play machines: If RAM record+play is triggered on the same step, would that work as a pass-thru, i.e. play the sampled sound as if the sampled track was played directly, or are there any latencies ?

btw: I just read that you cannot choose an input track for the RAM recorder – it always records the entire output and/or external input. that seriously limits the use of these recorders. Is this correct or have I overlooked something ?

As you can see, I’m still thinking of exchanging my MD for an MD UW but so far I could not really convince myself that it would be worth it.

Are they any cool tricks you can do with the RAM recorders/players I have not thought about ?
The static sample playback is of no big interest to me since I already have an AR.

RAM recorder applications I can think of so far are:
[ul]
[li]Use the MD as a simple loop recorder[/li]
[li]Use it for stutter/beat repeat effects[/li]
[li]Use one pattern to build a sample-chain set in RAM, then use a different pattern to play these samples[/li]
[/ul]

One final question: Is the RAM for the 4 RAM recorders global ? I.e. is it possible to use one kit to record a sample-chain, then replay it using another kit ?

#3

You have not yet considered connecting one or more individual output to an external input in order to sample selected machines.

Having to use several tracks to achieve these kinds of outcomes is a necessary downside to the ingenious environment that Elektron built in the MD and MM.

Watch carefully for the CUE settings in the RAM REC machines.

I believe so, but if I ever tried that it was many years ago, so please don’t take this as a confirmation.

In general, have a read through Tarekith’s tips and tricks document or Veets’ “Next level” document, which should be downloadable easily, both digests of posts from experts on the defunct-but-still-online Elektron-Users forum.

#4

What about a LFO (square maybe) on VOL so that only the first hit is silenced ?

1 Like
#5

@PeterHanes: Good idea, hadn’t thought of that, indeed. I wonder what the latency of such a loopback would be (it would not matter for delaying tracks, though).

What I’ve been doing is far is using the “local off” feature of the MD and do the micro-delay in an external MIDI processor, which then sends the notes back to the MD.

This works suprisingly well, every hit sounds the same. The snaredrum sound I made for testing this consists of three tracks/layers. The first layer is played at +0 ticks, the second one at +3, the third one at +5 (@ppq=192).

Fortunately, “local off” on the MD means that not only the knobs just send out CCs while being decoupled from the sound engine, but also the internal sequencer just sends out notes. This is very good since the internal sequencer can still be used when pulling this trick.

Problem solved, I guess.

One drawback of this method is that you loose plocks which are not send out via MIDI. Only the MID machine does that but only if “local” is “on”. Guess I found my first MD bug :wink:

@Lying Dalai: I’m already using a square LFO to cut off the sound of the first two layers. Using another LFO to cut off the first part of the sound would not work so well since that would make the sound either very quiet because of the amp envelope, or it would result in a tremolo like effect if you’d increase the amp decay time. There is no one-shot square LFO, unfortunately (the AR has one).

#6

time to update your OS :slight_smile:

2 Likes
#7

I must confess I don’t know the MDUW+…
Yet.

#8

^^ no problem, just try it out yourself and you’ll see, I mean hear

@void: ah, the new trig condition feature. I’ll check that out later.

…in the meantime:

I put on my lab coat (once again) and did some timing measurements, with interesting results:

Test procedure:

[ul]
[li]MD/AR routed into DAW (Reason) via regular audio input (main out L on both machines)[/li]
[li]PC sequencer (not the DAW) triggers both MD/AR and a DAW softsynth (Antidote, also tried Subtractor) click sound[/li]
[li]Known PC seq->DAW latency is ~0.45ms (average, measured with Midi Latency Analyser v2.1.1)[/li]
[li]MD/AR/PCSeq BPM is 133.3, PCSeq ppq=192 => 2.34434 ms/tick[/li]
[li]MIDI songPos (16th=0) + continue commands are sent to MD/AR when replay starts[/li]
[li]Known audio interface input latency is 128 samples @44.1KHz = 2.9ms[/li]
[/ul]

Test results:
[ol]
[li]MD: internal sequencer / internal sync: 6 ticks late (14.06604 ms) (multiple restarts needed to hit exact start time, stays in sync for ~40 seconds)[/li]
[li]MD: internal sequencer / external sync: 1/64th note (12 ticks) late (28.13208 ms)[/li]
[li]MD: external sequencer / external sync: 4…5 ticks late (9.37736 …11.7217 ms)[/li]
[li]AR: internal sequencer / internal sync: 2 ticks late (4.68868 ms) (stays in sync for several pattern iterations, ~20 seconds, plays in sync everytime the replay is started)[/li]
[li]AR: internal sequencer / external sync: 1/64th note (12 ticks late) (28.13208 ms)[/li]
[li]AR: external sequencer / external sync: 2 ticks late (4.68868 ms)[/li]
[/ol]

(take the ms with a grain of salt, I did this by hearing / listening for flams/phasing)

You have to subtract the audio interface latency from the result times, maybe an additional millisecond or two for the (unknown) DAW processing latency of the input signal. The output latency does not matter since both the softsynth click and the input signal are subject to that.

Conclusion:

  1. and 6) look/sound good.
    1. are not too exciting.
  3. and 5) are totally off and unusable.

Has anyone else noticed this “internal sequencer + external sync” lag ? It seems unnecessary, in regard to the other results.

While you can always delay all other synths/sequences to compensate that lag, anything higher than ~5ms means that it will be quite hard to manually play+record keyboards/other controllers in sync.

If you step-sequence all your music, you probably won’t care but I thought some of you may find this interesting.

There is always the possibility of just sampling the output, of course.

#9

p.s.:

“unusable” was not quite the right word. There’s just a constant 1/64 note offset and that numbers also say that MD+AR play perfectly in sync.

I played around with the MD a bit more and also threw the AR into the mix.

I must say that it really was a lot of fun and the two machines complement each other nicely, i.e. the “lo-fi” like MD sounds and the “round” analog sounds of the AR.

Here’s what I came up with. I recorded the MD+AR into Reason and also added a pad and a delayed synth sound, and some EQ and compression:

After this wall of text, I felt it would be nice to post some music, even if it’s just a 2min “test”. At least I always like it when other people do that.

So my final conclusion is: I really like the MD a lot, it’s definitely a keeper.

1 Like
#10

Great review bsp…I just order the standard version.

I have one major question for you-

Because there is no microtiming (a huge disappointment for me) I’ll be triggering the MD from Ableton when I need specifically “off grid” timing. But when I’m triggering via midi will the MD still record automation ( from MD knob movements) or does there need to be a note on the sequencer in order to record knob movements (ie filter sweeps or effects changes)

Thanks

#11

Hi Upright,

the MD will not record any MIDI CC/knob events, just play them back.

Therefore you will need to record those events in Ableton.

What happens if you try to record automation on the MD itself while sending it MIDI notes is that the MD will record these notes, which will result in double triggers.

Apparently there’s no way to turn this off. It also happens when local control is off.

What you can do is use a CTR-8P machine and step-edit plocks (that’s the only way to get “trigless locks” on the MD).

You cannot do that on the voice track since that will automatically insert a note trigger (=> double triggers).

It’s also possible to record these CTR-8P plocks by turning knobs but you must place a trigger on each step manually beforehand.

After you’ve recorded the knob motions, you will have to delete the automatically created / MIDI recorded note triggers on the other tracks, so that can be a bit annoying, depending on how many tracks you use.

Alternatively, you can mute the voice tracks before recording the automation, this will prevent the note triggers from being recorded.

Personally, I think I’m just going to sequence on the MD and record it as audio.

I also have the standard version of the MD and figured that I’ll do the glitch/reverse/resampling/microtiming stuff ITB.

But hey, maybe someone here can still convince me to buy the UW version (because you can do “that” awesome trick with it).

One thing I keep wondering is the RAM machine pass-thru and global RAM behaviour I asked about earlier.

If the “pass-thru” is possible, a RAM machine could be used to filter the bit/rate crusher and also apply an additional envelope to a sound.

What essentially made me choose the non-UW version for now is that you cannot select an input track for the RAM recorder.

A hardware loopback via an individual output, like PeterHanes suggested, would be a workaround, but that limits this to two tracks plus you loose two indiv outputs, and finally there’s probably some sampling latency involved (could be an effect by itself, though).

#12

the UW is the Machinedrum to get. :slight_smile:

is RAM global? yep, you have 4 RAM buffers in the machine, these are shared between kits.

pass-thru? yep, it has a CUE parameter. but IIRC you can do this on the non-UW too, or doesn’t the non-UW have the INPUT machines? not sure.

the glitch stuff you can do with the RAM machines has a unique character… haven’t heard anything which comes close ITB…

the best thing about the RAM machines though, is that you can think of them as 4 independent “step-programmable delays” or something… recording what’s coming out of the machine, feeding it back with all the glitchy goodness. it rocks.

there’s a lot of “happy accident” stuff in there which expands your range of expression with the thing…

it’s the best thing about the MD by far.

#13

The original Mk 1, pre-UW, configuration had external inputs and INP machines.
http://www.soundonsound.com/sos/jan02/articles/elektronsps1.asp

#14

Thanks Bsp,

I appreciate the info. UW is completely off the table for me as I just don’t need the sampling at all, I have plenty samplers as it is lol. I just want highly customizable synth drums and the standard seems to be perfect for that.

#15

Thanks for the answers, void (and PeterHanes, just saw your post when I was about to post this pre-written reply).

With the 4 RAM being shared between kits, creating sample chains with one pattern and replaying them with another should be possible then, right ?
That would be quite cool since it would effectively increase the number of available sounds per pattern (not the number of simultaneous sounds, of course).
Well, 16 is plenty already but it might be worth a shot.

Thanks for the tip with the input machine. I totally overlooked this so far (was probably thinking of the AR inputs which you cannot do much with other than route them into the compressor).

I just played with this and routed outputs C and D back to the inputs. There seems to be no noticable latency (which answers another question of mine).
This is a real useful feature for extra-percussive sounds (two more amp envelopes!) and filtering the SRR.
The sound loses some volume, though, so the volume of all other voices needs to lowered.
The overdrive seems quite useful for bass drum sounds.
Retriggering the filter on a decaying sound is fun.
Last but not least, the INP-Ex machines have a hold parameter for their amp and filter envelopes, which the “voice” machines are missing.

Regarding the resampling effects of the UW: I think I have a pretty good idea what to expect there. A few years ago, I made a sampler softsynth that also supported resampling.
I did not know about the Machinedrum (or Octatrack) back then and my sampler did not have resample trigs (=no immediate feedback) but you could easily resample a whole sequence or even a whole song at the press of a button, then
use the resulting waveform in a sample instrument.
Some of the stuff I did with that used up to 8 feedback resample iterations :wink:
There really is a lot of fun in that, and happy accidents really do happen.
To be honest I’m thinking about making an updated version of that sampler, then route some of the individual outputs of the MD and AR into it.
Just thinking out loud, though :wink:

However, while it would sure be nice to have the resample feature in the box itself, that feature is more like an effect, not to be overused
(there used to be a time when certain music was full of glitches and stutter – I do love IDM but I “hate” that specific effect)
I really cannot imagine this to be the best thing about the MD (UW). It sure is the most unusual feature and I cannot think of any other (HW) synth that has this.
But yeah, there’s more than just glitch/stutter possible with that, of course.

What I find best so far is the EQ per track (+the adjustable filter bandwidth), this is invaluable for “mixing” tracks.
I wish the AR had that. Maybe Elektron could add a new bandpass filter type where the resonance knob acts as the bandpass width ?! That would be so awesome. Not sure if that’s technically possible, though.

What’s also great is the sound diversity of the MD. If you count out the input/mid/ctr/e12 machines, there are 33 different sound synthesis machine types (for comparison, the AR has just 17, and some of them have very few parameters).

In the end, the sound will always have a certain character, of course, and I think that for actual composing, using samples is just more practical.

Note by:
Someone should really make a sampler that takes finger prints of all your samples and is able to find similar samples at the press of a button…
This could work similar to texture generators where you can iteratively choose between several alternatives to finally arrive at the desired result (more or less;)).
Maybe that someone should be me but if anyone else wants to give this a try, go ahead. Good thing that at least the EU does not have software patents.
And now that it is written in a forum, nobody can patent this anymore :smiley:
Well, just thinking out loud, again.

I just found it very disappointing that you cannot at least categorize/tag samples in the AR.
I now have ~500…600mb of samples on it and finding a kick drum, for example, involves a lot of folder browsing.
I know, most software is like that, but that does not mean that it is the right way to do it.

Well, I have around two weeks left to swap the MD for an MD UW, I’ll think about this. At least now I think I have a good idea of the additional things you could do with the UW.

I will probably keep my current MD, though. The input machines already do what I initially thought I needed the UW for, and I suspect that I would not use the resampling that much in practice. Like Upright, I definitely have no need for more sample playback machines.

1 Like
#16

Wow, ingenious summary!

This will help a lot of people, thanks a lot.

#17

[ul]
[li]“No amp envelopes (just a simple decay and the ability to use LFOs as additional “ramp” style envelopes)”[/li]

[/ul]

. does anyone have any tips on how to make soft kick drums? i’m looking to remove the clicks at the start of bass drums,. i’ve just got MDUW and feel a little lost without an attack. i did try making some hits by closing the filter and adjusting the LFO ramp, makes it sound very muddy! ideally i would like another work around. there’s probably something i’m missing? thanks!

#18

Move the start point back past the click.

1 Like
#19

thanks for the input, but unless i’m missing something not all kits have start point?
EFM for example.

#20

Thats correct.
LFO’s are usefull, I mean use an LFO on amp or filter as you know, but use another LFO to control the first LFO for better control.
You have to get creative, you can actualy use all the LFO’s (from all tracks) to control just one track for example.

1 Like