Elektron-Users

Munchen wrote:
elektron is a top clothing manufacturer that ALMOST cracked the " best ever drum machine " .

I'm getting tired of paying full price for unperfected machines knowing full well I'm just paying to keep the R&D team in cool tech clothing while there getting blown out in the minimal tech club circuit . . .

Oh well , it's back to getting my hands on florians mr. lab
. . . Think I'll need to get out the chloroform . . .



:roll:

I just finished some tests on MDUW MKII. I'm witnessing maximal +79 sample gaps at 48khz sampling which translates to 1,65ms jitter.

I used heavy load (all 16 tracks going off 50% steps at random) and just GND:IMP going on. No noticiable difference in the pattern.

With 1,65 ms jitter I really can live with it.

Hope this helps. You do your own conclusions...

- sampsa

Just GND:IMP

Gap measured / diff to 24000
24037 +37
23933 -20
24002 -18
23968 -50
24107 +57
23933 -10
24002 -8
23968 -40

Heavy load

Gap measured / diff to 24000
24002 +2
24038 +40
23933 -27
24106 +79
23968 +47
24029 +76
24003 +79
23968 +47

EDIT: My jitter calculations are not correct, the actual jitter is smaller than the one I present here. Need to calculate jitter from the grid not from the last sample (which already might contain jitter)...

wooooo! sounds tight! :-D thats tight right?

I dont know so much about this stuff and am wondering what these jitter numbers translate to in the real world... is this number for the MD considered REALLY good, good, average,ok ,or sub par? i dont really know. I looked at the original timing thread, but it flummoxed my brain.

Using the same test, what would the jitter be for other gear... like roland drum machines... linn drum... MPCs????

texmex wrote:
I just finished some tests on MDUW MKII. I'm witnessing maximal +79 sample gaps at 48khz sampling which translates to 1,65ms jitter.



Thanx Tex - what Tempo?

regards - David

Rew wrote:
wooooo! sounds tight! :-D thats tight right?

I dont know so much about this stuff and am wondering what these jitter numbers translate to in the real world... is this number for the MD considered REALLY good, good, average,ok ,or sub par? i dont really know. I looked at the original timing thread, but it flummoxed my brain.

Using the same test, what would the jitter be for other gear... like roland drum machines... linn drum... MPCs????

Hi - the original Timing Thread does get a bit over the top but have a look at my Litmus Page to get some comparisons:

http://web.webhost4lif...sp?action=page&name=34

Remember - it's all down to what you like - don't get too hung up on the numbers - everyone hears and appreciates things differently and there is no right or wrong - I set up the Litmus page for close friends really who couldn't get straight answers about timing performance and it turns out many people are interested in the same things.

Regards David.

Wow David, very interesting. I've wondered about how my sh-101 timing was because it's audible in it's loss of sinc from digital performer, Live, emu sp-12, and tr-707. Always I've used the external clock in and never the external gate input. I'm stoked on this bit of info.
The way I look at my 2 drum machines is as drummers themselves. Each has their own feel, timing and style just as a real drummer does (of which I prefer). I often offset hits from the sp-12 (which I guess is a loose machine to beggin with) every few bars by 1/64 just to add a little more looseness like a real drummer - actually since that machine is always double timed that would be 1/128 variation...depending on how you think about it.
My other take on timing looks into the cosmos. Nothing is perfect, not even the sun resolves perfectly around it's own axis! No loop is perfect out there or down here, but don't quote little ol' me on that.
Anyway, cool tests and props to the art of timing! :-D

omriL wrote:
Wow David, very interesting. I've wondered about how my sh-101 timing was because it's audible in it's loss of sinc from digital performer, Live, emu sp-12, and tr-707. Always I've used the external clock in and never the external gate input. I'm stoked on this bit of info.
The way I look at my 2 drum machines is as drummers themselves. Each has their own feel, timing and style just as a real drummer does (of which I prefer). I often offset hits from the sp-12 (which I guess is a loose machine to beggin with) every few bars by 1/64 just to add a little more looseness like a real drummer - actually since that machine is always double timed that would be 1/128 variation...depending on how you think about it.
My other take on timing looks into the cosmos. Nothing is perfect, not even the sun resolves perfectly around it's own axis! No loop is perfect out there or down here, but don't quote little ol' me on that.
Anyway, cool tests and props to the art of timing! :-D

Thanks for the props omriL - yeah the 101 experiment was an eye opener for me in many respects - finally proof of what I had long suspected - that no matter how good your master sync might be and even with the facility to jack analogue voltage pulses in - a sequencer or drum machine - vintage/modern - will only ever be as tight as it's code/hardware allows it to be. Using your Cosmos concept - I see it as the Sync 'Hardware Event Horizon' - you can get close to tight sync but you can't improve beyond the design limitations of the box. All the best - David

innerclock wrote:

texmex wrote:
I just finished some tests on MDUW MKII. I'm witnessing maximal +79 sample gaps at 48khz sampling which translates to 1,65ms jitter.



Thanx Tex - what Tempo?

regards - David

It was the standard 120 BPM.

Need to do a simple layering test later because that's what I'm concerned about. If tracks deviate from each other it's going to be a slight problem because of the phase shifting...

so we can say that the timing is as untight as it was ?

texmex wrote:
I just finished some tests on MDUW MKII. I'm witnessing maximal +79 sample gaps at 48khz sampling which translates to 1,65ms jitter.

Gap measured / diff to 24000
24037 +37
23933 -20
24002 -18
23968 -50
24107 +57
23933 -10
24002 -8
23968 -40



Hey Tex - just had a look at your numbers again - how did you calculate your + and - variances? Are they how far away they fall from the theoretical grid? Some of the numbers don't add up by my sums.....

Also - variation from 'perfect' grid tempo is useful certainly but more significant in terms of what we actually hear as shifts in rhythmic feel is the push-pull differences between consecutive steps that should be even across the whole pattern.

By your first test these end up as:

24037 to 23933 = 104 samples
23933 to 24002 = 69 samples
24002 to 23968 = 34 samples
23968 to 24107 = 139 samples
24107 to 23933 = 174 samples
23933 to 24002 = 69 samples
24002 to 23968 = 34 samples

[Anyone feel free to double check these - it's late and I'm bushed so my maths might be suspect!]

Note the same cyclic nature of the push-pull as measured in the Mark 1.

The maximum 174 sample difference at 48 kHz equates to
3.6ms variation between what 'should be' evenly spaced steps/events.

Regards - David