Elektron Musical Instruments

Hi David,

Following your testing method, here's the first 8 quarter notes' spacing data from my hardware sequencers.  I used a single track of hard-quantized uniform velocity and gated rim-shot samples.  Do my thoughts/comments make sense to you?

Elektron MD SPS-1 UW OS v1.33 
22112 22017 22048 22113 22016 22049 22016 22113
max. differential: 95 samples
comments: somewhat uniform long/short variation pattern with a period of roughly every 1/2-2/3 measure

Korg EMX-1 latest OS
22001 22000 22100 21998 22100 22000 22100 22000
max. differential: 102 samples
comments: after the first two notes, the variation becomes uniform (i.e tight relative to its own timebase)

RS7000 latest OS
22051 22047 22050 22114 21984 22047 22050 22120
max. differential: 130 samples
comments: has a different short/long pattern with a period of roughly 1 measure

Roland MC-80 EX OS version unknown 
21998 22060 22059 22070 22047 22060 22065 22069
max. differential: 62 samples
comments: after the first two notes, the variation minimizes resolves to a similar pattern as the RS7000, but with a much smaller variation magnitude of 23 samples (i.e tighter)

Conclusion: the MC-80 EX is the tightest of my hardware sequencers.  So, does anyone want to buy it?  ;-)

It's ironic that the most fun and interactive of these sequencing devices (MD and EMX) have the most clock variation from ideal, whereas the most structured and rigid box (MC-80 EX) is the "tightest".  Coincidence?

Finally - have you done any measurements of software VSTi hosts playing the quarter-note pattern in one computer, and using another computer to record the host's audio output?  Of course there are many more variables in SW, such as audio interface latency, the VSTi plug-in itself, the host, the OS, etc... but that's SW for you.  It would be interesting to see how SW stacks up against HW...

I have a feeling you might experience different results with rendered audio tracks within the host as opposed to recording the audio output on another computer/device.  My guess is that rendering would yield a more accurate output than a recording of the audio output.  Do you agree?

Thanks,
Scott
 


  ----- Original Message ----- 
  From: innerclock2004 
  To: elektron-users@yahoogroups.com 
  Sent: Wednesday, April 04, 2007 3:42 AM
  Subject: [elektron] MD SPS-1 Timing Performance Issues


  Hi all - there is an interesting topic/thread regarding timing 
  stability of the MD SPS-1 at the Elektron Users Forum if anyone feels 
  like taking a look while the thread and interest is active.

  http://www.elektron-users.com/modules/newbb/viewtopic.php?
  topic_id=2155&viewmode=flat&order=ASC&start=0

  All the best - David
  www.innerclocksystems.com



   

[Non-text portions of this message have been removed]