Happy New year all
Can MD II UW+ do a similar Technique from Merlins OT guide–if so how ?

9.5.3 Abusing a Track for Live Input _
_This sample technique is nasty and not described in the manual as a seperate technique, but it works and is, I.m.h.o. very usable so I present it here. _
_As described earlier, track recorders and tracks are independent. A flex machine running on a track will not bite the track recorder on the same track and vice versa. . . There is, however, a way of setting up a track recorder and a flex machine in such a way that they do bite each other in a useful way. I set it up as follows: _
_Set up a track recorder and let it record the live input. Record length should match the pattern length. _
_Place a full trigger on step 1. By doing so, the recorder keeps on recording, boldly overwriting the last recording every time the pattern reloops. Although the recorder does it’ s job, I cannot hear the live input. _
_Set up a flex machine on the same track. _
_Assign the flex machine to it’ s own record buffer. In other words: if the recorder of track 4 was used, we now put a flex machine on track 4 and assign it to recordbuffer 4. _
_Trigger the flex machine on step 1. Indeed this is the same step on which the recorder was triggered. We have now set up a strange situation: the track recorder is writing to the buffer while the flexmachine is reading from it at the same time. . . So what happens? Well, and this is documented in the manual, in such a situation, you will hear the live inputs. This yields a couple of advantages: _
_Removing the recorder trigger breaks the read/write situation. Result: the flexplayer simply plays the contents of the track recorder. Live input has gone. _


as far as I know setting this situation up on MDUW won’t feed the live input through - unlike OT it appears to me that MDUW uses a double-buffering setup for the RAM machines (or at least it behaves that way), whereas OT is single buffer (i.e. the recording buffer and playback buffer are the exact same memory locations)

edit: each has unique advantages, sometimes I wish I could switch OT to double buffering like MDUW, and vice versa

additional edit: just to be more specific about how the MDUW functions, it won’t let you play back a recording until after it has finished recording; i.e. it doesn’t replace the recording in the associated playback machine buffer until the new recording is complete


I used MD that way, with realtime rec / play. I don’t remember any particular setting, except negative pitch or equal to 0, but it worked pretty well.

Happy new MD!


here is my test I perform just now

  1. record live input 1 bar loop
  2. set playback trig at beginning of bar to play recorded loop, remove record trig
  3. mute live input
  4. put record trig down
  5. record trig is triggered, however we still hear previously recorded loop until the next bar when the playback of the newly recorded silence replaces it

I would describe this as double buffering behavior, the Octatrack is not architected this way. Incidentally, the way the MDUW handles things is identical to the way the Tyme Sefari eurorack module works

edit to add: similarly this is why the playback of the RAM machines on MDUW do not drop out the way OT playback can potentially drop out when you start recording again to a buffer that you’re also accessing for playback


I was using RAM R1 and P1, without removing them. It was realtime because I was playing guitar with no delay. Maybe I missunderstood the OP?

This is also possible with OT. You can play previously recorded part of buffer while recording again, without interruption. I discovered that recently and I was very surprised.


that’s mostly the case, but I would qualify that statement somewhat - it’s more accurate to say the OT is reading/writing simultaneously to/from the same memory locations (i.e. not double buffered). MDUW works fundamentally differently since it is double buffered as I describe - my test shows it is not feeding through live input


Here is my very simple test to show it feeds directly :wink:

Default settings
1 track with RAM R1, 1 trig on step 1
1 track with RAM P1, 1 trig on step 1

You can hear / mangle input in realtime.


If I perform my exact test on OT it silences as soon as it hits the recording trig. I think you’re getting tricked, it simply does not put a new recording into the buffer until it is done with the new recording. It will persist the old recording until that point in all tests and usage I have ever performed. When I first got OT I was expecting it to behave like the MDUW I had borrowed for a while but it didn’t


If you use Src3 and record the track or start recording the inputs but have the record trig microtimed one notch to the right of the play trig, the OT will play the previous buffer as its recording a new one…


that’s playing the part of the buffer it hasn’t wiped yet from new recording, it’s not double buffering

edit: interestingly as soon as the OT starts recording to a recording buffer, if you’re looking at the waveform display it will actually wipe the display but that’s a lie, there is still audio in that chunk of the buffer until it gets overwritten or the recording finishes (assuming the new recording is shorter than the old buffer)


Yeah, that’s what I heard / saw with rec / play cursors.


Alrighty then… :smile:


If I designed OT mk3 it would have option to pick single or double buffer settings :heart_eyes:

edit: oooohh, now I’m imagining a ‘mixed’ mode with advantages of each approach…


In your MD test, is RAM P1 played by a track with lower number than RAM R1 ?
Because behaviour is not the same in that case.

In my test, RAM P1 is played by a track with higher number.


interesting, yes the P1 is lower number track than R1 in my test




re-did my test, very cool! I guess on MDUW you get to pick if you want to wait to replace the buffer in the playback machine until after recording is complete, or send it right through! I always put my RAM records all the way on right so I don’t hit them by accident, but never considered the order of operations on sequencer trigs could affect this so dramatically :sweat_smile:

still a bit odd how this all works out though, and why, because the kind of behavior I am used to with MDUW when having the record machines on higher number tracks you’d need to use 2 recording buffers on OT to replicate (i.e. persist the entirety of the old recording in memory until the new recording is completely finished recording - double buffering)

edit to add: once again the MDUW proves superior in some respects over more modern designs :grin:

edit to add: thinking about it, having the RAM record machine on a higher sequencer track presumably makes those triggers execute a tiny bit later, which makes it basically like the Octatrack microtiming situation that Open_Mike described, but clearly the result is quite different on MDUW vs OT


thanks for all reply
Can you please give me the exact procedure to achieve above
I have a MD UW+ and wish to test on a blank 4 bar pattern i have ready.
My turntable preamp out >>>>left MD input A and right input B


Simple as that, for mono input A+B.

2 more tracks needed for stereo, and you need R1 with IBAL on the left to record input A, R2 with IBAL on the right to record input B.

1 track with RAM R1, 1 trig on step 1, IBAL left
1 track with RAM P1, 1 trig on step 1
1 track with RAM R2, 1 trig on step 1 IBAL right
1 track with RAM P2, 1 trig on step 1


interesting that in the manual the tutorial for the RAM machines has you put the record machine on a lower number track, but otherwise I’m not aware of any mention in the manual about track order mattering for these machines

The concept of how the RAM machines function can
be a bit hard to grasp since they work so differently
from the rest of the machines in the Machinedrum.
Here follows a short tutorial of how to set them up for
recording and playing back the sound from the main
outs. This tutorial assumes you are familiar with the
basic operation of the Machinedrum UserWave.

  1. Make sure that the main out of Machinedrum UserWave
    is connected to your mixer or sound card. You can also
    plug a pair of headphones into the phones output. Compose
    a 1-bar pattern consisting of a bass drum, a snare
    drum and some hi hats. The pattern should not contain any
    other machines.
  2. Assign the first RAM RECORD machine (RAM-R1) to
    track M1.
    This machine will be used to capture the beat.
  3. Assign the first RAM PLAY machine (RAM-P1) to track
    This machine will play back the captured sample of
    the RAM RECORD machine.
  4. Scroll the SOUND SELECTION wheel until track M1,
    the track with the RAM RECORD machine, becomes
  5. Enter the SYNTHESIS page of the RAM RECORD
    machine and make sure that MLEV is set to zero. Turn
    down ILEV to -64. Doing this you will record sound from
    the main outs and nothing from the external inputs. LEN
    should be set to 64 and RATE to 127. This will capture one
    bar with maximum recording quality.
  6. Press the [REC] button and place a trig on the first step
    of the pattern, then press [PLAY]. When the leds
    of the pattern sequencer moves over the RAM RECORD
    trig, the machine will start to record sound. Remove the trig
    right after it has been trigged. This is because we don’t
    want the machine to record a new take. The loop is now
    stored in the RAM RECORD machine. It is time to play it
  7. Scroll the SOUND SELECTION wheel until track M2,
    the track with the RAM PLAY machine, becomes active.
  8. Make sure that the led is still lit and the
    place a trig on step 5 of the pattern sequencer. You should
    now hear the captured loop start playing over the already
    programmed beat, changing the rhythm considerably.
  9. Try experimenting with the loop. Place a trig on step 9 of
    the pattern sequencer and parameter lock the STRT
    parameter to a value of 16. The recorded loop will now
    begin from a different start point further changing the style
    of the beat.