Timing Issues?

Hey guys,

new ER-101 user here and fired it up last night for a test run. I noticed that when I was sending it a clock from the 4ms QCD to send it clock (getting this from the Doepfer Darktime, which in turn is getting clock from my DAW via MIDI). Seems like there is a lag and also seems to drift - all other clocked modules work fine when getting this clock, most likely something I’m doing.

I just following the quick start from the manual to setup a part (I’ve tried using only 1 step and have gone up to 8 steps). I’ve tried Duration and Step settings at 1 and 2 and I think 4 to no avail. Any recommendations on what could be causing this? Thanks!

Ypsi Kid

Are you sending a reset signal too?

:slight_smile:

No reset signal being sent. I ended up sending a reset to keep in time with the sequence I was playing around with, but when I was initially playing with it, I was feeding it just a straight clock from 4ms QCD.

Ypsi Kid

I would not expect a clock to be in sync without a reset, it might be, but not necessarily, as far as I know a reset pulse is pretty much essential.

The plus side of this is that you get a lot more functionality and playability because you can use your reset as a musical device like anything else :slight_smile:

I don’t want to say this is definitely it, I guess there could be some other problem, but I’d try setting this up as a first before looking for anything else!

Thanks Kel, I’ll give this a try tonight when I get home and update. Appreciate the quick reply!

Ypsi Kid

I don’t mind admitting I used to do the same thing and not bother, but then got into resets and well… I imagine you will probably be like me, and probably most other folks too, and never look back once you have it up and running :wink:

If it doesn’t resolve the timing issue come back and we can dig further, plenty of folks round here with loads of experience and knowledge and I am sure together we will get to the bottom of it because it should work perfectly.

Hi Kel,

I’m confused - why would you need a reset? The clock input should be constantly re-calibrating itself based on the incoming signal. Either like a tape tempo, averaging the time between in the gates, or just stepping forward when a new gate arrives.

If you need to use a signal for both clock and reset, are you really just using 2 clocks? And then you also have to make sure that the reset reaches before the clock…

Oh I see, yeah, it’s not that simple, there are lots of things to consider, perhaps try to think of it in relation to your sequence and where the play head is in your sequencer, bearing in mind that it is not likely to be a one clock pulse per step relation unless you have specifically set it up that way*.

It’s quite hard to explain and it’s really late, so here’s a little thought exercise that might help everything make sense; if you are playing an 8 note sequence over 4 bars at 24ppq and you change the clock division or multiplication or stop and restart your clock, where is the beginning of your sequence in relation to your master clock?

And no, I don’t think it’s helpful in this instance to start thinking of the reset as another clock, although you’re right conceptually and in practice it could be, I’d consider that a more advanced topic to explore after you have fully understood why you need the two signals to make everything work correctly.

I’ll check in tomorrow when my brain might be working a bit better and be able to explain it in a better way.

  • I wouldn’t particularly recommend a one clock pulse per step setup, ideally you should have 24ppq if you are wanting to sync to midi.

Good morning. Clarification please!

  • The clock path is DAW > Darktime > QCD > ER-101?
  • When you hear lag/drift from the ER-101, what are you using as a comparison?

I may have solved it. Think it was a combination of things:

  1. Didn’t have the sequence set up right (I now have it set to 1 step, 1 gate, duration 4)
  2. Midi timing was slightly off, adjusting via track delay in the DAW
  3. Now sending a reset directly from the Darktime to the ER-101

The clock path is DAW > Darktime > QCD (Also was trying direct from the Darktime). I was using a drum loop I made to compare timing, think I’m good now - at least the sequence I have running now is synched nicely.

Appreciate all the quick responses and sorry for such a simple question!

Ypsi Kid

I should state that there is still some weird behavior. I used a manual gate to reset the ER-101 whenever I stopped the sequencer (I also ‘reset’ my DAW to begin on the ‘1’ of bar). Sometimes I’d see the ER-101 falling in between quarter notes (so hitting where a HH should hit as opposed to a kick in a 4x4 pattern). Once I reset again, seems to fall in line, but every once in a while I do see this delay - no idea where this could be coming from?

Ypsi Kid

So at least you now get why the reset is required? :slight_smile:

I don’t know what the problem could be after that,I’d be interested to hear what your clock settings are, i.e. what is the DarkTime doing with the midi?

p.s. never any need for apologies for asking questions here, we’re all very understanding folks and have all been there at some point, well, actually even after three decades of messing about with this stuff I still get things wrong and ask simple questions :joy_cat:

Sorry to hijack this thread, but I still don’t see why you’d need a reset signal.

I mean I understand if you have some sort of master clock which is also controlling other sequences and you stop in the middle of a sequence, yes you’ll then need to reset then to get everything back to the 1. But that’s not what I’m talking about, and that’s not what I think you’re talking about, so in a normal clocking situation with the ER-101, 1 incoming gate signal is equal to 1 step of the sequencer, right? So if my sequence is 4 steps long, after that 4th gate, I’ll be back on the 1. I don’t see where the drift would come from or why a reset would help fix that.

Any other drift which is coming from dividing or multiplying 24 or however many PPQ to a clock pulse that’s then feeding the ER-101, the drift is coming from whatever the source of that conversion.

The only reason I see for bad drift other than this is if the clock source into the ER-101 is not consistent, and even then, that should happen on a sequencer that takes an incoming clock signal and averages its timing against the previous timings of the clock. If a sequencer just goes to the next step, simply, after each clock pulse, there shouldn’t be any drift at all, as long as the signal is reaching it correctly.

I can see how a reset might be helpful as a fail-safe, an extra measure taken to ensure the best possibility of 2+ things being synced, but at that point that’s really just using an extra clock. And, in most circumstances all I can see a reset doing is actually complicating the situation, by further shifting the sequence if the clock and the reset signals don’t arrive in a specific order relative to each other.

So I’m curious, how does the reset come into play for you? I’ve never heard anyone mention this technique before.

Let’s establish some terminology first so that we don’t confuse each other. There are 2 kinds of synchronization loss that I know of:

  • Clock drift: Two clocks gradually go out of sync because their periods are not exactly the same.
  • Counting: A sequencer is not counting clock pulses correctly, or it was not properly sync’ed when started (i.e. it didn’t start counting from 1).

In general, a clock reset is necessary when you want to sync one clock with another clock because there is no such thing as 2 independent clocks with exactly the same tempo (i.e. all clocks drift). You also need a sequencer reset when you only have one clock driving a bunch of sequencers and you need to get them to all start together. Loss of synchronization is always with respect to something, and we need to know that something to make a diagnostic.

The ER-101 indeed acts like an analog sequencer in that it needs clock pulses to advance through a sequence. However, if you have activated a track’s clock multiplier (by setting the multiplication factor to something other than 1) than the ER-101 will generate additional virtual clock pulses based on the observed incoming clock period. Here is an excerpt from the ER-101 manual (p.29) explaining the procedure.

Clock Multiplication

When the clock multiplier is set to a value greater than one, then the last seen period from the external clock is used to generate a faster clock. For example, suppose track 1’s multiplier is set to 8 in the track options screen. The following will occur:

  1. First the ER-101 waits until the next pulse on the external clock.
  2. When this pulse arrives, output a pulse to track 1 and note the period. Let’s call this period T.
  3. Wait T/8 seconds.
  4. Output another pulse to track 1.
  5. Repeat step 3 and 4 until we have outputted 8 pulses including the pulse from step 2.
  6. Go back to step 1.

(Throughout the above procedure, whenever an external pulse arrives, abort and start again from step 2.)

The above procedure generates 8 pulses for every single pulse from the external clock, thus running track 1 at 8 times the speed of the external clock. In other words, a step in track 1 with a DURATION=8 will last one
cycle of the external clock. In addition, the above procedure guarantees two beneficial properties:

  • The synthesized clock is always in phase with the external clock.
  • The synthesized clock period tracks the external clock period but with a lag of one full period.

Caveat for synchronizing with various clock multipliers

If you change the clock multiplier or quickly change the external clock frequency while the sequencer is playing then a multiplied track will go out of sync with any other tracks with different clock multipliers. In these cases, just send the ER-101 a RESET to get all the tracks in sync again. This happens NOT because the multiplied clocks are out of phase with the external clock (they are guaranteed to be in sync) but rather because you changed the frequency of the clock in the middle of a step. This is typical (and probably unavoidable) behavior for patches that involve multiple clocks.

Assuming that the post-multiplied clock is less than 3kHz, the caveat at the end is the only way that I know that the ER-101 can and should “lose count”.

Yes, assuming all of your steps have DURATION set to one.

There are cases where a reset is not necessary. For example, you have two sequencers running from the same clock and you have arranged for them to start together, than after that they will not need a reset to stay in sync. The ER-101 has this property.

1 Like

It is hard to diagnose without knowing how the Dark Time generates a clock from your DAW’s MIDI, especially around a start/stop event. For example, is the Dark Time dividing down the MIDI clock? If so, does the Dark Time restart the clock divider’s internal counter when you start/stop on your DAW? I would think it does but it needs confirming. What about the QCD? Are you resetting it also, for the same reason?

A clock divider is a (very simple) type of sequencer. It has the same synchronization needs as a full-blown sequencer.

1 Like

Morning - I see @odevices beat me to it and probably did a far better job of explaining than I would have too!

An epic tale of reset woe is the MakeNoise Tempi, check out the MW thread :wink:

Re. the Dark Time, I had a look at the manual and the information about the clock divider setup is on p19:

You may use Dark Time as a simple MIDI-Clock-to-Sync interface. When running as a slave, synced to an external MIDI-clock (master), Dark Time will send out a clock signal from its Clk-Out socket according to its clock divider setting (please refer to section 3.1.4, page 13)

So it really depends how you have your clock division setup.

The good news is that from what I have read (I’ve never used a Dark Time) it seems as though it has been very well implemented and it also has a reset output so should work as you wish with all the right settings.

Okay, that makes sense - it’s when one clock changes in relation to the other clock, and that other clock is also doing some sort of division or multiplication, then a reset signal will keep everything synced up.

I’ll check out the Tempi thread - not super familiar with that module - but I can see how it gets complicated when you have a bunch of derivative clocks all generated from one or more other clocks, especially if they’re syncopated or poly-rhythmic.

I haven’t spent too much time playing with the clock division/multiplication features of the ER101. I’ll need to give that a try.

Thanks again @anon83620728 and @odevices - some great info in there and helps me understand better why I’m seeing the behaviour I’m seeing. Sending a reset from the Dark Time synchs everything up nicely!

Cheers and thanks again for all the knowledge, you guys are awesome!!

MM

Cool! :smiley_cat:

Another good read is the Pam’s New Workout thread on MW, lots of discussion about timing, sync, clock division, ppqn etc… there!

1 Like