OS Updates! Rytm MKII 1.40C, Digitakt 1.07, Digitone 1.02


I’ve sent the Digitakt 1.07 file to my Digitakt and it just says ‘receiving’ for the last 10 mins. I am guessing it’s not actually updating. Can I reboot the Digitakt and try again without risk?


i think, you can abort with “NO”. it seems you are not sending anything. be sure you have set your midi ports right. in the DT you have to enable usb at the inport. with which application are you updating?


Thanks, worked instantly the second time.


Don’t get too excited, it still doesn’t work for non Elektron boxes. They need to implement instant pattern change


What would that entail? You want a pattern change on the next 16th step after changing instead of waiting for the next bar?


No, it should change pattern instantly when recieving a PC. Most hardware sequencers will send a PC before the first note of a pattern is even played, and they expect the recieving unit to respond immediately, not wait until 1, 2 or 4 bars have finished playing. This way you can use the midi sequencer to switch patterns on your entire rig and program complete songs.


so you are saying most hardware sequencers are waiting before the first note if there will a pc change coming in?
i don’t think so…
if they send a PC before the first note of a pattern, the first note will lag behind because it is not on 0 0 0 0.
i never had a problem with sending a pc change BEFORE the next pattern, so it will change right on time…
i don’t see your problem.


That is correct, MIDI messages happen very fast and PCM is sent/received before MIDI Note data. The receiving unit should respond to the PC change BEFORE the first note is even played. This is the case with patch changes too. If I send the SH-01A a patch change at the beginning of a pattern it will change the patch instantly AND before the first note on step 1 is played.


Good stuff, innit?

My AK/OT/DN are happily synchronized.


ok if you don’t think its necessary that the first note message should be on 0 0 0 0.
when do you expect the second note message in a quarter pattern on 0 1 0 0 or on 0 1 0 48?

in short it will happen in the last pattern - everything else would be a big mess - midi is NOT so fast that it can react immediately cause its serial.
there must be a musical measure for a pattern to change. be it a 1/16 or a 1/1 or a 4/1, otherwise everything would drift apart.
even in the A4 and AR where you can change a pattern “immediately” its not immediately it happens on the next step.


Right, it will never be immediate. The sequencer needs to schedule Trigs and P-locks in advance of the next step. I think his main complaint is that it doesn’t happen on the next possible scheduling cycle (the next 16th step?).


as long as i set the change length of my DT and DN to the same value the pattern changes are always in sync.
so when i change the pattern somewhere between the 1 and the last step of a pattern. the pattern changes after finishing

i know the behaviour of the a4, where you could change a pattern ig after the 5 th step and the next pattern continuous at the 6th step - of course thats great, but i have really no problem like it is on the dt
but this pc changes at some esoteric minus time, i don’t get…


MIDI messages don’t all come through on 16th steps, it’s much faster than that. There is plenty of time for a synth to react to a PC before the first step.


wow such fast? … but still there is nothing before 0 :wink:


It’s a sequencer though…things need to be in time. If you allowed pattern changes at some unquantized time, things would be a mess. You could make an argument to change on the next 16th or 32nd (whatever multiplier you have on the sequencer, etc), but in absolute realtime doesn’t make sense for a sequencer. Besides that, you can P-Lock a hundred parameters to a single step, have those slide, and nudge that step forward / backward. The sequencer has to schedule all of that data ahead of time.

The smallest time resolution would be the minimum time you could get two trigs next to each other on the same track (using micro-timing), at the fastest BPM.


I don’t really want to have to explain how MIDI messages are sent so I’m gonna stay out of this thread now. I’ve said what I need to, hopefully Elektron will add instant PC in an update.


Over a DIN MIDI connection, a Program Change message takes 1ms in the most ideal of circumstances.

Let’s assume we have two sequencer A and B running perfectly synchronised.

Let’s also assume sequencers B reacts to Program Change messages by immediately starting playback of the pattern indicated by the program number, which is the behaviour you’re advocating if I’m not mistaken.

If sequencers A now sends a Program Change message to sequencer B, and sequencer B starts playback of the pattern indicated by the program number, sequencer B will now be at least 1ms behind sequencer A.

This assumes the most ideal of circumstances. In practice, sequencer B will more likely be somewhere in the 5 to 10ms range behind sequencer A.

At 120BPM, a single beat (one quarter note) will have a duration of 500ms. A sixteenth note (or one step on a 16-step sequencer) will have a duration of 125ms.

It’s not easy to think of circumstances in which you’re going to notice this as audible glitches, especially on a fully digital box that does sample playback or where the FM algorithm might change.

If I were to manufacture such a box, I wouldn’t implement this kind of behaviour simply to avoid customers complaining about these glitches and insisting I’d fix that “bug”. :slight_smile:


Absolutely! This was the only real show stopper for me. I’m happy to have my Octatrack, Digitakt, and Digitone all follow pattern changes. Hopefully I’ll have a lot of output now over the next years!


again, on the analog 4 instant pattern change means that the pattern changes anywhere within the beat, in the moment it receives the Pattern Change Command, while continuing from the step it was (eg. not restarting from step 1) . This way two sequencers wont go out of timing.


True, but that means that the first step will be from the previous pattern.