Jomox ALPHA BASE

Any music gear discussions that don't fit into one of the other forums.
Post Reply
User avatar
dubonaire
Super Deluxe Wiggler
Posts: 10622
Joined: Sun Jan 17, 2010 10:45 pm

Re: Jomox ALPHA BASE

Post by dubonaire »

stice wrote: Tue Jan 24, 2023 11:01 am Regarding ‘groove’ I’m entirely in the fence: it’s been proven pretty conclusively that clock/timing differences don’t explain itZ. Yet subjectively I can’t deny that some machines feel a lot better than others, without even trying. Is it envelope shapes? Sample selection? Distortion in the amps or mixers?

It’s a curious situation for sure.
Well the clock is only one element of a drum machine.
User avatar
stickman
Ultra Wiggler
Posts: 774
Joined: Mon Jan 08, 2018 7:45 am

Re: Jomox ALPHA BASE

Post by stickman »

dubonaire wrote: Tue Jan 24, 2023 3:43 pm
efluon wrote: Tue Jan 24, 2023 8:42 am
BRE wrote: Tue Jan 24, 2023 7:44 am The only groove any drum machine has is the one that's been created. There are no drum machines with a special groove. The math is the same when it comes to time division- Roger Linn ended that debate quite a while ago.
It feels like it is still ongoing though? Machines are mostly deterministic, but the results may still vary if there are degrees of freedom inside the algorithms.

E.g. there is something nice about just the right amount of jitter. Especially when it is not the same for all voices. (I really find bad european samba bands more groovy than tight brasilian ones). There are other possible sources of groovy deviations from the precise, and as the Jomox software has other issues, it might include groovy accidents?

I've known people who actually sampled the timing from real drum machines and even from Logic on Atari ST to create groove templates. The R-8 had a feature to apply grooves to whatever patterns you programmed.
effluon many of the machines have been tested by innerclock Systems https://www.innerclocksystems.com/litmus
That's v interesting data..

Anecdotally, folks consider elektron machines commonly to sound "rigid" yet machines like Behringer RD seemingly have lower audio jitter.

Might it be safe to dismiss jitter as a variable here and focus more on how the sound is generated?
User avatar
dubonaire
Super Deluxe Wiggler
Posts: 10622
Joined: Sun Jan 17, 2010 10:45 pm

Re: Jomox ALPHA BASE

Post by dubonaire »

stickman wrote: Tue Jan 24, 2023 4:09 pm
dubonaire wrote: Tue Jan 24, 2023 3:43 pm
efluon wrote: Tue Jan 24, 2023 8:42 am
BRE wrote: Tue Jan 24, 2023 7:44 am The only groove any drum machine has is the one that's been created. There are no drum machines with a special groove. The math is the same when it comes to time division- Roger Linn ended that debate quite a while ago.
It feels like it is still ongoing though? Machines are mostly deterministic, but the results may still vary if there are degrees of freedom inside the algorithms.

E.g. there is something nice about just the right amount of jitter. Especially when it is not the same for all voices. (I really find bad european samba bands more groovy than tight brasilian ones). There are other possible sources of groovy deviations from the precise, and as the Jomox software has other issues, it might include groovy accidents?

I've known people who actually sampled the timing from real drum machines and even from Logic on Atari ST to create groove templates. The R-8 had a feature to apply grooves to whatever patterns you programmed.
effluon many of the machines have been tested by innerclock Systems https://www.innerclocksystems.com/litmus
That's v interesting data..

Anecdotally, folks consider elektron machines commonly to sound "rigid" yet machines like Behringer RD seemingly have lower audio jitter.

Might it be safe to dismiss jitter as a variable here and focus more on how the sound is generated?
A drum machine's groove is not due to jitter. This is a point Roger Linn has made clear, and to quote that website "jitter is not human feel".
User avatar
efluon
Wiggling with Experience
Posts: 367
Joined: Wed Mar 29, 2017 8:19 am
Location: There is no planet Earth

Re: Jomox ALPHA BASE

Post by efluon »

stickman wrote: Tue Jan 24, 2023 4:09 pm
dubonaire wrote: Tue Jan 24, 2023 3:43 pm ...
effluon many of the machines have been tested by innerclock Systems https://www.innerclocksystems.com/litmus
That's v interesting data..

Anecdotally, folks consider elektron machines commonly to sound "rigid" yet machines like Behringer RD seemingly have lower audio jitter.

Might it be safe to dismiss jitter as a variable here and focus more on how the sound is generated?
I would also add that it is not safe to trust this interesting data, that unfortunately comes without a description of what is measured.

Elektron machines do blockbased processing (the source of most jitter) but of internally generated (sequenced) data. There is not even analog envelopes in e.g. the Digitakt. So no source of jitter unless it is externally sequenced. Yet it is listed as having a jitter of 64 samples.
I will measure this myself, by just recording a plain single drum quarters beat.

And I agree that jitter most certainly is not human feel. (I once wrote a humanize function and explicitly modified a gaussian randomization with some easy rules. That works quite well.)

But while I always loved the sound of the Jomox machines I never liked their user interface, so I do not own an AB and can not actually comment on its groove.
User avatar
dubonaire
Super Deluxe Wiggler
Posts: 10622
Joined: Sun Jan 17, 2010 10:45 pm

Re: Jomox ALPHA BASE

Post by dubonaire »

efluon wrote: Wed Jan 25, 2023 3:05 am
stickman wrote: Tue Jan 24, 2023 4:09 pm
dubonaire wrote: Tue Jan 24, 2023 3:43 pm ...
effluon many of the machines have been tested by innerclock Systems https://www.innerclocksystems.com/litmus
That's v interesting data..

Anecdotally, folks consider elektron machines commonly to sound "rigid" yet machines like Behringer RD seemingly have lower audio jitter.

Might it be safe to dismiss jitter as a variable here and focus more on how the sound is generated?
I would also add that it is not safe to trust this interesting data, that unfortunately comes without a description of what is measured.
It is completely safe to trust that data, which does come with a clear description of the inputs and measured outputs and the methodology. Innerclock Systems knows what it is doing. The only variable would be differences within individual instruments across a product line.

How is this not a description of how and what was measured?
At a tempo of 120 BPM, continuous 16th events should occur exactly 12000 samples apart from each other when captured at 96kHz.

Jitter values as reported in the tests below are the maximum random timing error measured between consecutive sixteenth step events over a continuous sixty second capture period.

Audio Out Jitter >
96 samples / 1.00ms

MIDI TX Jitter >
32 samples / 0.33ms
User avatar
efluon
Wiggling with Experience
Posts: 367
Joined: Wed Mar 29, 2017 8:19 am
Location: There is no planet Earth

Re: Jomox ALPHA BASE

Post by efluon »

dubonaire wrote: Wed Jan 25, 2023 3:16 am How is this not a description of how and what was measured?
It misses the source of the events (i.e. internal vs external). Also missing is the actual measurement (i.e. after recording, how are the on-beats counted), but as it describes using step pulses that probably is rather safe.

I'll be in the studio this evening and will make some recordings with the internal sequencers on Digitakt and TR-8, which I'll later measure manually in the daw and report very briefly here. Unless that is not wanted.
User avatar
dubonaire
Super Deluxe Wiggler
Posts: 10622
Joined: Sun Jan 17, 2010 10:45 pm

Re: Jomox ALPHA BASE

Post by dubonaire »

efluon wrote: Wed Jan 25, 2023 3:41 am
dubonaire wrote: Wed Jan 25, 2023 3:16 am How is this not a description of how and what was measured?
It misses the source of the events (i.e. internal vs external). Also missing is the actual measurement (i.e. after recording, how are the on-beats counted), but as it describes using step pulses that probably is rather safe.

I'll be in the studio this evening and will make some recordings with the internal sequencers on Digitakt and TR-8, which I'll later measure manually in the daw and report very briefly here. Unless that is not wanted.
No it doesn't. It indicates whether the source is sync master, sync slave (what you are calling internal and external) (inc ppq and mode), or voltage trigger, and if you look closely and understand what you are looking at you'll see certain patterns emerging. To be honest I don't think you understand what you are reading.
User avatar
revtor
Dialing it in..
Posts: 2716
Joined: Sat Nov 29, 2008 10:31 am
Location: North Jerz

Re: Jomox ALPHA BASE

Post by revtor »

efluon wrote: Wed Jan 25, 2023 3:41 am
dubonaire wrote: Wed Jan 25, 2023 3:16 am How is this not a description of how and what was measured?
It misses the source of the events (i.e. internal vs external). Also missing is the actual measurement (i.e. after recording, how are the on-beats counted), but as it describes using step pulses that probably is rather safe.

I'll be in the studio this evening and will make some recordings with the internal sequencers on Digitakt and TR-8, which I'll later measure manually in the daw and report very briefly here. Unless that is not wanted.

Please do! Data like this is always welcomed…
Fellow North Jersey Synthaholic
ObsoleteModular
Ultra Wiggler
Posts: 903
Joined: Tue May 17, 2016 4:24 am
Location: Bight Ron

Re: Jomox ALPHA BASE

Post by ObsoleteModular »

BRE wrote: Tue Jan 24, 2023 7:44 am The only groove any drum machine has is the one that's been created. There are no drum machines with a special groove. The math is the same when it comes to time division- Roger Linn ended that debate quite a while ago.
You're wrong.

The math is theory. The reality is electronic engineering for analog systems and programming for digital systems. How these are approached and their effectiveness differ wildy.

A hardware interrupt in a microprocessor versus a timer implemented in code. Serial versus parallel triggering of voices. There's many examples.

All can contribute to a different feel for the same pattern programmed on separate machines.

Hell, we even had two 808s in the old studio that weren't the same. Program the same beat on each and they didn't feel the same at all. One was always wonky. It was probably bust but that kind of makes the point.
User avatar
m0rb1d
Filtersweep Abuser
Posts: 1179
Joined: Sun Jul 18, 2010 1:30 pm

Re: Jomox ALPHA BASE

Post by m0rb1d »

I do agree with BREs point about what it does to samples. The analog envelopes have an extreme effect on the sound of the samples and can almost sound like a crazy compressor with the right settings. Another big thing is how much the individual voice volume setting has on the sound of the selected voice.
User avatar
stickman
Ultra Wiggler
Posts: 774
Joined: Mon Jan 08, 2018 7:45 am

Re: Jomox ALPHA BASE

Post by stickman »

dubonaire wrote: Tue Jan 24, 2023 5:39 pm A drum machine's groove is not due to jitter. This is a point Roger Linn has made clear, and to quote that website "jitter is not human feel".
Also, a cool interview with Roger Linn on the subject here (didn't see it elsewhere in the thread):

https://www.attackmagazine.com/features ... pc-timing/
User avatar
efluon
Wiggling with Experience
Posts: 367
Joined: Wed Mar 29, 2017 8:19 am
Location: There is no planet Earth

Re: drummachine jitter

Post by efluon »

So, I made some recordings. This was all internally sequenced 16th. Once with sync to external midi (E-RM Multiclock), once internal midi, both at 120 bpm. I used short hihats. Recorded at 44.1kHz (I like my old converters). I measured distances from waveform onset visually (i.e. before first non-zero sample) in Ocenaudio. Also, I did not write a program to measure but did only look at random samples, i.e. can't give you a standard error.

Swing was adjusted to 50% on Digitakt. Nonetheless there are two bins into which all distances between onsets can be sorted. Same result for external and internal sync source. That is 5497 samples alternates with 5526 samples. No variance at all. While this could be interpreted as jitter and scales to about 64 samples at 96kHz (the number given at that strange website) it is not random, and after rereading the Digitakt manual I find the description unclear ("Turn DATA ENTRY knob E to set the SWING ratio to 51-80%. The default setting is equal spacing, 50%."), so my interpretation is that the minimum is not equal spacing but 51%.. I will try to get clarification for that.

On TR-8 swing can only be adjusted by potentiometers which i set to their center detent. So I am not that surprised that the results fall into two bins again. Again, no difference within the statistical significance of my samples between internal and external sync. One bin varies between about 5550 and 5558, the other bin between 5468 and 5471. This may be in there on purpose, I can't say.

The reason I made these measurements was that I have the impression that the Digitakt is extremely tight, and tighter than the TR-8. The data on that site did surprise me. My own measurements, while sporadic, hint at my impression being correct. My takeaway is to not believe the absolute values of the swing settings of both devices.
blech
Learning to Wiggle
Posts: 50
Joined: Fri Jul 22, 2016 3:48 pm

Re: drummachine jitter

Post by blech »

efluon wrote: Thu Jan 26, 2023 7:11 am ...after rereading the Digitakt manual I find the description unclear ("Turn DATA ENTRY knob E to set the SWING ratio to 51-80%. The default setting is equal spacing, 50%."), so my interpretation is that the minimum is not equal spacing but 51%.. I will try to get clarification for that.
Not sure what you mean here. In the Elektron sequencer, 50% = no swing, and the range of possible swing (above none) is 51% to 80%. So, the minimum (default) setting for the swing parameter is 50% (equal spacing), but the minimum setting for applied swing is 51% (first value resulting in unequal spacing), if that makes sense.
User avatar
dubonaire
Super Deluxe Wiggler
Posts: 10622
Joined: Sun Jan 17, 2010 10:45 pm

Re: drummachine jitter

Post by dubonaire »

efluon wrote: Thu Jan 26, 2023 7:11 am So, I made some recordings. This was all internally sequenced 16th. Once with sync to external midi (E-RM Multiclock), once internal midi, both at 120 bpm. I used short hihats. Recorded at 44.1kHz (I like my old converters). I measured distances from waveform onset visually (i.e. before first non-zero sample) in Ocenaudio. Also, I did not write a program to measure but did only look at random samples, i.e. can't give you a standard error.

Swing was adjusted to 50% on Digitakt. Nonetheless there are two bins into which all distances between onsets can be sorted. Same result for external and internal sync source. That is 5497 samples alternates with 5526 samples. No variance at all. While this could be interpreted as jitter and scales to about 64 samples at 96kHz (the number given at that strange website) it is not random, and after rereading the Digitakt manual I find the description unclear ("Turn DATA ENTRY knob E to set the SWING ratio to 51-80%. The default setting is equal spacing, 50%."), so my interpretation is that the minimum is not equal spacing but 51%.. I will try to get clarification for that.

On TR-8 swing can only be adjusted by potentiometers which i set to their center detent. So I am not that surprised that the results fall into two bins again. Again, no difference within the statistical significance of my samples between internal and external sync. One bin varies between about 5550 and 5558, the other bin between 5468 and 5471. This may be in there on purpose, I can't say.

The reason I made these measurements was that I have the impression that the Digitakt is extremely tight, and tighter than the TR-8. The data on that site did surprise me. My own measurements, while sporadic, hint at my impression being correct. My takeaway is to not believe the absolute values of the swing settings of both devices.
Did you notice the TR-8s result on the Innerclock site is the tight one? The TR-8 result is not so tight.
User avatar
efluon
Wiggling with Experience
Posts: 367
Joined: Wed Mar 29, 2017 8:19 am
Location: There is no planet Earth

Re: jitter or swing?

Post by efluon »

blech wrote: Thu Jan 26, 2023 4:00 pm So, the minimum (default) setting for the swing parameter is 50% (equal spacing), but the minimum setting for applied swing is 51% (first value resulting in unequal spacing), if that makes sense.
It is what i would have read, but why even insert that 51 into the text instead of giving a parameter range from 50 to 80? The measured swing is actually less than 51% but it is more than 50%, so I was looking for a reason there.
dubonaire wrote: Thu Jan 26, 2023 4:48 pm Did you notice the TR-8s result on the Innerclock site is the tight one? The TR-8 result is not so tight.
I only noticed that there was significant jitter given for both machines as well as the digitakt, and I have a TR-8 at the Studio. None of these machines have any reason for Jitter when being internally triggered and internally synced (clocked), as internally the offset of the events regarding block borders is known (which it is not for external triggers which always will be played only in the next block). I might look at the TR-8s when I find the time.
Post Reply

Return to “General Gear”