how many ms back and forth is the jitter in your system with your plugin?
Im gettIng zero on Op-XY and .1 both directions showing on Oxi MKII. Analog Rytm occasional .1
Thats all 3 running simultaneously
Struggling with TR1000 but not sure if user error but using its own sync delay instead. Someone above says theirs works so guess its me
1 Like
that sounds great!! i missed your comment 
I guess peoples computer setups are diff so worth a test with trial before pulling the trigger. Personally I find this easier then playing with daw latency offsets
1 Like
I`m normaly using my âmulticlockâ and testing now the omniclock. So far so good, but if you using inside Ableton PlugIns producing latency then you have to adjust the clock again. Same with âmulticlockâ! Further I canÂŽt use every Synth showing up in the menu of Omniclock. For examble Moog Sub37, PRO1, Novation PEAK and some more, but I use it for Keystep and OXI MKII, perfect. Free space at the multiclock. Do not understanding the issue, maybe because I use âmioXLâ!
PlugIn compansation will be great?
Yeah the compensation is huge. The Nome2 can run with some plugins in the session and still nail the timing so Iâll stick with that for now. Itâs still important with any clock solution to run a very light plugin setup on a session.
2 Likes
What iÂŽm not understand is, why Ableton stuff not able to bring a solution? Only 3third Partys smart to do it!
1 Like
iâm getting significant wobble my end - I can produce real data if anyone interested
1 Like
Because achieving reliable, jitter-free, multi-port shiftable precision tempo-sync in a hybrid system isnât as simple as it would seemâŠ
I donât want to hijack the thread and weâre not fishing for sales at all - very different price point obviously, but the idea being pushed around â that you can get precision, hardware-grade clock timing with zero hardware and at a tenth of the price of some options, purely in software â is the bit worth testing, because thatâs the whole premise. We released our Sync-Lock hardware system around 2009, and since then others like E-RM, Floatingpoint, Nome II and ACME have offered similar excellent products with different features and specs, all with the core aim of delivering stable tempo-clock to hardware in a DAW environment. Notably, none of them have ever offered a software-only equivalent. This is the stuff I do all day, so rather than argue it, here are actual numbers.
Plugin â Mac Studio M4 Max â Antelope Orion 32 Gen 3 Audio Interface â Ableton 12.4.1 â MIOxl â 5-pin DIN out, 100 BPM, 24 PPQ clock, measured right at the TX pin â 80-bar continuous capture into Saleae Logic Pro 16 analyser @ 500 MS/s:
- Jitter Max: +600 / â681 ”s
- Push-Pull Max: 1.28 ms
- Tempo deviation, tick-to-tick: +2.78 / â2.36 BPM (instantaneous 97.7â102.8 BPM)
- Ticks within ±5 ”s: 0%
- Long-term average tempo: spot on, no drift
Structured ±350 ”s early/late wobble on every tick, quantised to the DIN byte period.
To sixhandâs question â why canât Ableton (or any DAW) just sort this itself â hereâs the honest answer, and itâs not really the pluginâs fault either. Itâs the limitations of the architecture and delivery method. A DAW generates MIDI Clock accurately against its audio timeline; âsample-accurateâ is true at that point. But the moment the clock event leaves the buffer it crosses the OS MIDI stack, USB framing, and finally the 5-pin DIN line in the MIDI interface. At that point the DAW and plugin have no control over timing delivery once itâs handed over. A third-party plugin like this can give you a handy per-port manual offset to nudge a device into âthe pocketâ â but remember the base clock jitter is what really defines âthe pocketâ.
Hereâs the part that matters for the shuffle/groove side specifically. Applied clock âgrooveâ is itself a set of deliberate, precise timing offsets. For these to really matter, your base clock stream has to be very stable â jitter spikes down in the double-digit-microsecond range ideally. The further the source clock sits above that, the more any applied groove just blurs into the baseline jitter.
One measurement tip worth more than any spec sheet, including anything Iâd ever publish: ignore RMS/average jitter figures. A clock can average perfectly over time and still throw individual spikes hundreds or thousands of ”s off. Your slaved hardware fires on the individual edge of each clock, not the average â so itâs the spikes that affect the feel. Always look at Maximum Jitter and Maximum Push-Pull (sequential event jitter sum).
None of this makes solutions like this pointless â for nudging a device or two into time by feel, or the shuffle stuff, it does things the regular DAW wonât. Just worth knowing what âprecisionâ means at the final MIDI output. Grab a cheap logic analyser and measure your own rig. Far more useful than taking anyoneâs word for it, mine included.
martibs - just adding this here as the forum considers me a new member 
8 Likes
I mean, you donât even need to know all this technological jargon - a plugin, a software running on an operating system, can not reduce any instability introduced by that same operating system.
On a slightly off-topic, but still related note: how does the MIDI Din clock stability on Elektron sequencers (Octatrack for instance) measure vs an average audio production system, running Omniclock or the like?
EDIT: Found some information here
1 Like
The link from Modwiggler discussion with that information is absolutely rubbish from 2017, we are not going to get anywhere with people just coming out of left field with numbers. We had tight midi, sync, in pro studios throughout the 80,90s before everyone jumped into the in the box and lost all their groove then spend hours mousing around to create swing. E-rm, Mioxl, Expert Sleepers etc etc, loose their precision as you building your sequence, I have tried all this, the only solution is Gridlock II from InnerClocksystems, the guy does not even talk about who his clients are because professionals dont have time with BS, studio time is money and everything has to be on and on the Grid. I do not know how much time and money I spent with all these mediocore interfaces and their clocks, when you pressure them to publish results they go silent, just look at the drama with Roland TR-1000, its a drunk sailor, oh we will fix in the computer, then whats the point and Roland knows how to do it, but they wont cause its not the same company that it was when they made all that great gear. If you serious about having accurate timing just buy a Gridlock II, I did early January, running 17 machines all in tight sync, the offset is accurate for over 3 hours, yes I recorded 3 hours of 17 machines firing off onto another computer. You do not have to print audio, you can just keep working on your midi parts and arranging and if you print audio and bring it back in, it is in total sync, it so exciting and tight that you end up playing well, you create your OWN POCKET, it becomes your swing, this is not rocket science, no one is making product to help musicians, they all trying to hit the $49 plugin price and $399 for synth drum machine price, because market research shows that. If anyone wants more information, I can help out no problem or reach out to Innerclock, we are all in this together.
1 Like
Hi Innerclock â great post and thanks for sharing the traces and numbers. Seeing logicâanalyser captures is very interesting, so big thumbs up for the time put into this.
A quick read of the numbers you posted:
- Your reported jitter max of roughly +600 / â681 ”s and PushâPull max around 1.28 ms show exactly what weâd expect from a DAW â driver â interface â DIN chain: longâterm tempo is stable, but individual ticks can have noticeable spikes. This is highly dependent on whether USB or 5-Pin MIDI is used, as we state in the manual. Of course the users OS tuning, background processes etc. heavily influence this, as I am sure everyone can imagine.
- Your tempo range and the âspot onâ longâterm average confirm thereâs no systemic drift â the timing errors are short, occasional spikes.
Putting the numbers in human terms: if a bass player stands at the far edge of a 5âmetre stage while the drummer is in the middle, the bassist is already dealing with about 2.5 m / 343 m/s â 7.3 ms of acoustic latency just from sound travel. The ~0.6â1.3 ms electrical spikes you show are a fraction of that. That doesnât make the spikes irrelevant, but it helps give scale: theyâre small compared with many everyday musical timing factors weâve existed within as musicians.
A few clarifying notes from our side:
-
Measured performance: Omniclock uses 0.25 ms step resolution (1x) and, in real DAWâinterfaceâDIN chains, we see peak physical output spikes typically between 0.5 -1.5 ms (depending on setup) with much lower averages. That aligns with what you captured.
-
Musical context: many classic machines (TRâ909, SPâ1200, MPC60, etc.) had noticeably larger microâtiming variation than the peaks you and we measure. That imperfect timing is part of their character
-
I had a conversation about this with Rob Acid when we started development on Omniclock where he shared that modern timing perfection in his opinion was making electronic music feel slightly stiff and lifeless. He shared his TR-909 had much more movement and drift in the output of its sounds, leading to a better feel.
-
This chat was the inspiration to explore adding intentional timing drift to Omniclock on the shuffle patterns inspired by vintage machines. We decided to give users a choice: tight, repeatable timing at the edge of what non sync box setups allow when set to accurate mode, or controlled, musical microâmovement when you want character with the vintage settings. Micro timing variations are added into the output clock, even if shuffle is not engaged.
-
(Sidenote: I almost fell out of my chair when I saw what used E-MU SP-1200s sell for a few days ago ($8000+) and that has one of the âloosestâ sounding timings of all the legendary hardware sequencers/samplers. I believe it is loved (and so expensive) exactly for this reason, and of course its 12 bit low sample rate sound.
-
TR-909, MPC 60, Lindrum and many other legendary machines behave with exactly these âlooseâ characteristics, breathing life into otherwise very static patterns.
-
We have the highest respect for companies building dedicated hardware timing solutions. Those devices are essential for people who need absolute deterministic timing, and the engineering behind them is important and very impressive! Our companyâs focus, meanwhile, is on musicality âputting proâlevel timing and groove tools within reach for as many users as possible, regardless of budget. I think there is a space in the market for both.
Thanks again for the rigorous testing â real measurements like yours are exactly what the community needs to understand the tradeoffs between absolute hardware determinism and musical feel.
8 Likes
Also I think a big point that confuses people is if you give a drum machine super stable clock, it will still have its own internal sequencer feel/jitter. Etc. So giving stable clock doesnât quantize the feel of the machine and make it stiff - it just gives an accurate feel from the machine, instead of feeling loose and off.
I see some people say they record their machine with no sync clock and then quantize the audio in the daw to fix it later, and that is a feel killer.
3 Likes
Are you inferring this plugin is basically snake oil?
Thanks Erik â appreciate the response, and genuinely good to see a developer engage with timing measurements rather than dismiss them out of hand. No disagreement on the philosophy: intentional micro-timing for rhythmic feel is a legitimate pursuit. Our interfaces and matching plugins have provided sub-sample precision MIDI Clock / DIN-Sync and Voltage Clock transmission with variable swing and blur for nearly 20 years.
One nuance on the stage-latency analogy, since it gets used so often as a counter argument to the importance of rhythmic jitter reduction in improving and maintaining tight feel:
The 7 meters of acoustic distance is a constant offset, and the ear adapts to fixed latency completely â itâs just where youâre standing. Rhythmic jitter is variable, step to step, and we automatically entrain to a steady pulse far more readily than the ear reacts to one thatâs merely late. Static latency and dynamic rhythmic jitter arenât on the same axis â and the dynamic one is what matters for feel. That distinction aside, fully agree thereâs a space for both. Good luck with it.
PS â one thing worth adding to the above, since the âprecision = stiff / no soulâ idea comes up so often and I think it inverts the truth.
We run MPC-60 / MPC-3000 / MPC-4000 / Linn LM-1 / CR-78 etc in our facility so we know
Roger Linn himself put it well in an interview with Attack Magazine:
âThe playback timing should be very accurate⊠if the note dynamics and swing are right, then the groove works best when the notes are played at exactly the perfect time slots.â
Stiff does feel good in the right place. My opinion is that itâs actually essential. A solid rhythmic performance always has a few elements bolted down hard â a rimshot, a kick/snare spine. That solid and predictable core isnât the absence of feel; itâs the reference the feel is built against.
The mechanism is really an expectation/reward thing. A steady rhythmic core sets up a prediction â our brain knows where the next hit should land. Feel is the play against that prediction: a deliberate push, a drag, a hair behind. That violation of whatâs expected, and its resolution, is the tension-and-payoff that makes rhythm actually feel like something. But this mechanism only works if there is something to lock to. Lock the rhythmic core down and every deliberate deviation reads as intentional and expressive. Let the core itself wobble and the brain canât form a stable prediction in the first place â itâs all mush, so the push-pull has nothing to play against, and it stops reading as expression and starts reading as slop.
So itâs not that loose / jittery = human feel. Feel is the deliberate violation of a confidently-predicted grid. Take away the confident prediction and the looseness has nothing to violate â you donât get more feel, you get less resolution on the feel youâve got. Tighter foundation, more the subtle movement registers; looser, the more it dissolves. Thatâs why precision and human feel arenât opposites â a tight rhythmic core is exactly what lets the expressive stuff land.
To conclude â Prince (arguably a master of feel without question) put it better than any spec sheet, on Joy In Repetition (1990) where he sings âa little bit behind the beat, just enough to turn you on.â
Behind the beat (micro-timing placement), regardless of machine or human, is a concept of relativity â one element placed fractionally behind another. But this placement can only reveal itself as intended when that deliberate lateness is set against a strictly stable, predictable reference. Take away the stable reference and âbehind the beatâ loses its meaning.
And the inverse applies â the more stable and predictable the rhythmic reference, the more micro-timing placement matters.
As I said in my first post I did not want to hijack this thread so I will back off. In the end we are all just trying to make our tools better at enabling our human expression through music â there is no right or wrong.
7 Likes
Not at all.
It does real things some DAWs wonât â per-port manual offset, and clock-level swing/groove for feel. For coarse skew / offset correction it works fine.
My point was much narrower: the âhardware-grade precisionâ language just doesnât hold up when you measure at the output pin on any standard class-compliant MIDI interface. Thatâs a claim issue, not a âthe thing is worthlessâ one. And the Digitakt numbers I posted cut both ways â standard MIDI clock from respected hardware sits in a similar band. So itâs not software-bad / hardware-good. Itâs that MIDI clock generation, from anything, isnât always as tight as it could be â and in the end, that matters. How much that matters is a purely personal thing.
Think of a photographer choosing between two lenses, or a graphic artist picking a pencil for detail work. Both lenses are good, but one resolves more detail and clarity. A blunt pencil still draws fine â but give the artist the choice and thereâs no argument. Same concept; our creative medium just happens to be time.
3 Likes
Its snake oil, so is E-rm, Mioxl, and the ones I tried, I started on Atari in the late 80s, it did good, some analog sequencers where also amazing, we had tools in the studio to fix offsets of many machines, that gear is gone because it went all into the DAW and that is when we lost the groove, it became cut and paste, then plugins, then plugin latency, cheap plugins, pro plugins, then abelton with launcing clips live, dont get me started on AD/DA converters because all of sudden everyone is a mastering engineer. If you are serious about your music, you want to create really nice productions that have a feel and you want to hear your gear, that you spent thousands of dollars on firing off in the analog domain via midi then just get a Gridlock II, I even went into eurorack for 3 years and the only sampler that could give me tight drums was the Rosssum Asssimila8or and it sounds good. I hear someone saying the Sp-1200 has loose sequencer, what you need to understand that triggering audio sample from a well built drum machine will always be faster, tighter than you triggering an audio sample with your midi keyboard to the DAW, that sp-1200 has very little to process as appose to your DAW. This is getting really old and we got a new generation who think we had computer back in the day like we have today. Its 2026 and we still have native plugins that introduce latency.
I rather have a precise machine that does exactly what its suppose to do so that when I play off, on, or ahead I dont have to quantize, because I know the machine is tight. Tr808,909,303,sh101,606,707 and couple more, all have the same amount of latency so when you sync these machines up it feels tight, but the moment you introduce say an Akai mpc or S sampler, or Emu Eos sampler, you will feel that those machines are lagging compared to what you triggering from the sampler, Atari +tracker program+ Akai S sampler= tight drum and base, jungle, etc.
Tr-1000 is a drunken sailor, they never will be able to fix it, Tr8s is 10 times tighter, they screwed up the design.
At the risk of seeming foolish after all the highly technical talk in the preceding posts, here goes: I donât get it.
Iâm trying the Omniclock plugin now and it seems nice - particularly the swing control - but when it comes to midi clock sync, is me setting the latency on an output port to -15ms in the Ableton MIDI settings functionally any different to me setting the Omniclock plugin to -15ms? Is it more stable/reliable than the Ableton clock in ways that donât come across in my quick testing?
To be clear Iâm not being a hater; I genuinely donât know!