A couple tips that I learned long ago back in my hardware synth days that might ease your mind (or drive you crazy) but at least you might know more about why this is how it is: 1. MIDI timeclock is not sample accurate, and will never be sample accurate from a device just due to the limitations of the time-code standard and the individual inaccuracies of the internal clock of the device you're using (drum machine, sequencer etc...). This is not something you can fix, and this is why most producers usually sampled their sounds into a single sequencer or program for timing accuracy. You'll notice this when DJ mixing old-school dance music too, this isn't just a tape issue, sample accuracy is just that imperfect to begin with out-of-the-box. 2. MIDI events are also inaccurate. Because of the fundamental way midi sends the data, you simply can't get all the data to arrive all at the same time. Basically even if it's seemingly simultaneous, there's several events happening in quick succession and therefor will never be sample-accurate due to this without having each device have some sort of buffer built-in that is sample accurately matched to the same clock (which again will never be sample accurate 100%) 3. There's 2 main ways devices sync. MIDI Time-clock and MIDI Timecode. One sends out a sync message (usually about 16th or 24th divisions) and so all the space in between those times can easily fall out of sync when going against the perfection of sample accurate DAWs. Timecode is based on FPS in relation to tape sync and uses reference points to sync the clocks of two devices, and any time they drift in anyway will result in the imperfection you're seeing in your DAW. 4. Additionally on top of MIDI imperfection, you're also dealing with audio jitter at all points of the AD/DA chain, so that's also another point of inconsistency. So all that combined together means that timeclock syncing different device clocks in by its very nature never going to be sample-accurate for the foreseeable future. This is also why often times going in/out of the digital realm when mixing/mastering you'll get timing variations and imperfections that are not sample/phase accurate because of the delays in all the electronics and the digital-to-analogue conversions going on. If you're all in-the-box you wont notice this because you're hearing the same imperfections equally at the same time and rendering is all software based so the sample accuracy is uniform without live processing jitter. Ultimately for synth parts and stuff it's not a huge issue and provides some nice imperfection, but for drum machines, it's usually a sample-back-in or make loops from devices on their own clock and fine-tune and import grooves to get perfection. Brian Transeau has talked about this at length over the years due to his love of outboard gear, and also why most of his stuff has to be sampled in and nudged to perfection to get the results he gets. Love the channel, stay strong.
Ricky, this is the answer. MIDI clock is especially inaccurate and inconsistent. If you want a fighting chance, you have to record all outputs simultaneously, and if you edit them, move all tracks together to improve phase alignment with Live (or any other DAW) clock/grid. Many of the sequencers in these lovely hardware devices have their own micro timing quirks as well. A Linndrum pattern has very different microtiming compared to an 808, for example. The more you pull the thread on this sweater, the worse the situation will become. I recommend making only single take recordings, using all outputs to capture all tracks, and then shifting these to best align to the DAW you record into.
@@zedd_centauri This is also why peeps back in the day were so in love with the Atari-ST trackers with midi, because it was a single midi trigger source so timing could all go simultanious out and just capture live to tape. Other than the expensive timing boxes you can get are more or less simulating the tape clock signal at high resolution but it's as suited for the type of beat-clock timing that alot of drumachines use (some do it better than others I admit) but yeah. If someone wants the swing, you might record dry in and save that as groove data in ableton (one of its more killer features) so you could apply it sample or loopwise. Kinda takes the fun out of the hardware-based production, but this is how we all had to live with it, even in early all-digital setups.... quite frankly I don't miss it (besides the fun gear though).
I am so so happy that someone else sees this stuff happen. I have fought with weird MIDI glitches like this for so long and it bugs me out how nobody else ever seems to talk about it...like their workflow and creativity aren't just totally bombed by stuff like this all the time.
I never synchronize a machine to Live. I type the tempo I want in the hardware, hit record on Live, hit play on the machine a moment later, warp/adjust the beginning of the clip afterwards. No weird sync nonsense, no latency problems.
Bang on 🎯. Similar experience of frustration. Just a thought, would you yeild similar results reversing the process ie. Slaving Mac-mini & Abelton to the drum machine? Awesome stuff 💯‼️
Same here. No matter how loose the clock in a piece of hardware is (and some are pretty loose) I've never encountered one that's anywhere near as bad as MIDI over USB (whether it's a USB interface or direct MIDI-over-USB). Some OSes are better than others but it's baked in to the USB protocol (at least up to USB 2, which is still what most hardware is going to use) - data packets over USB are sent at a 1kHz frame rate, which means the absolute tightest possible timing is going to only be accurate to 1ms over USB, plus any timing issues happening in software (usually at the OS level) get rounded to 1ms increments. By comparison, a solid hardware sequencer (MPC2000/1000/etc; Octatrack; many others) generally has jitter under 0.5ms. I've measured the jitter on my own MIDI interface at about 7ms! For me, the best solution has been clock over audio (originally SMPTE timecode into my 2000xl, more recently an Expert Sleepers USAMO), which is as close to sample accurate as MIDI/SMPTE and your hardware's timing allows.
I know right. Been wondering for ages why it sounds all nice and groovy and then you record everything separately and that groove is somewhat or completely gone. Now just record everything at once in a multitrack which took years to figure out to be fair but certainly worth investing in a mixer and audio interface that can do that.
Hey dude! I've been a long time watcher of your videos - I appreciate how you are constantly digging into the nitty gritty and the unglamorous parts of production - I've learned a lot from you and you've kept me really inspired the past couple of years! Thanks - keep it rocking! 🤘
@@ceounicom It isn't jitter. Windows and Mac OS have never been real time OS's and have no way of guaranteeing events will happen on time. If they were, most consumer software wouldn't work, and they're designed for usability. Also, USB is inherently terrible. Back in the 90s when you plugged in a MPU401 your MIDI input was connected directly to your CPU by an IRQ line, and PC MIDI had the fastest possible response. It's been downhill since. When using a Sound Blaster or other cheap MIDI card without the IRQ, I had exactly this same problem on my 486. I could even make my system crash by plugging in my MC303 and hitting "play." Remember that MIDI was invented before most instruments *had* a CPU, and it's been good enough for performance for almost fifty years now. I think a Keyboard Magazine article in the late 80s described the situation as "in the time it takes a MIDI note to be transmitted, sound travels a few inches in air. If this delay is too much for your composition, then you might need to put your instrument inside your head." Jitter is an entirely different thing, and if you have one millisecond of it, what comes out of your speakers will not sound like music at all.
@@stickyfox I know first hand that these aren’t real-time OSs, at first I thought it was my software, then my audio interface, and finally when I tested my built in mic I knew it was windows. It’s only a few moments off but that’s enough to hi 1/8th of a beat off. Currently trying to find a fix.
Man, I relate to this so much, I just abandoned trying to track everything into a daw. If I choose to produce on MPC, I do everything in the box and record the full track and if I don't feel like it I do everything in a daw. The MPC mix can be dirty and unbalanced due to its limitations but I prefer the swing and feel over the clean sound.
Rick great video, the best way to get everything tight is to have a midi card sitting in your pcie slot on your computer like a RME Hdspe etc. I have the Erm Multiclock, that is maybe the best way to get computer, external gear, sequencer etc to talk to each other and when you go to print audio, you wont do any adjustment accept the beginning. As of right now Erm plugin for Cubase is not functioning correctly, according to them there is an update coming to fix that, hopefully this not the case for other DAWs as it looks like Abelton can be solid with some offsetting, working on Bitwig to see how it works. Unitor 8 for Emagic(90s)still is pretty solid and Mac is providing USB midi driver support, that is pretty tight. Another company from Germany called Miditech 16x16 USB/MIdi has slightly better results than the Edirol. There are also loads of folks using Roland SBX-1 doing all the sequencing dawless and then printing from DAW clock or just Rolands clock and trimming the downbeat. I just cant believe that in 2021 we are talking about midi not being what it used to be. Also if you can please make a video about the Erm Multiclock and how to use it as USB/interface, still have not figured out all that potential and that manual needs to be thicker. There are some folks using the ERM plugin from daw feeding something like Kenton Midi through and getting great results when printing the audio. If you have been creating a track all Dawless and love everything about it, and start printing the audio into the DAW you are going to loose that vibe, there is nothing like all the gear firing independently, this is why so many of 90s dance/hiphop has so much human feel to it, because lot of folks just record straight from a mixing desk onto DAT or tape machine or adat. We need to help everyone get this thing sorted, the amount of videos about timing/latency/jitter etc is just crazy. Thank you for all your great work. Cheers.
This is exactly why I love all of your videos. I have ran into things like this with no way to explain it. I actually have a M-AUDIO USB to MiDi adapter and it has never failed me when I needed it!!!!
really enjoying your videos. i‘m facing the same problems at the minute getting my hybrid setup working. makes me feel a bit better to find out that latency and everything being off really is a thing
This wasn't boring at all. This is extremely similar to a rabbit hole I went down on 2-3 separate occasions and it's nice to know others have encountered this.
Thanks for making this video. I’ve started recording a separate 4x4 kick from every machine just so I can make sure each track is tight once it’s in ableton. It’s maddening and I spend more time on this that the actual track.
And this is why after 20 years of cubase and Ableton etc I went back to outboard. That and I’m a computer engineer and sitting in front of a computer in my spare time just sucks my creativity and work flow 😔
Midi over USB for any sequencer kind of thing seems to always be funky. But the model cycles is just pretty bad over midi or usb. The only way I manage to get it perfectly timed even on the first few beats is to hit record in the DAW but have the cycles play an empty pattern. Then launch the pattern you want to record and that extra bar helps to get it all synced in time for the second bar.
i don't own any elektron gear, but this is my standard practice. it doesn't matter if i'm sending MIDI to DIN devices via mio4 or direct to MPC live via USB - an extra bar at the beginning is mandatory when i'm tracking MIDI devices into ableton.
I love going down these rabbit holes with you even when they seemingly lead nowhere lol. I still learned various things and got a kick out of watching you wrestle with this. Hey, maybe your "eureka" moment will come later. Keep these types of videos coming. ✌️
This is my favourite channel on RUclips. It's funny cause I never would have thought of this kind of video, if you pitched it to me I would have said it wouldn't attract an audience, but I love every video.
I recently connected and mixed all my hardware on my father's mackie analog mixing console. Although the noise floor was significantly louder, I couldn't believe the difference I felt in playing the gear. The only way to describe it is how immediate and responsive everything was. DAW latency might be just a few ms but when playing through it for a decade you can certainly feel the difference when you move over to electrical signals rather than converted signals.
It was very insightful. You made several basic not talked about issues with syncing and recording from hardware into Hybrid setups. I never considered running the same sound side by side to check the sync. Genius! I always end up snapping to grid....
Hey Ricky, I can't say, how happy I am, that you made this video. For me. it's an confirmation, that I'm not such a great foolish at all... :-) . I've used Logic and now Studio One is my DAW of choice (again on Mac). The timing is way better, but you can't get totally rid of jitters and latency. My workaround is the same, as you explained: 90% of my midi-sequencing I'm doing with my Squarp Pyramid (btw: it's great!); then I'm recording my sessions and songs via an multi-input-interface direct in Studio One (or Capture), set the tempo there manually as it should be and drag the recorded tracks all together in the right grid. That works great for me. My opinion: if you're working with timing-sensitiv hardware like drummachines or grooveboxes and you will grab them into a DAW (without the multiclock...), then that's the way it goes...unfortunatly. Stay healthy!!
this video is of value ricky bc we do struggle with these issues but without the understanding or equipment to isolate exactly what is causing these hiccups, it helps to see problems like this elsewhere so we are more informed with respect to those random software related issues as well as products that are fails, esp when those products are very expensive tho it may mean little and be unseen by you al together thanx for absorbing the impact of buying an expensive dud, it sux and im sure many will not end up in the same spot bc you did already preesh mane.... tempo issues are very confusing in the beginning. i had a real issue learning that the Orba's 100 BPM is MPC beats 97.7..... it took me a long time to accept there is an odd gap in actual tempo so i could just move on with the recording (i wasted a lot of time thinking i was doing something wrong when i just didnt know software kind sucks sometimes with these issues) i have never mutli tracked into a DAW. im just scratching the surface on software at all and find this info very re assuring. i wish i saw this sooner, PLEASE do a beginners guide to recording tracks including these types of issues that may trip up a noob like myself. the issues that wont be mentioned in the user manual such as this tempo discrepancy with software and hardware. this video will help me harbor additional patience when recording in the future and i think you have enough others to speak on that would help someone like me, thanx ricky... keep it Gnar Gnar
I believe the dots at that zoom level represent each “sample”. I’ve always wondered about sync issues like this especially with the Virus TI2. Thanks for making this video! 🙏🏼
I love this video because this is me right now trying to multi track sequence and record my Syntakt thru over bridge in Ableton. I felt very alone and frustrated. I no longer feel alone.
Watching this really makes me appreciate knowing I am not insane, as I did the same experiments a while back before finally cracking and buying an E-RM. I love that you are looking for a cheaper alternative though, I just haven’t found anything besides the E-RM that works. If you want to try a USAMO just for fun, I can send you mine. It didn’t like my RME interface and is just sitting around my studio.
Good subject. I have ran into this problem lots over the years . I now only use usb after I've arranged a song using a squarp pyramid to sequence then I bring the computer in for live instruments and vox. I thought it was just me hearing usb slip in timing :)
Years ago while I was just switching to computers from hardware I learned about this issue And my fix was to use the computer sequencer to trigger my hardware using midi notes And record the output from my hardware into my DAW This was the only way I could make sure the clocks were rock solid
I've experienced this too. I use Windows and the DAW mangles midi out.. Don't always record your parts as individual parts. E.g., it can be better to record your drums in 1 stereo mix. You lose the ability to edit the midi and different volumes, but you retain the groove and vibe. When I use the DAW and create drum parts then I always have to fix the audio wave. Timing is much better from the Akai, or from an analog clock. Analog really grooves for locking synthesizers and drum machines together.
I've been stuck in latency/jitter hell for a while now, also thinking of just tracking to tape and learning the banjo lol... thanks for the relevant content Ricky
@@longlostrobots8146 yes I know! :) It’s the clock jitter that drives me crazy at this point, when recording multiple takes it can create issues... mostly with percussion where I want really tight transients that are consistently in time. Oh well, it is what it is 🤷
I’ve been using the Innerclock Sync Lock for years now to sync midi and din-sync hardware to Ableton as clock source. With some clever routing you can make the setup work even with monitoring via Live, allowing adding insert plugins, sends, etc.
Seems like it’s around the same price as the erm and I don’t see any knobs to adjust the clock delay, I thought that was the way to get all your gear in sync
Absolutely not boring dude! I'm actually having lots of trouble to sync my Mother 32 to Logic Pro's clock and recording something in time... I don't know what's happening, I've tried everything too, DIN, USB, but the sequence almost never starts on time and is all over the place. Computers mess tempo up for sure!
I usually skip the ads but Nikkolas Smith is legit. excellent shout out Ricky. 👍 I was in art class with that dude in middle and high school. He was amazing even back then. I was trying to make bongs out of ceramics. 😆
I've noticed it too. You can hear it drift even without external syncing. It's messed up. The worst is syncing two laptops together, both running Live. It's all over the place. Sloppy mess.
Seriously though, I'm having much better luck running my 2K XL as the master clock into a Motu Express XT and clocking Ableton from that. I can see the timing jitter from the MPC replicated in the Ableton BPM preview. The key was honestly getting a dedicated midi interface for cleaning things up. BUT, when you disengage the MPC as the master clock you lose that organic swing / jitter. So it's now rock solid in Ableton which is bittersweet. If there was a way to record the BPM drifting as automation that would solve the last part of the puzzle for me in regards to maintaining the MPC "groove" in a hybrid setup.
try overbrige. take a midi-signal from the digitakt to clock another devise while running overbrige. thats my way to "not buy a multiclock". but if i would have the money.........
I agree. OB is the 2nd best solution. Still not as a good as the E-RM since OB adds a lot of latency and chews up the CPU, but it’s the best alternative I found as well.
I have the same with the 4k but i guess that some machines are more musical with it's jitter than others. Best way to record using multiple outputs of the mpc's so you can mix it. Problem starts again when you want to add things later....
damn this is basically hard proof I need an ERM. Even multitracking into bitwig using the pyramid as the main clock for my synths with no midi sync in the daw will leave me with weird jitter
Really. I was going to buy a Pyramid thinking it would improve jitter. Guess the money is better spent on the ERM in that department. I have seen the midi data jitter and timings for a lot of devices, the Pyramid is better than average at that, the Elektron stuff is satisfactory. But who can afford a Sequintix Cirklon or even wait for it to be build as they are back ordered by 2 years. Lol
This reminds me a nightmare that's a bit of topic, but related to timing problems and computers. At the time I was working with computer as the heart of my live set-up (with NI Mashine), I did many records with my band recording live with the audio coming from my "live" laptop to my studio desktop in two tracks while recording the other instruments. Afterwards, I opened Mashine as plugin on the desktop computer and re-recorded all the sounds from Mashine in separated tracks for accurate producing and mixing. No problem with this method but one time. The recording messed up somewhere and the timing between the audio-recorded tracks and ITB Mashine tracks where totally off. It started good enough at the beginning of the tune but turned more and more out of sync. At the middle of the tune, all the groove where lost and after the break, it was just a mess. We tried many things : record again from the laptop, do the stuff in the box in the laptop, slightly change the tempo of the whole project and so on. But but nothing worked. It remained far from the groove of the live recording. It looked like he tempo had slightly changed during the recording. All we had to do was to keep the electro stuff done and record again the other instruments. It only happened once, but it taught me definitely that computer timing can mess up. Now I've turned my live set-up dawless and we record live with all the instruments at the same time. No more timing nightmare.
Though you did not provide a solution, I was relieved to find that I am not the only one in this very frustrating predicament. Yea, I guess latency will always be there, we just gotta choose how we work around it...sucks
20:59 YES, always run your audio interface (and of course midi interface) on a separate USB, directly to the computer. Everything else like keyboard, dongle, drive, etc... can run from USB hub. This is definitely a source for this kind of timing issues.
@@casecold1864 The good old firewire... It feels so good to have my old Ultralite firewire running rock solid... midi from my Ultralite in FW is ultratight... and OP-1 usb-audio-midi code is the best thing you could possibly imagine... boutiques on the other end... nevermind...
After a few years of struggling with midi jitter out of computers i think the Roland SBX1 has solved the problem. the ERM is next level though with it's timing shifting...
Thanks for the video Ricky, wasn't that boring :) This shows the world of hybrid hardware/daw setup is still an utopia nowadays. The best thing to do is either choose between DAW-only or hardware only in your project, or if you want the best of both worlds, go DAW/powerful groovebox and sample some synths when needed, knowing you'll have to manually warp/sync them.
That was a fantastic video. The nuts and bolts of it all can really stifle creativity. I sometimes get a blast of inspiration but, by the time I've got everything working it's gone. By the way, does the little chap with his tongue out have a name?
@@arjanpetersen not really, latency can be compensated. the big issue isn’t about it. the real problem is jitter, which can’t be compensated in any way. you can calculate how many samples of delay you’re having over each buffer size and just compensate. but overtime you’re gonna be off anyway at some point because of the jitter. I simply solve the issue by using the device called USAMO by Expert Sleepers, there are other options but this is probably the cheapest that actually works.
@@formerself4372 it probably sounds compensated, but it's definitely not sample accurate compensated and definitely will jitter a little bit, making it off the grid sometimes. that's the difference. it's useful because I can simply edit all my jams and everything will be synced from start to finish, i don't have the need to ajust the samples because they were always on time with no fluctuation in time.
You'll drive yourself nuts with this Ricky if you go down the MIDI timing rabbit hole, you'll start to hate the timing of DAW's. Coming from the 90's though, hardware sequencers and the first generation of computer sequencers ( they weren't even DAWS back then ) please be reassured that although this problem still exists it's exponentially better then it was back then. Midi timing from computers and sync ( post Atari ) literally sucked big time back then. For sure. We did managed to make tracks with it though :)
your actually introducing latency by recording with in on its supposed to be strictly used for monitors hence the in terminology the idea with Ableton is you would have two tracks one for monitoring another for recording seedtostage has a good video in it.
I had an email converstion with the creator of Expert Sleepers USAMO - Andrew Ostler,and he said "there are two sides to every MIDI connection - the transmitter and the receiver. The USAMO gives you a near-perfect transmitter, but it can do nothing about the behaviour of the receiver." so Your synth,or drum machine have to a good/perfect midi receiver also to get accurate timing,no matter if You send perfect midi signal,I think this is one of the deepest rabbit hole,I end up with other method CV - control voltage, I sequence my Moog Slim Phatty,and Arturia Microbrute with control voltage,but not every device have CV input.
That's the reason we went to recording everything at once using a mixer that has outputs and then going straight into 8 channel interface. Certainly not the cheapest option and if you have a small home studio set up might not be possible but highly recommended if you can do it. Also saves tone of time since you just record everything all at once too.
@@formerself4372 well you record all the instruments at the same time rather than doing multiple takes so there might be midi timing issues but you don’t loose that swing and groove and relationship between the instruments
You actually have two problems here. - One is the audio clock issue on basic audio interface, that can be solved with an expensive wordclock; If you’re recording lot of audio, it makes a huge difference; even decent audio interface that are supposed to have internal clock are bad and have timing oscillation and jitter. But not everyone can afford an Apogee or Antelope and such... - The second issue is actually from USB MIDI interface that is often out of sync with internal clock, it’s often due to poor drivers on mass market midi interface. The more robust midi interface that I’m aware of is the Motu Midi Express; while back it was the only interface that had solid driver that would avoid this kind of issue. I had to do plenty of test for my work to link computer system (OpenLabs keyboards) with MPCs. We had to spend a LOT of time to develop in-house drivers for the Presonus interface to be rock solid for our system to work. it’s been a while now that I haven’t look into it, but I’m pretty sure the issue is still the same. I doubt any mass market audio/midi interface developer spent time to develop rock solid midi drivers for their interface. Only guys like Motu, Avid and such might have decent solutions... not the kind of gear we find in (bedroom) home-studios.
Since I saw a vid from you for the first time about hardware latency issue's you got me busy diving into it.. So now I want the multiclock too. Guess that's the only device to fix latency for 4 synths at the time.. (should have kept hardware out of the door LOL) Anyway, thanks for all your vids (on this topic) been really helpful.
I have had good luck with the Mio too, it works well with old synths for SYSEX (DX7, D50, POLY 800 for example). USB to Midi is basically USB to serial and uses similar chips to FTDI USB to serial parts. I did a deep dive a few years ago on this stuff because I was having issues with timing from an application over USB to a (non music) product. Basically USB has a "frame rate" and doesn't process the request as soon as it is received. It will wait up to 1 milli-second to process and send the request. The request can have a lot of data in it, but it is only sent on a 1msec boundary. (I may be paraphrasing here). Older MIDI implementations in the pre USB days would have had a Serial IC "UART" such as the 16550 style that talked directly to the CPU and interrupted the CPU the instant a message was received. Now a MIDI message must navigate the USB subsystem first which adds an unpredictable amount of latency (Jitter). The MIDI standard itself for clock is pretty amazing, it is a single byte (0xf8 I think) transmitted to sync devices. This means if you are just looking for sync, you should be able to generate an interrupt per byte, and just look for 0xf8. That is very fast, a few micro-seconds, for modern processors. Compare that to the USB latency of up to 1 milli-second and you start to see where the USB system makes things worse. MIDI clock is 24 PPQN, so at 100BPM we would see 40 pulses per second, or 1 pulse every 25 milli-seconds. A 1 milli-second jitter is 4% error on the individual 24 PPQN pulses. You wouldn't think you would notice, but it's there! Coupled with sampling latency of the audio, and you have to manually time shift a lot of stuff. I'm just getting into Ableton, and I'm finding stuff like this really kills my motivation. I'm trying to generate enough talent (i.e. buy enough stand alone gear!) that I could just track to final mix. Or I'll get rid of the clock altogether and get really really into ambient stuff.
My clock setup right now is an actually having my drumbrute's MIDI out hooked into my iConnectivity MioXM, which distributes MIDI Clock to all of my devices as well as my computers over the network (using RTPMidi). It's impressively solid compared to other options and I didn't need to buy any new things!
I have given up trying to track external instruments in Ableton, particularly if I'm using instruments in the DAW as well. Unless I'm 100% in the DAW I get absolute crap. Sounds like I should stump up for an E-RM, I've been using a launchpad pro MK3 as a master sequencer and using Ableton as a glorified tape machine lately just to keep things in time.
the ironic thing is "record to tape and learn the banjo" solution is why I was thinking to move to hardware lol but now i realize i won´t have a smooth transition from daw to dawless, because right in the midst of the hybrid phase is where most issues happen so i guess with latency is either best to be daw-exclusive with maybe some midi keyboard _or_ full on dawless
I had the same syncing issues recording my elektron digitakt via USB/overbridge - things were not lining up consistently and I was getting weird timing hiccups. I found that if I set all my ableton audio inputs to "IN" and turned ON "Reduced Latency While Monitoring" in the options menu (I think ON is the right setting - could be wrong)... then things generally would line up perfectly. Another suggestion I heard someone make about syncing problems is to set your hardware sequencer as the master (ie. Digitakt) and slave ableton to that when you record.
Simple technique here: add a 1 bar 'count in' sequence before your song plays, like how bands do to cue the start of a performance. It can be a simple 4/4 kick pattern or whatever. Then once multitracked in your daw, it becomes much easier to line up your tracks on the 1 as they all share the same count in. Once aligned you can then just trim out that first bar
I always thought this was because of Midi to Control Voltage converters in analogue machines. I don't mind, I kind of like it, but I don't have it as severe as in this video. Never had any problems with missing notes or crazy delay's and it doesn't ruin the swing, some this might be an other problem, like midi jitter.
Great video. The worst issue is trying to keep two kick drums from different sources in sync.... Another possible workflow might be - 1. do your sequencing in the hardware sequencer / drum machine -2. export the midi only to the daw . 3. use Expert Sleepers USAMO, which sends sample accurate midi (that's the claim anyway) out of the daw as audio which the USAMO box converts back to midi.... but I haven't tested this! Another issue with Daws (or Pro Tools at least, I presume Live does it too) is they skip a sample every couple of bars (there's a formula that dictates how often this happens, something like dividing the sample rate by the tempo or the other way around).
Don't monitor in the daw. Set monitoring to off. There is always latency on live monitor It used to be recommended to give an empty bar after midi start to avoid the 1.1.1 jittering between synced units. Mpcs don't perfectly phase cancel over takes, ever (even on master) I guess everything won't if the cycles doesn't. It all has to go down at the same time.
20:28 We may not be able to hear less than ~10/1000 of a second of latency, but of course as soon as any kind of phase shift from parallel processing comes into play, even a completely undiscerning ear is going to be able to hear differences of just a single sample, eg 1/44100 of a second. I think a lot of these latency issues come down to how well the device drivers are written as much as the hardware itself. I'm a big fan of RME gear on that front, always absolutely rock solid.
I spent the last 6 years chasing midi sync with a computer after buying the squarp pyramid sequencer to sequence bitwig. Now after years of experimentation and measuring (and basically quitting making music over all these fucked up timing issues) I finally got the recipe right: - Get a modern soundcard with a midi interface, because this way your signals will be just a bit more "in sync" with each other. Dont expect 20 usb devices to work together in any meaningful way. - Get a DAW that will run off midi clock instead of being the midi master - Let your midi master clock be your device (eq. octatrack, squarp pyramid etc) Now to the more esoteric: - Avoid plugins that introduce latency, bitwig for example has an indicator on individual channels that lets you know how much milliseconds of delay you get for using certain limiters or reverbs - Run the lowest latency you can afford, but not so low that your system is always on the brink of collapse, as this will actually increase jitter ***UNCONFIRMED SUPER VOODOO TECHNIQUE**: Run at 96k because ehh..i dont know something about USB packets being handled differently and shit. This one actually brought down the midi jitter to reasonable levels and get a 4/4 Kickdrum pattern within about 1/4096 or 1/2048 which is good enough for me ;)
i am a beginner producer and i want to get rid of the topic latency early on to build a functional workflow around it. ideally, i want to control (via midi clip or keyboard) an external sound source and capture it with the audio in of my interface. i also want to use ableton effects. now i have some ableton drums playing and the external instrument receives midi to play, which i monitor in ableton. now i adjust until it sounds good together. now i want to record the synth to audio and when i start the drum clip and the audio synth clip together, it should sound exactly the same as when i was monitoring before. i can record with ableton effects or i can record the dry signal at my input. the dry signal is faster than the monitored one (with effects). now, i put the same effect chain on the recorded audio. all 3 variations (playing it live, recording the monitored audio with effects and putting effects on the recored dry signal) all sound slightly different... i do not think, i care too much if something has latency, but i need to recreate the exact same timing i adjusted to in the "jam" phase. with or without latency, it needs to sound the same. i do understand this topic gets much more complex when introducing a global clock, but i am already struggling with this basic step...
to me there is a change in the groove of any machine as soon as it's clocked by another one. I also believe that each machine receiving a clock, is constantly trying to follow the clock it receives, and then has very slight changes in its timing, but i don't see it as a problem, to me that's the interest of hardware, that it doesn't play exactly the same way everytime. Btw I think that not all hardware equipment will start it's clock properly if you don't start recording at 1.1.1 , i mean the beginning of the arrangement view. Happens to me with some drum machines and synth, mainly the older one but some recent synth too. Also if you wanna record with minimum latency, you have to trigger the "off" button, not the "auto" or "on", that let the sound go through Ableton, then to your speakers. You need to use the sound directly from your audio interface, to the monitors. I proceed that way with an apollo twin and i have 0 latency problems. Of course that means that you are processing your sound outside your daw, or recording, then processing. Love the channel keep up the nice videos and music 💪🏻
I don't think it's latency when you record the machine directly with no clock and align manually. But it's definitely weird. Could be anything, even the Mac.
As tedious as it is, I appreciate you posting it. So many forum goons that don’t make any music have such awesome clocking and all their problems are solved. This isn’t a solved problem, it’s not gonna BE a solved problem. Our perception really exceeds what technology is capable of, and in ways we’re bumping up against the limits of physics and CPUs when we start fine tuning milliseconds.
Good vid, thanks. To deal with it I just turn off swing in my gear and able the swing in Ableton. I actually find the Groove feature in Ableton really powerful and it also quantizes before applying the swing so for the most part it fixes any issues.
I'd like to see the same test done with the ERM. I mean you say it just works but have you done the same test and confirmed it? It might be possible that it feels like it because it cost 600 or so bucks :D I have no doubt that the device itself does a good job (it's a german product afterall and if we are known for something then it's for being punctual lol) but I'd still like to see the endresults and the grand scheme things. Greetings from Germany
@@RickyTinez looking forward to the next video 😉 btw love those clock and sync related videos from your channel. I was afraid I was the only one to be all ocd about this stuff because it is hardly ever discussed on RUclips and I really wonder how all the many many other producers out there tackle the problem. Maybe a future solution to this problem could be a specialized OS which handles audio AND midi prioritized. Whishfull thinking haha
I bought an ERM years ago and it's never given me a single issue. Totally worth the price to never have syncing problems ever again, and the ability to adjust per track timing and swing is just icing on the cake. Granted, it's my master clock with all my outboard gear and Live slaved to it. It works flawlessly.
Is this really necessary, checking the millisecond with the mouse? It’s so ocd. What about the popular mantra, that we should be producing with our ears and not our eyes.. Let’s just take that micro latencies as a humanize effect and be done with it. Don’t check it and zoom it in.🙂 Think about how they recorded back in the days when there was no daw’s and it was recorded live to tape. The bpm fluctuated a few bpm’s across the whole track. And it sounds great.
marko polo has a workaround for latency, if I'm not mistaken. to track the song out of an MP without losing swing, create a dummy pattern on the MP with 1 quantized kick on 1 and put it before everything. track your stuff into the daw so that each track starts with that dummy pattern. then take tracks and align those kicks perfectly. this should bring the swing back
Wtf I thougth I was fking something up when I tried to record my two novatiom circuits and a TR-09 separately and got this exact problem. In this particular case is when Roland products like the MC-707 or the MX-1 help a lot. Since they work as audio interfaces you can make them send the main tempo and record multitrack while you are jamming. Hope more manufacturers do this cause' is a live saver.
Hey Ricky... I think it's Elektron Model boxes. I'm running an AMD 3.4gz with 16gb ram. I used to have the Circuit and Model:Samples as my main hardware center pieces. I had both plugged in via USB and the Circuit would be tad off but the same each time, while the M:S would be off and start late every time. I sold both pieces of gear to try and let new energy into the studio, as well as learn sound design instead of relying on presets/patches. I now have the M:C and just reordered the M:S. I Love Ableton and have heard ppl say the Model boxes suck over USB. Maybe someone will figure it out one day. As always, thanks for the video.
There was a take that the model samples didnt start at one. Ive had this problem before with hardware when you stop the transport and when you press play again the machine doesnt start on 1 but instead starts on where it stopped before. Other than that i had this problem with the digitakt before using overbridge, with overbridge everything seems to line up fine every time. Also for the last technique you showed, if your device has metronome build in, i usually turn the metronome on and record some bars of the metronome (you can also just sequence a 4 on the floor sequence as metronome) so you can use it as reference to cut all the audios.
@@RickyTinez no, what i meant is before starting to use overbridge i went though this torture all the time. with overbridge never noticed anything going out of sync which is amazing!
USB involves multiplex switching and the packets can fall behind or end up ahead of others depending on a number of factors i.e how many devices you are using that share a controller, if you are using a hub. I'm not sure of everything, but there could also be the implementation of USB on the device and clock jitter on the device itself because USB is synchronous. Someone needs to make a PCIE card for just MIDI, which would be fabulous to have. Some sound cards do have a direct MIDI DIN asynchronous connection, but they are often expensive. I always get clock drift when syncing via USB from Ableton to my Beatstep, and then using MIDI DIN to sync everything else. I can't recall if I have tried without anything else on the hub or my motherboards controller, but because of how USB works, I think you are likely to get some amount of drift or possibly jitter from the device.
Ive spent so much time with these types of tests lol. It’s midi jitter via usb that’s the issue (Although polyphonic midi also has a 1ms latency per note) cuz usb is sent in lump sums every ms or more. A good solution is serial midi with an audio-source sending the data. I have had good luck with the expert sleepers esx-8md and ES40 via spdif. The learning curve for that was quite insane for me, but there is truly no jitter once you get it working. I am now trying to sequence everything and record it live one take.
Latency can be accounted for with a timing offset, jitter is the real problem. MIDI over USB causes jitter thanks to things like multiple packets per frame, device contention, and host OS priorities. I'd be curious to see what kind of timing guarantees different OS class-complaint USB MIDI specs have, if any. Class-compliant USB audio is tuned for accuracy (at least for Apple).
Usb does matter, I used to have horrible jitter issues on my ensoniq asr X which I fixed by making my roland tr505 the midi master, at the time my setup was the original bass station, a tr505 and the asr x. But although the asr sounded great it was useless as a midi master. These days I use CV gate and sync heavily and only use midi via usb but with these volcas I record directly into a br80 and avoid the daw. Then I chop up those samples into ableton sample packs but I isolate my hardware from my daw and only grab the hardware as samples.
Most of the latency you experience in a DAW is caused by the OS routing audio through the DAW itself. Try switching all your monitoring to off on the channels in Ableton and monitor the audio directly from the your hardware. I do this via RME TotalMix, which is a routing utility that comes with RME interfaces. I'm sure you have shown a similar thing for your Focusrite (?) interface in a previous video. The key is not to monitor ANYTHING through the DAW. It's not a slight against Ableton, you will get the same extra processing and in turn latency using any additional audio routing. Hope this helps!
Appreciate Ricky's search for alternatives, but latency/jitter problems between DAWs and hardwares are the exact thing that products like the USAMO or the E-RM MultiClock were made to correct. They're pricy solutions, but definitely worth the money if you have a hardware sequencer that needs to live in harmony with your DAW.
you might not be sponsored by ERM but my takeaway was BUY A MULTICLOCK. Also, the dots are the actual sample amplitudes at the sampling frequency, the line represents the audio after it is converted back to analog.
I wish Ableton would have an option to send clock independently of transport. That would fix the issue of hardware needing to "catch up" to the clock when transport starts. Anyone knows of any sequencers, DAW or otherwise, that have this feature? Having clock separate from transport would also remove the need for something like the Multiclock, if you already have a MIDI interface. Or does the Multiclock provide other features? AFAICT it's "just" a MIDI interface that also generates clock. Great video, BTW. Exactly the kind of things I've been fighting with.
When using USB there is a minimun jitter that can happen because the USB Driver has limited "resolution" ( i.e. it can only send messages up to a certain speed, after that it can't so there will be a bit of jitter on the clock regarding that with any USB solution ) that's one of the reasons sync through USB will never be perfect, that doesn't mean it's unusable, it's just that there is a minimum jitter when using USB. That's the reason why some clocks will use AUDIO cable and send AUDIO TRIGGERS instead, this guarantees that you don't have that sort of limitation when using AUDIO signals. I believe another way of solving that problem would be to use ETHERNET cable instead of audio cable. Regarding ableton some of the latency can come from "Reduce latency when monitoring" option, generally if you turn that option to OFF or mark the track monitor to "OFF" instead of "IN" and "AUTO" you will get a "tighter" recording that isn't influenced by the latency plugins / DAW is creating. This issue overall is very complex and confusing, so it's very hard to be really fix it. Find something that works for you with your setup and stick to it.. That's definitely a tough problem to solve..
one fix is to use a hardware device as master clock sync everything to that and also use it to clock DAW. Therefore all external devices are in time and DAW plays in time with them. Then track the output of your external devices all at the same time.
Akai did it themselves too with the MPC software version. After owning a few MPC's over the years ( from the 60 to the 2000XL ) I bought an MPC studio black few years ago, took it out of the box, tested the swing, put it back in the box and sent it straight back. Awful.
I was hoping you’d also compare using the Elektron as the master for comparison. I guess I can always test and compare on my rig with my various devices.
Try making ableton slave over midi-usb(straight to pc usb, not usb hub) . Use the main drum machine with a tight clock as master to sync rest of gear with midi din thru a midi hub. My setup is master Drumbrute-midi hub - slaves ableton(midi USB) , rest of synths(midi din). Ableton link is a nice feature if you have some old ipads ✊
A couple tips that I learned long ago back in my hardware synth days that might ease your mind (or drive you crazy) but at least you might know more about why this is how it is:
1. MIDI timeclock is not sample accurate, and will never be sample accurate from a device just due to the limitations of the time-code standard and the individual inaccuracies of the internal clock of the device you're using (drum machine, sequencer etc...). This is not something you can fix, and this is why most producers usually sampled their sounds into a single sequencer or program for timing accuracy. You'll notice this when DJ mixing old-school dance music too, this isn't just a tape issue, sample accuracy is just that imperfect to begin with out-of-the-box.
2. MIDI events are also inaccurate. Because of the fundamental way midi sends the data, you simply can't get all the data to arrive all at the same time. Basically even if it's seemingly simultaneous, there's several events happening in quick succession and therefor will never be sample-accurate due to this without having each device have some sort of buffer built-in that is sample accurately matched to the same clock (which again will never be sample accurate 100%)
3. There's 2 main ways devices sync. MIDI Time-clock and MIDI Timecode. One sends out a sync message (usually about 16th or 24th divisions) and so all the space in between those times can easily fall out of sync when going against the perfection of sample accurate DAWs. Timecode is based on FPS in relation to tape sync and uses reference points to sync the clocks of two devices, and any time they drift in anyway will result in the imperfection you're seeing in your DAW.
4. Additionally on top of MIDI imperfection, you're also dealing with audio jitter at all points of the AD/DA chain, so that's also another point of inconsistency. So all that combined together means that timeclock syncing different device clocks in by its very nature never going to be sample-accurate for the foreseeable future. This is also why often times going in/out of the digital realm when mixing/mastering you'll get timing variations and imperfections that are not sample/phase accurate because of the delays in all the electronics and the digital-to-analogue conversions going on. If you're all in-the-box you wont notice this because you're hearing the same imperfections equally at the same time and rendering is all software based so the sample accuracy is uniform without live processing jitter.
Ultimately for synth parts and stuff it's not a huge issue and provides some nice imperfection, but for drum machines, it's usually a sample-back-in or make loops from devices on their own clock and fine-tune and import grooves to get perfection. Brian Transeau has talked about this at length over the years due to his love of outboard gear, and also why most of his stuff has to be sampled in and nudged to perfection to get the results he gets.
Love the channel, stay strong.
Dude this is awesome, thank you man
Dang haven’t heard BT mentioned in a minute!
brilliant explanation, thank you.
Ricky, this is the answer. MIDI clock is especially inaccurate and inconsistent. If you want a fighting chance, you have to record all outputs simultaneously, and if you edit them, move all tracks together to improve phase alignment with Live (or any other DAW) clock/grid. Many of the sequencers in these lovely hardware devices have their own micro timing quirks as well. A Linndrum pattern has very different microtiming compared to an 808, for example. The more you pull the thread on this sweater, the worse the situation will become. I recommend making only single take recordings, using all outputs to capture all tracks, and then shifting these to best align to the DAW you record into.
@@zedd_centauri This is also why peeps back in the day were so in love with the Atari-ST trackers with midi, because it was a single midi trigger source so timing could all go simultanious out and just capture live to tape. Other than the expensive timing boxes you can get are more or less simulating the tape clock signal at high resolution but it's as suited for the type of beat-clock timing that alot of drumachines use (some do it better than others I admit) but yeah. If someone wants the swing, you might record dry in and save that as groove data in ableton (one of its more killer features) so you could apply it sample or loopwise. Kinda takes the fun out of the hardware-based production, but this is how we all had to live with it, even in early all-digital setups.... quite frankly I don't miss it (besides the fun gear though).
gotta love that slightly reassuring feeling of seeing Ricky struggle with tech like a normal person
yes honestly i love the minimal edit aesthetic of it, feels more human and so relatable
@@Yuluco trick is to talk nice to your gear
I am so so happy that someone else sees this stuff happen. I have fought with weird MIDI glitches like this for so long and it bugs me out how nobody else ever seems to talk about it...like their workflow and creativity aren't just totally bombed by stuff like this all the time.
I never synchronize a machine to Live. I type the tempo I want in the hardware, hit record on Live, hit play on the machine a moment later, warp/adjust the beginning of the clip afterwards. No weird sync nonsense, no latency problems.
Bang on 🎯. Similar experience of frustration.
Just a thought, would you yeild similar results reversing the process ie. Slaving Mac-mini & Abelton to the drum machine?
Awesome stuff 💯‼️
Seems like the solution, thanks mate!
Same here. No matter how loose the clock in a piece of hardware is (and some are pretty loose) I've never encountered one that's anywhere near as bad as MIDI over USB (whether it's a USB interface or direct MIDI-over-USB). Some OSes are better than others but it's baked in to the USB protocol (at least up to USB 2, which is still what most hardware is going to use) - data packets over USB are sent at a 1kHz frame rate, which means the absolute tightest possible timing is going to only be accurate to 1ms over USB, plus any timing issues happening in software (usually at the OS level) get rounded to 1ms increments. By comparison, a solid hardware sequencer (MPC2000/1000/etc; Octatrack; many others) generally has jitter under 0.5ms. I've measured the jitter on my own MIDI interface at about 7ms!
For me, the best solution has been clock over audio (originally SMPTE timecode into my 2000xl, more recently an Expert Sleepers USAMO), which is as close to sample accurate as MIDI/SMPTE and your hardware's timing allows.
@@mudsh4rk thank you. 🙏
THIS IS THE WAY
The dots are the individual sample plot points 👊
(Edit - great video!)
Dude how is it that you make videos on exactly what I`ve been thinking about all week.
👀
I know right. Been wondering for ages why it sounds all nice and groovy and then you record everything separately and that groove is somewhat or completely gone. Now just record everything at once in a multitrack which took years to figure out to be fair but certainly worth investing in a mixer and audio interface that can do that.
Hey dude! I've been a long time watcher of your videos - I appreciate how you are constantly digging into the nitty gritty and the unglamorous parts of production - I've learned a lot from you and you've kept me really inspired the past couple of years! Thanks - keep it rocking! 🤘
Btw, just to be a know-it-all... the problem you're facing isn't latency, it's clock jitter.
So I can watch clock jitter on flixzone? No need, I watch latency tests on Netflix with my wife every other day or so
Came here to say jitter
People used to claim Macs were immune to it and vastly superior to PC for that reason. Maybe that was back in the 90s tho.
@@ceounicom It isn't jitter. Windows and Mac OS have never been real time OS's and have no way of guaranteeing events will happen on time. If they were, most consumer software wouldn't work, and they're designed for usability. Also, USB is inherently terrible.
Back in the 90s when you plugged in a MPU401 your MIDI input was connected directly to your CPU by an IRQ line, and PC MIDI had the fastest possible response. It's been downhill since. When using a Sound Blaster or other cheap MIDI card without the IRQ, I had exactly this same problem on my 486. I could even make my system crash by plugging in my MC303 and hitting "play."
Remember that MIDI was invented before most instruments *had* a CPU, and it's been good enough for performance for almost fifty years now. I think a Keyboard Magazine article in the late 80s described the situation as "in the time it takes a MIDI note to be transmitted, sound travels a few inches in air. If this delay is too much for your composition, then you might need to put your instrument inside your head."
Jitter is an entirely different thing, and if you have one millisecond of it, what comes out of your speakers will not sound like music at all.
@@stickyfox I know first hand that these aren’t real-time OSs, at first I thought it was my software, then my audio interface, and finally when I tested my built in mic I knew it was windows. It’s only a few moments off but that’s enough to hi 1/8th of a beat off. Currently trying to find a fix.
Man, I relate to this so much, I just abandoned trying to track everything into a daw. If I choose to produce on MPC, I do everything in the box and record the full track and if I don't feel like it I do everything in a daw. The MPC mix can be dirty and unbalanced due to its limitations but I prefer the swing and feel over the clean sound.
Rick great video, the best way to get everything tight is to have a midi card sitting in your pcie slot on your computer like a RME Hdspe etc. I have the Erm Multiclock, that is maybe the best way to get computer, external gear, sequencer etc to talk to each other and when you go to print audio, you wont do any adjustment accept the beginning. As of right now Erm plugin for Cubase is not functioning correctly, according to them there is an update coming to fix that, hopefully this not the case for other DAWs as it looks like Abelton can be solid with some offsetting, working on Bitwig to see how it works.
Unitor 8 for Emagic(90s)still is pretty solid and Mac is providing USB midi driver support, that is pretty tight. Another company from Germany called Miditech 16x16 USB/MIdi has slightly better results than the Edirol. There are also loads of folks using Roland SBX-1 doing all the sequencing dawless and then printing from DAW clock or just Rolands clock and trimming the downbeat. I just cant believe that in 2021 we are talking about midi not being what it used to be. Also if you can please make a video about the Erm Multiclock and how to use it as USB/interface, still have not figured out all that potential and that manual needs to be thicker. There are some folks using the ERM plugin from daw feeding something like Kenton Midi through and getting great results when printing the audio.
If you have been creating a track all Dawless and love everything about it, and start printing the audio into the DAW you are going to loose that vibe, there is nothing like all the gear firing independently, this is why so many of 90s dance/hiphop has so much human feel to it, because lot of folks just record straight from a mixing desk onto DAT or tape machine or adat.
We need to help everyone get this thing sorted, the amount of videos about timing/latency/jitter etc is just crazy. Thank you for all your great work. Cheers.
This is exactly why I love all of your videos. I have ran into things like this with no way to explain it. I actually have a M-AUDIO USB to MiDi adapter and it has never failed me when I needed it!!!!
really enjoying your videos. i‘m facing the same problems at the minute getting my hybrid setup working. makes me feel a bit better to find out that latency and everything being off really is a thing
This wasn't boring at all. This is extremely similar to a rabbit hole I went down on 2-3 separate occasions and it's nice to know others have encountered this.
Thanks for making this video. I’ve started recording a separate 4x4 kick from every machine just so I can make sure each track is tight once it’s in ableton. It’s maddening and I spend more time on this that the actual track.
And this is why after 20 years of cubase and Ableton etc I went back to outboard. That and I’m a computer engineer and sitting in front of a computer in my spare time just sucks my creativity and work flow 😔
I finally went outboard, too. The latency also changes with plugin load, which makes it completely unpredictable. And yeah, fuck screens.
As a programmer, I support this message. Oddly enough computing in my spare time is all good, just not when it comes to music.
Midi over USB for any sequencer kind of thing seems to always be funky. But the model cycles is just pretty bad over midi or usb. The only way I manage to get it perfectly timed even on the first few beats is to hit record in the DAW but have the cycles play an empty pattern. Then launch the pattern you want to record and that extra bar helps to get it all synced in time for the second bar.
I second this. Old trick, but it works.
i don't own any elektron gear, but this is my standard practice. it doesn't matter if i'm sending MIDI to DIN devices via mio4 or direct to MPC live via USB - an extra bar at the beginning is mandatory when i'm tracking MIDI devices into ableton.
Bounce and manual adjust. No other solution. To ear your groove is impossible other ways. It's painful.
Can you explain how to have the cycles play an empty pattern?
Or you could record it as is and ignore Live’s grid, keep it loosy goosy
I love going down these rabbit holes with you even when they seemingly lead nowhere lol. I still learned various things and got a kick out of watching you wrestle with this. Hey, maybe your "eureka" moment will come later. Keep these types of videos coming. ✌️
most relatable production video on youtube
This is my favourite channel on RUclips. It's funny cause I never would have thought of this kind of video, if you pitched it to me I would have said it wouldn't attract an audience, but I love every video.
I recently connected and mixed all my hardware on my father's mackie analog mixing console. Although the noise floor was significantly louder, I couldn't believe the difference I felt in playing the gear. The only way to describe it is how immediate and responsive everything was. DAW latency might be just a few ms but when playing through it for a decade you can certainly feel the difference when you move over to electrical signals rather than converted signals.
It was very insightful. You made several basic not talked about issues with syncing and recording from hardware into Hybrid setups. I never considered running the same sound side by side to check the sync. Genius! I always end up snapping to grid....
Hey Ricky, I can't say, how happy I am, that you made this video. For me. it's an confirmation, that I'm not such a great foolish at all... :-) . I've used Logic and now Studio One is my DAW of choice (again on Mac). The timing is way better, but you can't get totally rid of jitters and latency. My workaround is the same, as you explained: 90% of my midi-sequencing I'm doing with my Squarp Pyramid (btw: it's great!); then I'm recording my sessions and songs via an multi-input-interface direct in Studio One (or Capture), set the tempo there manually as it should be and drag the recorded tracks all together in the right grid. That works great for me. My opinion: if you're working with timing-sensitiv hardware like drummachines or grooveboxes and you will grab them into a DAW (without the multiclock...), then that's the way it goes...unfortunatly.
Stay healthy!!
this video is of value ricky bc we do struggle with these issues but without the understanding or equipment to isolate exactly what is causing these hiccups, it helps to see problems like this elsewhere so we are more informed with respect to those random software related issues as well as products that are fails, esp when those products are very expensive tho it may mean little and be unseen by you al together thanx for absorbing the impact of buying an expensive dud, it sux and im sure many will not end up in the same spot bc you did already preesh mane.... tempo issues are very confusing in the beginning. i had a real issue learning that the Orba's 100 BPM is MPC beats 97.7..... it took me a long time to accept there is an odd gap in actual tempo so i could just move on with the recording (i wasted a lot of time thinking i was doing something wrong when i just didnt know software kind sucks sometimes with these issues) i have never mutli tracked into a DAW. im just scratching the surface on software at all and find this info very re assuring. i wish i saw this sooner, PLEASE do a beginners guide to recording tracks including these types of issues that may trip up a noob like myself. the issues that wont be mentioned in the user manual such as this tempo discrepancy with software and hardware. this video will help me harbor additional patience when recording in the future and i think you have enough others to speak on that would help someone like me, thanx ricky... keep it Gnar Gnar
I believe the dots at that zoom level represent each “sample”. I’ve always wondered about sync issues like this especially with the Virus TI2. Thanks for making this video! 🙏🏼
I love this video because this is me right now trying to multi track sequence and record my Syntakt thru over bridge in Ableton. I felt very alone and frustrated. I no longer feel alone.
"just track to tape" LOL
Watching this really makes me appreciate knowing I am not insane, as I did the same experiments a while back before finally cracking and buying an E-RM. I love that you are looking for a cheaper alternative though, I just haven’t found anything besides the E-RM that works. If you want to try a USAMO just for fun, I can send you mine. It didn’t like my RME interface and is just sitting around my studio.
What's up Landon! Thanks for being a member
Good subject. I have ran into this problem lots over the years . I now only use usb after I've arranged a song using a squarp pyramid to sequence then I bring the computer in for live instruments and vox.
I thought it was just me hearing usb slip in timing :)
Years ago while I was just switching to computers from hardware I learned about this issue
And my fix was to use the computer sequencer to trigger my hardware using midi notes
And record the output from my hardware into my DAW
This was the only way I could make sure the clocks were rock solid
I've experienced this too. I use Windows and the DAW mangles midi out..
Don't always record your parts as individual parts. E.g., it can be better to record your drums in 1 stereo mix. You lose the ability to edit the midi and different volumes, but you retain the groove and vibe. When I use the DAW and create drum parts then I always have to fix the audio wave. Timing is much better from the Akai, or from an analog clock. Analog really grooves for locking synthesizers and drum machines together.
Yes. I'm experiencing this midi jitter and for recording drums for metal songs this is horrible. One take for the whole song is the way.
I've been stuck in latency/jitter hell for a while now, also thinking of just tracking to tape and learning the banjo lol... thanks for the relevant content Ricky
You can adjust the latency in your settings . It just need a little bit
@@longlostrobots8146 yes I know! :) It’s the clock jitter that drives me crazy at this point, when recording multiple takes it can create issues... mostly with percussion where I want really tight transients that are consistently in time. Oh well, it is what it is 🤷
Very cool seeing it tested like this. I've always struggled with MIDI over USB.
I’ve been using the Innerclock Sync Lock for years now to sync midi and din-sync hardware to Ableton as clock source. With some clever routing you can make the setup work even with monitoring via Live, allowing adding insert plugins, sends, etc.
Seems like it’s around the same price as the erm and I don’t see any knobs to adjust the clock delay, I thought that was the way to get all your gear in sync
Absolutely not boring dude! I'm actually having lots of trouble to sync my Mother 32 to Logic Pro's clock and recording something in time... I don't know what's happening, I've tried everything too, DIN, USB, but the sequence almost never starts on time and is all over the place. Computers mess tempo up for sure!
I usually skip the ads but Nikkolas Smith is legit. excellent shout out Ricky. 👍 I was in art class with that dude in middle and high school. He was amazing even back then. I was trying to make bongs out of ceramics. 😆
This is my favorite video on the internet
another dope video and so happy to see the coffee in the intro again!
We're all in this together! I thought it was just me spotting this stuff!
Clock from Ableton always jitters. It’s known fact. I don’t recommend using ableton as a master clock
What about Logic? (Asking for, um, a friend...)
been bouncing my tracks of the mpc to logic for years and never had a problem
I've noticed it too. You can hear it drift even without external syncing. It's messed up. The worst is syncing two laptops together, both running Live. It's all over the place. Sloppy mess.
Seriously though, I'm having much better luck running my 2K XL as the master clock into a Motu Express XT and clocking Ableton from that. I can see the timing jitter from the MPC replicated in the Ableton BPM preview. The key was honestly getting a dedicated midi interface for cleaning things up. BUT, when you disengage the MPC as the master clock you lose that organic swing / jitter. So it's now rock solid in Ableton which is bittersweet. If there was a way to record the BPM drifting as automation that would solve the last part of the puzzle for me in regards to maintaining the MPC "groove" in a hybrid setup.
When you went "I got the ERM multiclock... so I went ahead and bought one of these things" my first thought was... "Ah shit, here we go again!". :D
same here! hahah
@@RickyTinez I got the best results with RME and MPC4000. Still needs some tweaking.
try overbrige. take a midi-signal from the digitakt to clock another devise while running overbrige. thats my way to "not buy a multiclock". but if i would have the money.........
I agree. OB is the 2nd best solution. Still not as a good as the E-RM since OB adds a lot of latency and chews up the CPU, but it’s the best alternative I found as well.
thx fot that, just thought i have to overthink my whole recording setup for my EP, but OB, Digitakt and RYTM are my things
I have the same with the 4k but i guess that some machines are more musical with it's jitter than others. Best way to record using multiple outputs of the mpc's so you can mix it. Problem starts again when you want to add things later....
damn this is basically hard proof I need an ERM. Even multitracking into bitwig using the pyramid as the main clock for my synths with no midi sync in the daw will leave me with weird jitter
same here with the pyramid into ableton
Really. I was going to buy a Pyramid thinking it would improve jitter. Guess the money is better spent on the ERM in that department. I have seen the midi data jitter and timings for a lot of devices, the Pyramid is better than average at that, the Elektron stuff is satisfactory. But who can afford a Sequintix Cirklon or even wait for it to be build as they are back ordered by 2 years. Lol
This reminds me a nightmare that's a bit of topic, but related to timing problems and computers. At the time I was working with computer as the heart of my live set-up (with NI Mashine), I did many records with my band recording live with the audio coming from my "live" laptop to my studio desktop in two tracks while recording the other instruments.
Afterwards, I opened Mashine as plugin on the desktop computer and re-recorded all the sounds from Mashine in separated tracks for accurate producing and mixing. No problem with this method but one time. The recording messed up somewhere and the timing between the audio-recorded tracks and ITB Mashine tracks where totally off. It started good enough at the beginning of the tune but turned more and more out of sync. At the middle of the tune, all the groove where lost and after the break, it was just a mess. We tried many things : record again from the laptop, do the stuff in the box in the laptop, slightly change the tempo of the whole project and so on. But but nothing worked. It remained far from the groove of the live recording. It looked like he tempo had slightly changed during the recording.
All we had to do was to keep the electro stuff done and record again the other instruments. It only happened once, but it taught me definitely that computer timing can mess up.
Now I've turned my live set-up dawless and we record live with all the instruments at the same time. No more timing nightmare.
Though you did not provide a solution, I was relieved to find that I am not the only one in this very frustrating predicament. Yea, I guess latency will always be there, we just gotta choose how we work around it...sucks
20:59 YES, always run your audio interface (and of course midi interface) on a separate USB, directly to the computer. Everything else like keyboard, dongle, drive, etc... can run from USB hub. This is definitely a source for this kind of timing issues.
Just avoid everything MIDI and USB in one sentence. It's the worse protocol EVER to use for stuff like this. It will NEVER be perfect.
@@casecold1864 The good old firewire... It feels so good to have my old Ultralite firewire running rock solid... midi from my Ultralite in FW is ultratight... and OP-1 usb-audio-midi code is the best thing you could possibly imagine... boutiques on the other end... nevermind...
I would never buy a usb audio interface. Never.
@@re8et355 FF800 here, it's perfect.
After a few years of struggling with midi jitter out of computers i think the Roland SBX1 has solved the problem.
the ERM is next level though with it's timing shifting...
Thanks for the video Ricky, wasn't that boring :)
This shows the world of hybrid hardware/daw setup is still an utopia nowadays. The best thing to do is either choose between DAW-only or hardware only in your project, or if you want the best of both worlds, go DAW/powerful groovebox and sample some synths when needed, knowing you'll have to manually warp/sync them.
That was a fantastic video.
The nuts and bolts of it all can really stifle creativity. I sometimes get a blast of inspiration but, by the time I've got everything working it's gone.
By the way, does the little chap with his tongue out have a name?
The little chap's name is Goobletone.
Turn off monitoring in Ableton. Timing is never gonna be tight if your monitoring is not set to off in Ableton.
Monitoring needs to be off anyway. But it doesn’t solve latency. It is only solving latency in ableton itself.
@@arjanpetersen not really, latency can be compensated. the big issue isn’t about it. the real problem is jitter, which can’t be compensated in any way. you can calculate how many samples of delay you’re having over each buffer size and just compensate. but overtime you’re gonna be off anyway at some point because of the jitter. I simply solve the issue by using the device called USAMO by Expert Sleepers, there are other options but this is probably the cheapest that actually works.
of course that monitoring needs to be off in Ableton, otherwise none of this will ever work.
AFAICT, if you're using the External Instrument device, latency compensation works correctly, even with Monitor enabled.
@@formerself4372 it probably sounds compensated, but it's definitely not sample accurate compensated and definitely will jitter a little bit, making it off the grid sometimes. that's the difference. it's useful because I can simply edit all my jams and everything will be synced from start to finish, i don't have the need to ajust the samples because they were always on time with no fluctuation in time.
You'll drive yourself nuts with this Ricky if you go down the MIDI timing rabbit hole, you'll start to hate the timing of DAW's. Coming from the 90's though, hardware sequencers and the first generation of computer sequencers ( they weren't even DAWS back then ) please be reassured that although this problem still exists it's exponentially better then it was back then. Midi timing from computers and sync ( post Atari ) literally sucked big time back then. For sure. We did managed to make tracks with it though :)
your actually introducing latency by recording with in on its supposed to be strictly used for monitors hence the in terminology the idea with Ableton is you would have two tracks one for monitoring another for recording seedtostage has a good video in it.
I had an email converstion with the creator of Expert Sleepers USAMO - Andrew Ostler,and he said "there are two sides to every
MIDI connection - the transmitter and the receiver. The USAMO gives you a near-perfect transmitter, but it can do nothing about the
behaviour of the receiver." so Your synth,or drum machine have to a good/perfect midi receiver also to get accurate timing,no matter if You send perfect midi signal,I think this is one of the deepest rabbit hole,I end up with other method CV - control voltage, I sequence my Moog Slim Phatty,and Arturia Microbrute with control voltage,but not every device have CV input.
That's the reason we went to recording everything at once using a mixer that has outputs and then going straight into 8 channel interface. Certainly not the cheapest option and if you have a small home studio set up might not be possible but highly recommended if you can do it. Also saves tone of time since you just record everything all at once too.
I don't understand, how does having an audio mixer fix MIDI timing issues?
@@formerself4372 well you record all the instruments at the same time rather than doing multiple takes so there might be midi timing issues but you don’t loose that swing and groove and relationship between the instruments
@@swamppeople ye but adding the mixer in there just mixes the levels, rather than affecting latency haha
Thank you for this! I'm just about to install a usb midi mix bus and this is good food for thought.
You actually have two problems here.
- One is the audio clock issue on basic audio interface, that can be solved with an expensive wordclock; If you’re recording lot of audio, it makes a huge difference; even decent audio interface that are supposed to have internal clock are bad and have timing oscillation and jitter. But not everyone can afford an Apogee or Antelope and such...
- The second issue is actually from USB MIDI interface that is often out of sync with internal clock, it’s often due to poor drivers on mass market midi interface. The more robust midi interface that I’m aware of is the Motu Midi Express; while back it was the only interface that had solid driver that would avoid this kind of issue.
I had to do plenty of test for my work to link computer system (OpenLabs keyboards) with MPCs. We had to spend a LOT of time to develop in-house drivers for the Presonus interface to be rock solid for our system to work. it’s been a while now that I haven’t look into it, but I’m pretty sure the issue is still the same. I doubt any mass market audio/midi interface developer spent time to develop rock solid midi drivers for their interface. Only guys like Motu, Avid and such might have decent solutions... not the kind of gear we find in (bedroom) home-studios.
Since I saw a vid from you for the first time about hardware latency issue's you got me busy diving into it.. So now I want the multiclock too. Guess that's the only device to fix latency for 4 synths at the time.. (should have kept hardware out of the door LOL) Anyway, thanks for all your vids (on this topic) been really helpful.
I have had good luck with the Mio too, it works well with old synths for SYSEX (DX7, D50, POLY 800 for example).
USB to Midi is basically USB to serial and uses similar chips to FTDI USB to serial parts. I did a deep dive a few years ago on this stuff because I was having issues with timing from an application over USB to a (non music) product. Basically USB has a "frame rate" and doesn't process the request as soon as it is received. It will wait up to 1 milli-second to process and send the request. The request can have a lot of data in it, but it is only sent on a 1msec boundary. (I may be paraphrasing here).
Older MIDI implementations in the pre USB days would have had a Serial IC "UART" such as the 16550 style that talked directly to the CPU and interrupted the CPU the instant a message was received. Now a MIDI message must navigate the USB subsystem first which adds an unpredictable amount of latency (Jitter).
The MIDI standard itself for clock is pretty amazing, it is a single byte (0xf8 I think) transmitted to sync devices. This means if you are just looking for sync, you should be able to generate an interrupt per byte, and just look for 0xf8. That is very fast, a few micro-seconds, for modern processors. Compare that to the USB latency of up to 1 milli-second and you start to see where the USB system makes things worse.
MIDI clock is 24 PPQN, so at 100BPM we would see 40 pulses per second, or 1 pulse every 25 milli-seconds. A 1 milli-second jitter is 4% error on the individual 24 PPQN pulses. You wouldn't think you would notice, but it's there! Coupled with sampling latency of the audio, and you have to manually time shift a lot of stuff.
I'm just getting into Ableton, and I'm finding stuff like this really kills my motivation. I'm trying to generate enough talent (i.e. buy enough stand alone gear!) that I could just track to final mix. Or I'll get rid of the clock altogether and get really really into ambient stuff.
My clock setup right now is an actually having my drumbrute's MIDI out hooked into my iConnectivity MioXM, which distributes MIDI Clock to all of my devices as well as my computers over the network (using RTPMidi). It's impressively solid compared to other options and I didn't need to buy any new things!
I have given up trying to track external instruments in Ableton, particularly if I'm using instruments in the DAW as well. Unless I'm 100% in the DAW I get absolute crap. Sounds like I should stump up for an E-RM, I've been using a launchpad pro MK3 as a master sequencer and using Ableton as a glorified tape machine lately just to keep things in time.
the ironic thing is "record to tape and learn the banjo" solution is why I was thinking to move to hardware lol
but now i realize i won´t have a smooth transition from daw to dawless, because right in the midst of the hybrid phase is where most issues happen
so i guess with latency is either best to be daw-exclusive with maybe some midi keyboard _or_ full on dawless
watched again for reference. this may be the "most boring video" like you said, but it is one of the most useful
I had the same syncing issues recording my elektron digitakt via USB/overbridge - things were not lining up consistently and I was getting weird timing hiccups. I found that if I set all my ableton audio inputs to "IN" and turned ON "Reduced Latency While Monitoring" in the options menu (I think ON is the right setting - could be wrong)... then things generally would line up perfectly. Another suggestion I heard someone make about syncing problems is to set your hardware sequencer as the master (ie. Digitakt) and slave ableton to that when you record.
Simple technique here: add a 1 bar 'count in' sequence before your song plays, like how bands do to cue the start of a performance. It can be a simple 4/4 kick pattern or whatever. Then once multitracked in your daw, it becomes much easier to line up your tracks on the 1 as they all share the same count in. Once aligned you can then just trim out that first bar
I always thought this was because of Midi to Control Voltage converters in analogue machines. I don't mind, I kind of like it, but I don't have it as severe as in this video. Never had any problems with missing notes or crazy delay's and it doesn't ruin the swing, some this might be an other problem, like midi jitter.
Great video. The worst issue is trying to keep two kick drums from different sources in sync.... Another possible workflow might be - 1. do your sequencing in the hardware sequencer / drum machine -2. export the midi only to the daw . 3. use Expert Sleepers USAMO, which sends sample accurate midi (that's the claim anyway) out of the daw as audio which the USAMO box converts back to midi.... but I haven't tested this! Another issue with Daws (or Pro Tools at least, I presume Live does it too) is they skip a sample every couple of bars (there's a formula that dictates how often this happens, something like dividing the sample rate by the tempo or the other way around).
Usamo looks like the same thing as ERM but only one MIDI output, and therefore no way to adjust individual clock delays
Don't monitor in the daw. Set monitoring to off. There is always latency on live monitor
It used to be recommended to give an empty bar after midi start to avoid the 1.1.1 jittering between synced units.
Mpcs don't perfectly phase cancel over takes, ever (even on master) I guess everything won't if the cycles doesn't.
It all has to go down at the same time.
That‘s right. ruclips.net/video/PT5mD2Zd7F8/видео.html is about monitoring latency
20:28 We may not be able to hear less than ~10/1000 of a second of latency, but of course as soon as any kind of phase shift from parallel processing comes into play, even a completely undiscerning ear is going to be able to hear differences of just a single sample, eg 1/44100 of a second.
I think a lot of these latency issues come down to how well the device drivers are written as much as the hardware itself. I'm a big fan of RME gear on that front, always absolutely rock solid.
I spent the last 6 years chasing midi sync with a computer after buying the squarp pyramid sequencer to sequence bitwig. Now after years of experimentation and measuring (and basically quitting making music over all these fucked up timing issues) I finally got the recipe right:
- Get a modern soundcard with a midi interface, because this way your signals will be just a bit more "in sync" with each other. Dont expect 20 usb devices to work together in any meaningful way.
- Get a DAW that will run off midi clock instead of being the midi master
- Let your midi master clock be your device (eq. octatrack, squarp pyramid etc)
Now to the more esoteric:
- Avoid plugins that introduce latency, bitwig for example has an indicator on individual channels that lets you know how much milliseconds of delay you get for using certain limiters or reverbs
- Run the lowest latency you can afford, but not so low that your system is always on the brink of collapse, as this will actually increase jitter
***UNCONFIRMED SUPER VOODOO TECHNIQUE**: Run at 96k because ehh..i dont know something about USB packets being handled differently and shit. This one actually brought down the midi jitter to reasonable levels and get a 4/4 Kickdrum pattern within about 1/4096 or 1/2048 which is good enough for me ;)
i am a beginner producer and i want to get rid of the topic latency early on to build a functional workflow around it. ideally, i want to control (via midi clip or keyboard) an external sound source and capture it with the audio in of my interface. i also want to use ableton effects. now i have some ableton drums playing and the external instrument receives midi to play, which i monitor in ableton. now i adjust until it sounds good together. now i want to record the synth to audio and when i start the drum clip and the audio synth clip together, it should sound exactly the same as when i was monitoring before. i can record with ableton effects or i can record the dry signal at my input. the dry signal is faster than the monitored one (with effects). now, i put the same effect chain on the recorded audio. all 3 variations (playing it live, recording the monitored audio with effects and putting effects on the recored dry signal) all sound slightly different... i do not think, i care too much if something has latency, but i need to recreate the exact same timing i adjusted to in the "jam" phase. with or without latency, it needs to sound the same. i do understand this topic gets much more complex when introducing a global clock, but i am already struggling with this basic step...
to me there is a change in the groove of any machine as soon as it's clocked by another one. I also believe that each machine receiving a clock, is constantly trying to follow the clock it receives, and then has very slight changes in its timing, but i don't see it as a problem, to me that's the interest of hardware, that it doesn't play exactly the same way everytime.
Btw I think that not all hardware equipment will start it's clock properly if you don't start recording at 1.1.1 , i mean the beginning of the arrangement view. Happens to me with some drum machines and synth, mainly the older one but some recent synth too.
Also if you wanna record with minimum latency, you have to trigger the "off" button, not the "auto" or "on", that let the sound go through Ableton, then to your speakers. You need to use the sound directly from your audio interface, to the monitors. I proceed that way with an apollo twin and i have 0 latency problems. Of course that means that you are processing your sound outside your daw, or recording, then processing.
Love the channel keep up the nice videos and music 💪🏻
I don't think it's latency when you record the machine directly with no clock and align manually. But it's definitely weird. Could be anything, even the Mac.
As tedious as it is, I appreciate you posting it. So many forum goons that don’t make any music have such awesome clocking and all their problems are solved. This isn’t a solved problem, it’s not gonna BE a solved problem. Our perception really exceeds what technology is capable of, and in ways we’re bumping up against the limits of physics and CPUs when we start fine tuning milliseconds.
That’s the hidden humanization feature that latency provides, only robots hit exactly on the mark.
Not even
Good vid, thanks. To deal with it I just turn off swing in my gear and able the swing in Ableton. I actually find the Groove feature in Ableton really powerful and it also quantizes before applying the swing so for the most part it fixes any issues.
6 min in and that's how I feel every time I deal with anything MIDI related. Pure confusion.
I'd like to see the same test done with the ERM. I mean you say it just works but have you done the same test and confirmed it? It might be possible that it feels like it because it cost 600 or so bucks :D I have no doubt that the device itself does a good job (it's a german product afterall and if we are known for something then it's for being punctual lol) but I'd still like to see the endresults and the grand scheme things. Greetings from Germany
.... i haven't and im afraid to.. haha
@@RickyTinez looking forward to the next video 😉 btw love those clock and sync related videos from your channel. I was afraid I was the only one to be all ocd about this stuff because it is hardly ever discussed on RUclips and I really wonder how all the many many other producers out there tackle the problem. Maybe a future solution to this problem could be a specialized OS which handles audio AND midi prioritized. Whishfull thinking haha
I bought an ERM years ago and it's never given me a single issue. Totally worth the price to never have syncing problems ever again, and the ability to adjust per track timing and swing is just icing on the cake. Granted, it's my master clock with all my outboard gear and Live slaved to it. It works flawlessly.
Is this really necessary, checking the millisecond with the mouse? It’s so ocd. What about the popular mantra, that we should be producing with our ears and not our eyes.. Let’s just take that micro latencies as a humanize effect and be done with it. Don’t check it and zoom it in.🙂 Think about how they recorded back in the days when there was no daw’s and it was recorded live to tape. The bpm fluctuated a few bpm’s across the whole track. And it sounds great.
@@LukezyM given that we are producing electronic music and not recording a band imho it totally is. This is what makes or breaks a groove.
marko polo has a workaround for latency, if I'm not mistaken. to track the song out of an MP without losing swing, create a dummy pattern on the MP with 1 quantized kick on 1 and put it before everything. track your stuff into the daw so that each track starts with that dummy pattern. then take tracks and align those kicks perfectly. this should bring the swing back
Wtf I thougth I was fking something up when I tried to record my two novatiom circuits and a TR-09 separately and got this exact problem. In this particular case is when Roland products like the MC-707 or the MX-1 help a lot. Since they work as audio interfaces you can make them send the main tempo and record multitrack while you are jamming. Hope more manufacturers do this cause' is a live saver.
Hey Ricky... I think it's Elektron Model boxes. I'm running an AMD 3.4gz with 16gb ram. I used to have the Circuit and Model:Samples as my main hardware center pieces. I had both plugged in via USB and the Circuit would be tad off but the same each time, while the M:S would be off and start late every time. I sold both pieces of gear to try and let new energy into the studio, as well as learn sound design instead of relying on presets/patches. I now have the M:C and just reordered the M:S. I Love Ableton and have heard ppl say the Model boxes suck over USB. Maybe someone will figure it out one day. As always, thanks for the video.
There was a take that the model samples didnt start at one. Ive had this problem before with hardware when you stop the transport and when you press play again the machine doesnt start on 1 but instead starts on where it stopped before. Other than that i had this problem with the digitakt before using overbridge, with overbridge everything seems to line up fine every time. Also for the last technique you showed, if your device has metronome build in, i usually turn the metronome on and record some bars of the metronome (you can also just sequence a 4 on the floor sequence as metronome) so you can use it as reference to cut all the audios.
you've expirienced something similiar in overbridge?! Its always been good to me but now maybe i'll double check a few times haha
@@RickyTinez no, what i meant is before starting to use overbridge i went though this torture all the time. with overbridge never noticed anything going out of sync which is amazing!
USB involves multiplex switching and the packets can fall behind or end up ahead of others depending on a number of factors i.e how many devices you are using that share a controller, if you are using a hub. I'm not sure of everything, but there could also be the implementation of USB on the device and clock jitter on the device itself because USB is synchronous. Someone needs to make a PCIE card for just MIDI, which would be fabulous to have. Some sound cards do have a direct MIDI DIN asynchronous connection, but they are often expensive. I always get clock drift when syncing via USB from Ableton to my Beatstep, and then using MIDI DIN to sync everything else. I can't recall if I have tried without anything else on the hub or my motherboards controller, but because of how USB works, I think you are likely to get some amount of drift or possibly jitter from the device.
Ive spent so much time with these types of tests lol. It’s midi jitter via usb that’s the issue (Although polyphonic midi also has a 1ms latency per note) cuz usb is sent in lump sums every ms or more. A good solution is serial midi with an audio-source sending the data. I have had good luck with the expert sleepers esx-8md and ES40 via spdif. The learning curve for that was quite insane for me, but there is truly no jitter once you get it working. I am now trying to sequence everything and record it live one take.
with overbridge, there is a slight latency of notes falling early, but if you adjust the microtiming 1/192 to the right it is tighter
Latency can be accounted for with a timing offset, jitter is the real problem. MIDI over USB causes jitter thanks to things like multiple packets per frame, device contention, and host OS priorities.
I'd be curious to see what kind of timing guarantees different OS class-complaint USB MIDI specs have, if any. Class-compliant USB audio is tuned for accuracy (at least for Apple).
What about record with monitor OFF? I saw a video that doing that kill that minimum latency
Monitor Off should indeed reduce latency !
Hey Ricky. Loving an honest look at your process. Inspiring me to do more of that myself. Cheers.
Usb does matter, I used to have horrible jitter issues on my ensoniq asr X which I fixed by making my roland tr505 the midi master, at the time my setup was the original bass station, a tr505 and the asr x. But although the asr sounded great it was useless as a midi master. These days I use CV gate and sync heavily and only use midi via usb but with these volcas I record directly into a br80 and avoid the daw. Then I chop up those samples into ableton sample packs but I isolate my hardware from my daw and only grab the hardware as samples.
Most of the latency you experience in a DAW is caused by the OS routing audio through the DAW itself. Try switching all your monitoring to off on the channels in Ableton and monitor the audio directly from the your hardware. I do this via RME TotalMix, which is a routing utility that comes with RME interfaces. I'm sure you have shown a similar thing for your Focusrite (?) interface in a previous video. The key is not to monitor ANYTHING through the DAW. It's not a slight against Ableton, you will get the same extra processing and in turn latency using any additional audio routing. Hope this helps!
Appreciate Ricky's search for alternatives, but latency/jitter problems between DAWs and hardwares are the exact thing that products like the USAMO or the E-RM MultiClock were made to correct. They're pricy solutions, but definitely worth the money if you have a hardware sequencer that needs to live in harmony with your DAW.
"..just track to tape, or ... learn to play the banjo.." 😂
you might not be sponsored by ERM but my takeaway was BUY A MULTICLOCK. Also, the dots are the actual sample amplitudes at the sampling frequency, the line represents the audio after it is converted back to analog.
I wish Ableton would have an option to send clock independently of transport.
That would fix the issue of hardware needing to "catch up" to the clock when transport starts.
Anyone knows of any sequencers, DAW or otherwise, that have this feature?
Having clock separate from transport would also remove the need for something like the Multiclock, if you already have a MIDI interface.
Or does the Multiclock provide other features? AFAICT it's "just" a MIDI interface that also generates clock.
Great video, BTW. Exactly the kind of things I've been fighting with.
When using USB there is a minimun jitter that can happen because the USB Driver has limited "resolution" ( i.e. it can only send messages up to a certain speed, after that it can't so there will be a bit of jitter on the clock regarding that with any USB solution ) that's one of the reasons sync through USB will never be perfect, that doesn't mean it's unusable, it's just that there is a minimum jitter when using USB.
That's the reason why some clocks will use AUDIO cable and send AUDIO TRIGGERS instead, this guarantees that you don't have that sort of limitation when using AUDIO signals.
I believe another way of solving that problem would be to use ETHERNET cable instead of audio cable.
Regarding ableton some of the latency can come from "Reduce latency when monitoring" option, generally if you turn that option to OFF or mark the track monitor to "OFF" instead of "IN" and "AUTO" you will get a "tighter" recording that isn't influenced by the latency plugins / DAW is creating.
This issue overall is very complex and confusing, so it's very hard to be really fix it. Find something that works for you with your setup and stick to it.. That's definitely a tough problem to solve..
one fix is to use a hardware device as master clock sync everything to that and also use it to clock DAW. Therefore all external devices are in time and DAW plays in time with them. Then track the output of your external devices all at the same time.
A problem shared is a problem halved it is said. Now I know it’s not just my setup. Definitely not a waste of time 😊
Ableton has always done this with every MPC I’ve had from the 1000 to the 60
Is any other daw better ?
Akai did it themselves too with the MPC software version. After owning a few MPC's over the years ( from the 60 to the 2000XL ) I bought an MPC studio black few years ago, took it out of the box, tested the swing, put it back in the box and sent it straight back. Awful.
I was hoping you’d also compare using the Elektron as the master for comparison. I guess I can always test and compare on my rig with my various devices.
My man finally back on the coffee!!!
I used Cakewalk 3.0 for DOS in the early 90s on and 8MHZ IBM pc with no midi problems lol
Try making ableton slave over midi-usb(straight to pc usb, not usb hub) . Use the main drum machine with a tight clock as master to sync rest of gear with midi din thru a midi hub. My setup is master Drumbrute-midi hub - slaves ableton(midi USB) , rest of synths(midi din). Ableton link is a nice feature if you have some old ipads ✊