TY! I am slowly getting into the groove of focusing on and enjoying the process itself. Results Will be What they may. Having a lot of fun with it at the moment. Very happy to have yours and everyones company on the Journey 🙏👊👊
@@Wintergatan my dude just stop stressing over perfection music is all about variation which will be inevitable with how mechanically complex such a machine will be just make something that will make good music and show us your process but dont over complicate yourself
If you set the baseline timing to the middle rubber band height you could have plus or minus 1.5 instead of 0-3. Meaning the hard notes would be 1.5 milliseconds early and the soft notes 1.5 milliseconds late instead of having the hard notes on time and the soft notes 3 milliseconds late.
Additionally, knowing the delay of any given rubber height, you can then program around it to either remove the delay, or lengthen it if desired. Provided the delay at any given height is consistent, which requires more testing.
This is what I was thinking too, that way it averages out to be tighter overall. A slight alternative to that would be to figure out what hardness would be most frequently used and center it on that instead. In other words mode average instead of mean average. So for example if you have say five hardness brackets like [Softest, Soft, Medium, Hard, Hardest], but Soft is used in 60% of cases, and Hardest is rarely if ever used, then it would make more sense to center it around Soft instead of Medium.
@@nalissolus9213 If you mean layered heights, then whatever machine he ultimately ends up building has to have extra moving parts to swap out the layered extra heights as needed (even simple sliding in/out mechanisms) rather than to move the singular (per-note?) height as needed. Whether that's a better solution or not is another question entirely, but it would definitely require more parts, potentially more engineering to work reliably, and have potentially more points of failure.
The other issue with these rubber band methods is your marble bounce is gonna be all over the place. So you're gonna need a huge funnel to make sure no marbles end up on the floor. Honestly I'd get some precision silicone or rubber sheets from something like mcmaser, and laser cut disks to stick on the vibraphone keys like you did with the latex last time. Yes it damps the notes slightly but it'll give a more consistent bounce than the hand made latex bumper. And you can get it with consistent tolerance so you can set the thickness of pad you want by just ordering the right material. And if you go with silicone, if it's sufficiently well cleaned, and the vibrophone keys are smooth, it'll stick without any adhesive through static interaction. So they could be peeled off and replaced fairly easily. Alternatively a drop of latex on the back or something would also work if you don't want them to ever move or come off unless you peel them off.
I'm in for the peeling idea. I think it is the best unless there's some magic material. Plus, making different pads enables more custom music decisions, making some notes softer, like the higher pitch, and harder in the lower note. And for last, a LASER CENTER FINDER!!! Noice!
Martin, for perspective on 3ms: 12 notes sent via midi to hit at the "same time" will have a 3ms spread due to the timing of midi at 38khz serial. Don't sweat the milliseconds. If you do you will be beaten by the machine again. Recognize when things are good enough and learn to accept them.
9:13, I love this realization! I think the slight variation can be considered a feature and not a bug. But if Martin prefers tighter music, I want him to choose what he thinks is best 👍
This creates an existential question. Is he trying to replace humans with a machine while trying to make the machine more human like? Either way I think he's brilliant and does an excellent job with video production too.
@@sungear It's a giant mechanical process, where errors and shortcomings all add up on each other. Martin found this out in his previous designs, and getting the machines to behave long enough to make a single music video for UA-cam was a battle all on its own. Designing a machine that can be broken down and set up multiple times on a world tour requires an exponentially greater effort in engineering, which is why Martin is going through all these design goals. So having it play tightly in every aspect allows for compounding errors to occur, and not effect the music.
Yeah, as soon as he mentioned the inherent delay to the dampening bands, it instantly made me recall an article I read many years ago about how to make very human-sounding rhythms in a DAW. The key insight was that harder hits tended to come earlier from real performers, and emulating that velocity vs timing relationship in the DAW gives rhythms a more organic sound. I really think it's right to regard this as a bonus --a bit of accidental elegance for the design.
Your original marble machine song is on the playlist for my car and I've been listening to it occasionally over the last month. That machine has lots of weird bits, it's character. The anomalies that pop up near the end can be part of the personality of the machine. Don't fear everything, you are still an artist and not only an engineer
Yes, he certainly proves beautifully that a person can be both; if this man's work isn't one of the best blendings of art and science that I've seen, then I'm having trouble thinking of better examples!
If you know the vibraphone will have a *consistent* delay, you could just program the music with that in mind and then it won't matter. The way you've designed the programming wheel means you could have the inserts for the vibraphone track take care of this for you. For programmed music, it is only inconsistent delays that are a problem.
@@JulianKingman Martin previously talked about the inserts and how they could be used for "control" channels as well as marble drops. I don't think he'd want to try and alter the actual drop release shape because it might affect the reliability of the drop itself. He could however have multiple channels per note with the first could be for the drop timing and additional ones to adjust the sound, such as the "hardness" of the sound.
I suspect it's an open question if "good" music is exactly on theoretic timing. Even if you take as given that you want to control the exact time, it might turn out that controlling it to be "off" by some exact amount is the way to make it sound best. I wonder what it would take to (while setting up arrangements) replace the programming wheel(s) with computer control servos? That would allow fast iteration (with arbitrary timing) while figuring out exactly what arrangement should be used and then that can be programmed into the mechanical system for actual use.
Dude I don't even care about the machine anymore, I am way beyond that point. The only thing I care about is seeing you succeed, that's why every Wednesday for the last 4 years I've waited for your videos, even in failure you don't give up. Don't change man
I really appreciate the "why am I doing this?" question. I suspect you enjoy the challenge and you don't want it to end. I feel like you are optimizing all the parts, maybe more than necessary, out of a joy of prototyping. And the goalposts are always moving because you keep thinking of ways to make it better. At some point you gotta say "it's good enough" and make the thing and call it finished.
I work on making synthesizer music instruments, and I definitely feel more interested in the process of making the instrument than using it to make music.
@@warpigs330 but the original goal of the Marble Machine was to make music, that was the whole goal from the start. We had MM, then MMX, now we are at MM3 and there seems to be too much optimisation and chasing perfection which doesnt exist. Martin needs to ask himself when is it "good enough"?
The feature monster - it hunts quite a few people. It whispers in your mind: "If I can do that, why not the other thing, too?" - And there the cyclus starts
I love the fact that after so long... the marbles are still in play. Your marble machine was part of my childhood and still to this day I am amazed with it.
Your ability to make music whether you are playing it or you have one of the marble machines playing the music you have programmed / created for it is truly awe inspiring.
I really enjoy the music dominated videos like this one. They are a piece of art in itself! Martin's passion for music and his talent combined with the technical aspects of the machine really bring together something unique in these videos that I have not seen anywhere else!
I think the rubber bands are a very cool idea! I don't know if this was ever a question to you, but personally I think it'd best to only use one type (thickness) of rubber band at all times. If you try to control a lot of variables in the sound, it'll just become a huge headache. The volume levels are very cool and easy to implement tough!
Being *able* to use different material is likely useful, but not at the expense of other things. Being able to change during a single song is likely unreasonable. Between songs might be a nice to have (but don't give up other stuff to get it). Between practice runs in the studio; sure. Maybe a good analogy is to treat the different bands like different brands of strings on a guitar: something you expect to play with while figuring out an arrangement, but that's left alone once you get it working.
I'm not a musician or an engineer. I'm a brute with everything I do. For what it's worth I came home depressed, exhausted, and tired but I watched your video and I saw your enthusiasm and it put a smile on my face. I am very thankful for your videos and I hope you can keep working on this amazing project however I understand if it slowly defeating you and I would not judge you if you call it quits. Thank you from the bottom of my heart for all these years of hardwork.
The vibraphone is such an unbelievably gorgeous instrument. I'm a saxophonist and bassist but if I could go back in time and do it all over, I'd pick vibes for sure. The rubber bands were such a terrific idea for getting consistent metal marble hits on the vibe keys! I am curious how to make it work across the entire keybed; I think you would need to prop up the band between each key so the tension is the same on the rubber band for each key. That might hurt the ability to control volume but it should act like the nut and bridge contact points on a guitar, giving you consistent tension, especially considering you'll always be dropping the marble right in the middle of the two contact points. That should also help send the bouncing marbles where you want them to go!
I think it's best to keep the obsession with tightness to the mechanism then the notes themselves. Not even the best ears will notice 3 milliseconds, but 3 milliseconds can turn into 30 in as complicated of a mechanism what you're making. I think it's a great mechanism and perfectly follows the KISS method (Keep It Stupid Simple). Keep doing what you do Martin, see you next Wednesday!
FYI less than 6ms is considered real-time for audio processing. At 7-8 ms you will start to hear "Doubling" and it will sound like two instruments/vocalists at the same time.
I think this is one of my all time favorite projects to watch. I've been watching this channel since the very beginning build of the MM1 and with everything that's happened so far I actually got a little bit emotional since we're back on sound design. I absolutely love that you never gave up! We can do this!
I've been with your channel for the last 5 years. I've seen all of the moments of the grinder, the welder, the "Good enough". For the last 2 years I have never missed a video. I was always subscribed but never commented or liked a single video. This one, somehow hit differently. I'm sitting here emotionally broken. When you said it's personal it resonated a lot. What I'm trying to say is that I've seen you come so far and your videos are just extraordinary. I'm probably not alone who watches them and feels a sort of emotional healing effect, calming. Don't give up on the dream. It's never worth giving up on it. Thank you. Thank you that you share this with all of us.
Videos where you go into these tangents, like creating whole new digital instruments is why I keep coming back. The end product will be cool but far more enjoyable is the process of designing, discovering, building, failing, trying, striving, improvising, and on and on. Loved watching your journey over these years.
These 3D prototypes are so beautiful in their own way. It's just so nice seeing devices that are designed from scratch for their purpose. Surely they don't look as whimsical as the MMX did, this is still soo beautiful in their own way!
I don't think that delays will be an issue, but there are two things that I think you should consider: 1) The different heights and rubber thicknesses will change how the marbles bounce, which could lead to more floor marbles if you don't account for that 2) This adds complexity which means that there are now more ways that the machine could break or fail I think that the rubber pads of the MMX are a better idea, since they are simpler and less likely to cause problems.
Absolutely. Not only direct complexity (X parts to adjust the height thickness), but also indirect (Y parts) for adjusting the marble reception/recycling. And if they should be adjusted individually, you multiply that by the number of channels, and still need to factor in the space needed for the parts in the machine...
Being able to control how the vibraphone notes sound is worthy of some additional complexity, in my opinion. But l hope Martin doesn’t get too far ahead of himself and sticks to his list of requirements.
Biggest issue is getting the same tension on the rubberband across all notes. It will be near impossible. Especially if you need to do it quick during a consert.
@@SuperCuriousFox It'd certainly sound good if he can make it work, but he's planning to add essentially a spring in front of each of the keys and is thinking of not only changing how much each spring compresses (the height above the key), but also the stiffness of the spring (the thickness of the rubber). This will make catching the marbles difficult. I'm sure that there's a way to make it work, but I want to make sure that Martin at least knows that it is something that he'll have to take into account if he decides to go with this. The complexity of the height adjustment mechanism, of any changes that he'll have to make to the marble catching funnels, and of anything that he'd add to tie this all in with the programming wheel would add additional ways that the machine could fail. All of that can be solved, but it would take time and effort to build and test all those possible solutions. If he decides to go ahead with it, it could end up very cool, but it's important for him to know what he's getting into. I think that going with the rubber pads makes more sense since it is almost as good, is already thoroughly tested, and will allow him to focus on other possible issues with bringing the marble machine to life. If it's worth the time and effort is all up to Martin, I just wanted to make sure he's considered what this change will entail.
What was the point of making suoer thin marble gates, if the rubber flaps would take up a ton of space? For this to work, each note would need its own rubberband and fork, because if you stertch one over the whole thing, the middle would be so much looser, than towards the edges. The work cannot be smaller than a certain amount, otherwise, there would not be enough rubber to stretch. We definitely have some space-management issues here.
I think the ability of a person to vary the timing on the soft notes and the loud notes actually is what puts the emotion into music. Keep it up Martin you are an inspiration. Watching your method is amazing.
The person has the ability to vary the timing because they know where the actual beats fall. Thus, the machine should also know precisely where the beats fall.
Might even make the mounting of the band simpler and easier to swap out thicker/thinner bands as it could more likely be just 1 structure instead of 1 for each note
This channel is the reason I keep coming back to youtube, honestly my favourite channel, always so intersting, entertaining, with beautiful music, and I never want this journey to really end, this journey is so much more fulfilling than the end product. I was so sad when the MMX stopped, kept looking for new videos, trying to find out if you would come back, but I am so so so excited to see how you are tackling the new machine, this is really the best channel out there!!
If you use the rubber, you also need to consider that you would have to put a holder between every notes and not just at the end, because the bounce would be deferent then and if you drop 2 marbles at the same time, the rubber would less cushion the fall
Really love the music added into these videos. I can imagine with some of the other, more engineering based videos that its hard to justify playing something but it makes the video that bit more special whenever you do
Martin, your journey has been a huge inspiration to me, and I am grateful for it regardless of the outcome. Watching your videos has changed how I work as a software engineer. We are cheering you on and learning along with you.
I feel like a process like this can become addictive. If you ever manage to completely finish the MM3 (or MM4, MM5 ... MMn), at that point, if I was in your position, I don't think I would be able to stop just there. I wouldn't be surprised at all if you tried to come up with a completely new idea and started a new journey by doing something just like this, but different. On a side note, I'm loving these videos, each day I'm eager to check on UA-cam to see if a new Wintergatan video pops up. Loving this, Martin. Doesn't matter if the project succeeds or fails, the content, documenting, development and reasearch of the entire project has been amazing. Been following since the early days of the MMX and I can't believe how far you already got, simply amazing. Keep it up! you have all our support and love! ♥
After all the years and setbacks, all the prototypes and angle grinding, I STILL BELIEVE WITH MY WHOLE HEART!!! I eagerly await the world tour. Life before death, strength before weakness, journey before destination. I love your music, Martin, and I'm honored to bear witness to your struggle.
The compromise would be to make sure the design is modular enough so the machine can get an upgrade to the rubber bands later if needs be. The pads seem absolutely fine though!
He explicitly mentions downsides with the old method at 3:40 The new is also more future-proof, it's far easier to change, swap and adjust as needed. In the old it's a pain to do any change so you're kinda stuck with whatever it ends up as
Yo Martin, long time follower here. Science and specifics of music are great , but can we for a moment just enjoy how beautiful it is to hear you play the glockenspiel. Like shheees-kebab, the tunes you play with the winter right around the corner, they really put me in a cozy warm place. If you find yourself procrastinating on the glockenspiel, hit the record button and throw out a quick single :)
I've followed the Marble Machin projects quite closely throughout my life, marveling at the mystical sound and wonderous machine that you composed a song for, designed a machine to play, and successfully ran it. I can't wait to see who wins round 3. Rooting for you!!!
I'm so glad that you found a perspective again. Last year I kinda thought this project comes to an end soon, but this machine feels a bit closer with every video and it's going to be a mechanical masterpiece!
I still think the simplicity of the original rubber pads is the ideal compromise. Less variety, but also not just less parts but more importantly less moving parts.
@@CrimsonQuandary I'll be honest I've kind of given up on Martin ever finishing another marble machine. He's gotten too caught in the weeds of perfectionism and "tight music". No musician is as tight as he wants this machine to be.
@@dylan1kenobi That's a really valid point. Rick Beato has spoken a bit about that idea too, specifically quantizing recorded tracks to make them "perfect". Then took some old great tunes, quantized them, and it took the life out of them. It makes more sense when you're making a machine like this that will play multiple instruments. But for sure there's something to be said about striving for perfection and how much is too much.
I've been watching the whole journey from the start of the MMX until now, the marble machine is basically a part of my childhood and I'm very happy to see you stay strong and even working towards a third marble machine. Video by video the project is looking more and more feasible and I hope you'll always be there to inspire us, and hopefully keep bringing the world of mechanical engineering and beautiful music together. Thank you for everything you are doing!
Greetings from Mexico! dude I understand your struggle, I have being following your content for two years now... It amazing! I love it, I don't like to see you fail, is just mesmerizing watching you and all the people that has help, try to solve nonexistent problems (I say this because your machine worked, from the very first version, It just was not perfect). Maybe try and change the way you see this "Martin vs the machine" to "Martin's Machine" it is an extension of you as a musician transiting in this world, if you succeed, It will always be Martin's Machine. Love your content, and I think this is what your doing right now, amazing content. In the mean time, you work on Wintergatan III that I know it will be an amazing machine!
I personally like the rubber pads from the MMX. It’s simpler, and it works. Is there a slight dampening of the notes/reverb? Yes, but it’s not enough for me to make that big of a difference. With the timing, I personally feel like the more perfect it gets, the more robotic it will feel. Of course you don’t want the timing so off it sounds bad, but there should be a “good enough” point. The centrifugal governor is pretty cool though, so idk.
I feel that he needs to watch out introducing more difficulties where things could go wrong. More customization also leads to more complications. Adjustible hights and rubber thicknesses leading to inconsistent delays which is exactly the thing he wants to be perfect. Sometimes you need to accept when something is good enough and focus on the parts that need priority, it might have a few downsides to use the pads, but at least it's simple and had been proven to work.
PollyBonanzas - I agree with you that making the timing "perfect" may make it sound robotic. Do excellent musicians have perfect timing? I think he should do a short study of live musicians he admires and see what he learns. He may find, as he said in this video, that timing varies with dynamics, and he may find that there is also minor random variation that makes the music feel more human and real. Jimmy Diresta says, "If it looks straight, it IS straight." One needn't always bring out the measuring tape and micrometer. Martin should maybe follow this maxim: If it sounds good, it IS good -- with no need for obsessive measurement striving for robotic perfection.
I agree. The rubber bands are very clever and cool, and I even agree that formalizing a tiny delay on softer notes could be viewed as a feature of the machine rather than a bug (although I disagree with you that "robotic" is a bad thing), but they do not belong on this machine. It's scope creep and the mechanical complexity, extra materials and failure points are not worth the gain of a slightly "truer" reverb from the vibraphone vs. simple rubber pads. Martin has an extremely hard time weighing various compromises. He keeps discovering flaws and wants to eliminate *all* of them without realizing that solutions to one flaw often introduces other flaws. He needs to pick a flaw to be happy with, rather than obsess over the perfect solution. And he needs to pick which flaw to accept based on his design guidelines: simple, fewer mechanical parts, serviceable, consistent performance. Or: He could change his design requirements to support doing this! That's fine, too, but he needs to recognize that's what he's doing, and be deliberate about it.
Hi there, MMX restauration team member here. One huge advantage of these rubber bands is that you can easily replace the band if it goes bad. With the pads you need to pour a very specific amount of latex rubber onto the vibraphone bars and let it cure for several days, and even then you can get the thickness wrong if the liquid latex did not have the correct dillution. Storing the bars incorrectly can damage the pads, hitting them with a lot of marbles can hurt the pads, ... Overall this rubber band thing sounds like a good idea.
@@jbert42 Thank you for this added context! I had no idea the rubber pads were that complex. Forgive my ignorance but what requires that? A "lesser" solution would probably require more frequent replacements, but it would also allow for easier swaps. It might even be worth testing something commercially available from 3M, like rubber or even felt pads to see how it sounds.
Martin I really think the manufacturing of these rubber bands for each and every note on the vibraphone is gonna be horrendous. The part count is gonna skyrocket if you decide that each note needs to have manually adjustable bands. It’s gonna take up a significant part of your assembly time, and then at the start of each concert you’ll need someone to “tune” the bands because they’ll have a different amount of tension depending on humidity, temperature, and how many times they’ve been used.
i feel like much more likely/reasonable would be a single layer of rubber that’s stretched above the whole vibraphone and can be easily raised/lowered to adjust volume and tone. Individually adjustable bands would be kind of silly because you wouldn’t be able to meaningfully adjust them live during a performance like you could with a single layer above the whole vibraphone
Wouldn't a single band cause things like chords to alter the trajectory of the bounces? Since the stretch and flex would not be constant when multiple balls hit it at the same time, might also change how hard the sound is with chords vs single notes. Would be cool if he could show how big of an issue the pads are, such a simple and easy solution otherwise
Things can be a failure in one aspect and not in the other. I don't think he sees the MMX as a failure in the aspect of a piece of art, which is probably why he entrusted it to other people he trusted instead of just burning it down.
1:23 I LOVE that sound! I really like the direction the marble machine is going with the noises it makes, and are looking forward to listening to songs that incorporate this sound particularly.
Instead of using different thicknesses of rubber band and different heights, you could use one thickness... one hight... but stretch it to change the amount of resistance the marble experiences.... You could have a pedal or something on the programming wheel to stretch/loosen the band to get the different sounds.
My biggest concern with the rubber banc idea would be rubber bands stretching... over time that will be a new ever changing variable that i think will introduce itself "inconsistently" based on which notes are played more... also how will he space the bands so that he can get one over each note. the more i think about this the more variables i keep thinking og getting added vs just the pads. question i think he will have to answer is more variables or a slightly better sound. imo less variables follows the trend of this marble machine. (also i really couldnt tell a sound difference in the high vs low bands so my ears may not be a good source lol)
Yes, it sounds like the rubber bands would introduce a lot of new parts. And the best part is no part... It also doesn't sound like a deal breaker like the other problems Martin was tackling lately.
@@safetymangames925 that's kinda like worrying about stretched strings on a piano tho, ultimately it's a musical instrument and little fiddly adjustable bits that need maintenance are just something that comes with the territory
this had so much energy coming back but in a different way - go martin go! PS: it was great to meet you and see your struggle but also your inspiration in Rüdesheim - and how you inspire so many other people to develop themselves further and deeper
a heartfelt and moving boxing match is probably the last thing i'd be expecting from the end of a wintergatan video. I love that i can always come back to this channel and see such interesting findings in what is probably the forefront of automaton made music.
This... is absolutely incredible. Me being musically inclined, I'm just amazed that you're able to find breakthroughs in the testing phase. I'm loving this
That was fascinating, to see the science behind your work. I agree that the bare metal impacts sound a bit harsher, but not discordant; the notes are still true. If anything, I think the bare metal gave more of an eerie tone to the notes, which would be nice for a creepier melody. With the bands, the difference is remarkable. It's so much more pleasant to the ear (apart from the hardest hits, but that's more due to a sensory disorder that I have). "Come in!" That gave me a nice giggle. :D And, yes, I imagine that a marble on the toe feels as harsh as a marble on bare metal, lol! And, as I'm typing three months after the video was posted, I'm sure that you've probably progressed nicely by now in your little "personal" battle; but, if not, I'm positive that you'll defeat it! Never give up!
could you use stretch (different materials have different elasticity and therefore stretch more/less with the same hardness) to reduce the amount of height required on the bands? EDIT: huh, I see you explored that, I actually think keeping the "human" timing is a good thing!
It depends on what Martin wants to achieve. If he used different materials, then he may have to swap them out per channel depending on the song. If he uses one material but changes the height only, then it could be altered during the song itself. He could also do a combination of both but that adds complexity which may affect reliability.
The only thing about a 3ms delay is this question: Does it create a delay when the marbles are returned? If there is a delay and the delay propagates through the marble mechanisms, that will throw off the music eventually. However, if you have a means of bringing marbles to a uniform state before they are dropped, then that can compensate for minor delays. Last but certainly not least, I noticed that some marbles were going wide when you dropped them onto the test rubber strip. Some experimentation to figure out if the marbles dropped from the machine fall in an expected pattern after bouncing off the rubber plus vibraphone is a worthy test IMHO.
This felt much longer than 11 minutes. What you do is amazingly interesting. I only now realised that I used to learn the HP theme on xylophone in school. Hälsningar från Berlin
I absolutely love hearing Martin making music. You are so gifted! please bring the band back together because I'm absolutely craving for new music!! Your journey (or quest if you like) in pursuit of mechanical perfection is far more than you've ever intended it to be. You've gained a lot of new skills, met a lot of fantastic new people and most important of all, by sharing this part of your life with all it's ups and downs, you've inspired me to pursue my own dreams. I actually went back to school and became the engineer that i've always wanted to be. Tack så mycket Martin!
You are doing amazing work on this one, just remember that sometimes the perfection is in the imperfections or you could say this adds character and feel. 3ms you will definitely get away with for the softer notes.
I love following along on this incredible journey. It's been 7-8 years of fascination for me, and even though most of this engineering still goes over my head, I can't help but pop in regularly to see what's new.
I'm worried about the cables you use; they are reliable when under tension, but the way some of your mechanisms work, the cables are being PUSHED and those types of cables don't like to be pushed since they buckle a small, but measurable amount, inside of the outer part.
The amount of buckling and ultimately tightness will be directly correlated to the length of the cable - Short cable, less buckling (but still present) long cable, more buckling, even less tightness.
You're right, it's no longer about the music, the marbles, or the machine, and it hasn't been about them for quite some time. They are only the method you are using for the journey, a personal journey of self development, self actualization, and self discovery. We're all on this journey, it's called life. We're watching you take your journey, and your steps sometimes even help us in our own journeys, which is why we'll all spending years watching a man make the perfect marble music machine. Thank you for sharing your journey.
Hey Martin, i just love the Music you make. To me it does not matter if your music is made by a Maschine, a Computer or played by hand. But i love listening to it. So please never give that up!
@@quakxy_dukx You are right. The Machine has to be playing vibraphone. But it is a neat sound for the future. I am very interested in what could be done with it)
I have absolutely nothing to do with music, but I love your videos! Please don't stop working on your next machine! You are extremely smart I'm so impressed with what you do. We all need to hear your next marble machine!!! Thanks stay strong
I love this series, and hope that you can produce a wonderful instrument. I’m only concerned that you are letting perfect get in the way of good. Your audience will not hear milli-second delays. We will listen to your performance because you are a human, and you have built a wonderful instrument!
I would love to second this, but in reality some of us would. It's like the uncanny valley of groove. Lateness in itself is not problematic. But just because it is milliseconds doesn't make it not count. Sorry for nitpicking, but some of Martin's reasoning is based in very real musician experience I think, both in life and with the two earlier machines. Building something that is musical is magical. Building something that is not is very very discouraging.
Excellence is is borne from the accumulation of attention to all of the details. We don’t know what we don’t know. I love seeing the effort being paid toward the testing this time around.
Martin these past couple months: "there is 0.3ms standard deviation, that's too much!" Martin today: "these are 3ms late, but maybe that's ok?" :D Also, very happy to see Martin's procrastination (samples all the notes :D on keyboard) coming back :D Jokes aside, I really hope you're enjoying this journey as much as we are :)
What if the marbles on the vibraphone channel are unique and covered with a thin coating of rubber, or just solid silicone all together? (Metal core to maintain mass)
this could be a thing however the friction coefficient of coated marbles is different and are also subject do a whole different level of wear and tear compared to full steel marbles. then there is the matter of how will the different weights of the marbles affect their movement trough the tracks and the drop timings
@@Sorestlor YEAH but i think some of us thought he was still doing seperate marble channels for the different instraments. either way, if the option was already rejected then I like this solution. otherwise, I feel like isolating the instrament to just have a consistent sound for each note is better than modularity of a band across the vibraphone... shug not a muscician but i feel like a lot could go into making this an extremely complex piece much like a piano
I do hope this machine gets done within the next 20 years, because boy oh boy then Martin can start composing music for it. Those little improvisations on the keyboard or the piano in this episode were alreadu better than most of the music I've ever heard. I do hope that you put out some of those out as full songs, maybe even before this machine is done. Tthank you for all the wonderful entertainment/engineering/music/everything really.
Such an interesting process that you have been going through and noticing. The issues that are within the machine are the same issues that musicians have but over time have changed the issue into 'Feeling'. This feeling then makes the piece there own and allows us to determine the musician of the current piece just by the details.
Here’s what I love when Martín builds a new marble machine: he tweaks some things that were having problems and makes it better. In fact, with each iteration of the marble machine, he tries novel methods to fix these problems just like an engineer would. Martin is my favorite musical engineer… in fact, he’s the ONLY musical engineer on UA-cam that I know of.
Martin: I need to make a simpler machine, so I will actually finish it. Also Martin: Hmm maybe I should add cam disks to the programming wheel to control marble impact velocity. Love it. Please keep going.
In this case, controlling marble impact is about making the instrument more musically expressive, which is very commendable. With that in mind, I seriously cannot understand why some commenters here have complained that this is supposedly "unnecessary complexity". It's not an issue of complexity for the sake of complexity and "ooooooo, what a cool feature, this will attract more viewers". This is an issue of complexity with a real valuable purpose. The inner workings of an acoustic piano are also complex. But it's not complexity for its own sake. There's a huge difference there.
When you demonstrated the very first marble with the rubber sheet I just said 'wow' out loud. Amazing. One forgets how much energy the falling marble has - more than enough to overcome the rubber trampoline in its way!
You’re exactly right!! It’s such a vibe to have the softer notes back on the beat. It would add such a nice human element to the dynamics of the machine.
This journey of yours is helping with with my business projects. I’ve been defeated many times over the few years after striking out on my own. Seeing you reiterate and try again and again gives me hope that I can find my solutions as well. Keep going Martin! 🎉
This whole series is not only about finding solutions to problems, but finding a methodology/system that best produces solutions. This is a great way to think when you are conducting big projects, teams, organizations, etc. I've learned so much with Martin and everyone here, so glad to be part of this! When the world tour comes to Brazil, I'll be first in line to see MM3 in action!
It's a consistent timesync issue so that should theoretically be able to be sorted by the programming wheel. You'd only be out of luck if the issue was inconsistent.
I have been following your journey since the very first marble machine and I just want to say that over the years I've loved watching you engineer your way to victory. Keep going man🎉🎉🎉
this journey is nothing short of one of the most incredible journeys I've ever seen. the lessons learned the trials and tribulations. Your journey to build this machine has been something I look forward to watching consistently and cant wait to see what comes next
Don't get discouraged, man, this journey is better than any final product...
TY! I am slowly getting into the groove of focusing on and enjoying the process itself. Results Will be What they may. Having a lot of fun with it at the moment. Very happy to have yours and everyones company on the Journey 🙏👊👊
@@Wintergatan my dude just stop stressing over perfection music is all about variation which will be inevitable with how mechanically complex such a machine will be just make something that will make good music and show us your process but dont over complicate yourself
I would say being robust and reliable should have much higher priority over the precision. But it's Martin 's machine. 🤷🏻♂️
Change is only direction not destination
@@Wintergatanyeah Martin! Don't lose your marbles!
If you set the baseline timing to the middle rubber band height you could have plus or minus 1.5 instead of 0-3. Meaning the hard notes would be 1.5 milliseconds early and the soft notes 1.5 milliseconds late instead of having the hard notes on time and the soft notes 3 milliseconds late.
Additionally, knowing the delay of any given rubber height, you can then program around it to either remove the delay, or lengthen it if desired. Provided the delay at any given height is consistent, which requires more testing.
This is what I was thinking too, that way it averages out to be tighter overall. A slight alternative to that would be to figure out what hardness would be most frequently used and center it on that instead. In other words mode average instead of mean average. So for example if you have say five hardness brackets like [Softest, Soft, Medium, Hard, Hardest], but Soft is used in 60% of cases, and Hardest is rarely if ever used, then it would make more sense to center it around Soft instead of Medium.
@@Chaogmu make clips that are meant for each rubber thickness and height, that system is modular and adaptable already anyways.
Came here to say the exact same thing ... With the new new programming pins and consistent and predictable timing issue could easily be accounted for.
@@nalissolus9213 If you mean layered heights, then whatever machine he ultimately ends up building has to have extra moving parts to swap out the layered extra heights as needed (even simple sliding in/out mechanisms) rather than to move the singular (per-note?) height as needed. Whether that's a better solution or not is another question entirely, but it would definitely require more parts, potentially more engineering to work reliably, and have potentially more points of failure.
The other issue with these rubber band methods is your marble bounce is gonna be all over the place. So you're gonna need a huge funnel to make sure no marbles end up on the floor. Honestly I'd get some precision silicone or rubber sheets from something like mcmaser, and laser cut disks to stick on the vibraphone keys like you did with the latex last time. Yes it damps the notes slightly but it'll give a more consistent bounce than the hand made latex bumper. And you can get it with consistent tolerance so you can set the thickness of pad you want by just ordering the right material. And if you go with silicone, if it's sufficiently well cleaned, and the vibrophone keys are smooth, it'll stick without any adhesive through static interaction. So they could be peeled off and replaced fairly easily. Alternatively a drop of latex on the back or something would also work if you don't want them to ever move or come off unless you peel them off.
didn't expect to see you on the wintergatan channel but im glad you're enjoying these videos with us :DD
surprise marble comment from the brain boys 🌝
Why not use drum skin?
I wouldn't trust glueless adhesion to stay put when being hammered by marbles. Let's see where his tests take him.
I'm in for the peeling idea. I think it is the best unless there's some magic material. Plus, making different pads enables more custom music decisions, making some notes softer, like the higher pitch, and harder in the lower note. And for last, a LASER CENTER FINDER!!! Noice!
Martin, for perspective on 3ms: 12 notes sent via midi to hit at the "same time" will have a 3ms spread due to the timing of midi at 38khz serial. Don't sweat the milliseconds. If you do you will be beaten by the machine again. Recognize when things are good enough and learn to accept them.
I appreciate his desire.
9:13, I love this realization! I think the slight variation can be considered a feature and not a bug. But if Martin prefers tighter music, I want him to choose what he thinks is best 👍
This creates an existential question.
Is he trying to replace humans with a machine while trying to make the machine more human like?
Either way I think he's brilliant and does an excellent job with video production too.
@@sungear It's a giant mechanical process, where errors and shortcomings all add up on each other. Martin found this out in his previous designs, and getting the machines to behave long enough to make a single music video for UA-cam was a battle all on its own. Designing a machine that can be broken down and set up multiple times on a world tour requires an exponentially greater effort in engineering, which is why Martin is going through all these design goals. So having it play tightly in every aspect allows for compounding errors to occur, and not effect the music.
Yeah, as soon as he mentioned the inherent delay to the dampening bands, it instantly made me recall an article I read many years ago about how to make very human-sounding rhythms in a DAW. The key insight was that harder hits tended to come earlier from real performers, and emulating that velocity vs timing relationship in the DAW gives rhythms a more organic sound. I really think it's right to regard this as a bonus --a bit of accidental elegance for the design.
exactly this is what separates us from the machines we make random micro errors in play :)
if the machine can play tight, you could make it play delayed on purpose for a more human feel, but you cant do it the other way around
Your original marble machine song is on the playlist for my car and I've been listening to it occasionally over the last month. That machine has lots of weird bits, it's character. The anomalies that pop up near the end can be part of the personality of the machine. Don't fear everything, you are still an artist and not only an engineer
Very much this! Otherwise he may as well write midi songs. The whole point of the machine is to be more than the sum of its parts
Yes, he certainly proves beautifully that a person can be both; if this man's work isn't one of the best blendings of art and science that I've seen, then I'm having trouble thinking of better examples!
If you know the vibraphone will have a *consistent* delay, you could just program the music with that in mind and then it won't matter. The way you've designed the programming wheel means you could have the inserts for the vibraphone track take care of this for you. For programmed music, it is only inconsistent delays that are a problem.
He would need to define "consistent."
The shape of the note release could control both the timing and dynamics of both the release and the rubber band
@@JulianKingman Martin previously talked about the inserts and how they could be used for "control" channels as well as marble drops. I don't think he'd want to try and alter the actual drop release shape because it might affect the reliability of the drop itself. He could however have multiple channels per note with the first could be for the drop timing and additional ones to adjust the sound, such as the "hardness" of the sound.
I suspect it's an open question if "good" music is exactly on theoretic timing. Even if you take as given that you want to control the exact time, it might turn out that controlling it to be "off" by some exact amount is the way to make it sound best.
I wonder what it would take to (while setting up arrangements) replace the programming wheel(s) with computer control servos? That would allow fast iteration (with arbitrary timing) while figuring out exactly what arrangement should be used and then that can be programmed into the mechanical system for actual use.
@@HouseBricksDoor187 I'm not sure Martin or the marble cult understands the basic premise of accuracy and precision.
Dude I don't even care about the machine anymore, I am way beyond that point. The only thing I care about is seeing you succeed, that's why every Wednesday for the last 4 years I've waited for your videos, even in failure you don't give up. Don't change man
I really appreciate the "why am I doing this?" question. I suspect you enjoy the challenge and you don't want it to end. I feel like you are optimizing all the parts, maybe more than necessary, out of a joy of prototyping. And the goalposts are always moving because you keep thinking of ways to make it better. At some point you gotta say "it's good enough" and make the thing and call it finished.
I work on making synthesizer music instruments, and I definitely feel more interested in the process of making the instrument than using it to make music.
@@warpigs330 but the original goal of the Marble Machine was to make music, that was the whole goal from the start. We had MM, then MMX, now we are at MM3 and there seems to be too much optimisation and chasing perfection which doesnt exist. Martin needs to ask himself when is it "good enough"?
Destination vs journey. Maybe a bit reductive, but not necessarily wrong.
The feature monster - it hunts quite a few people. It whispers in your mind: "If I can do that, why not the other thing, too?" - And there the cyclus starts
Seriously. The original video has 240m views and is the epitome of a lack of optimization
I love the fact that after so long... the marbles are still in play. Your marble machine was part of my childhood and still to this day I am amazed with it.
Got a little news for ya bud- If the first Marble Machine was part of your childhood, you are definitely still a kid. 😉
Well, its been 7 years
I feel ya. It's just, I'm turning 51 in a few days. 😄😉@@anatollegros3454
IT IS WEDNESDAY MY DUDES!!!! So happy to see Martin still plugging away!
aaaaaaAaaAAAaaaAAAAAAAAAAAAAAAAHHHHHH!!!
🐸🐸
Your ability to make music whether you are playing it or you have one of the marble machines playing the music you have programmed / created for it is truly awe inspiring.
It always amaze me how he play so good every instrument that he touch
I really enjoy the music dominated videos like this one. They are a piece of art in itself! Martin's passion for music and his talent combined with the technical aspects of the machine really bring together something unique in these videos that I have not seen anywhere else!
@@europhil2000yes. Like on his old videos
0:52 Upon hearing that sound, I let out a reflexive, very impressed *"Oooooh, shit!"*
I think the rubber bands are a very cool idea! I don't know if this was ever a question to you, but personally I think it'd best to only use one type (thickness) of rubber band at all times. If you try to control a lot of variables in the sound, it'll just become a huge headache. The volume levels are very cool and easy to implement tough!
Yes, smells a bit like feature creep.
@@d00dieb0x Martin even tried to introduce different types of marbles once during MM3. Thankfully it didn't go far to my knowledge.
Being *able* to use different material is likely useful, but not at the expense of other things. Being able to change during a single song is likely unreasonable. Between songs might be a nice to have (but don't give up other stuff to get it). Between practice runs in the studio; sure. Maybe a good analogy is to treat the different bands like different brands of strings on a guitar: something you expect to play with while figuring out an arrangement, but that's left alone once you get it working.
@@benjaminshropshire2900the other way to see it would be as different guitar styles. Depending on how the whole thing is set up.
Came here to say this. It would be a beautiful feature to have but seems like a lot of extra complexity. Best part is no part.
I'm not a musician or an engineer. I'm a brute with everything I do. For what it's worth I came home depressed, exhausted, and tired but I watched your video and I saw your enthusiasm and it put a smile on my face. I am very thankful for your videos and I hope you can keep working on this amazing project however I understand if it slowly defeating you and I would not judge you if you call it quits. Thank you from the bottom of my heart for all these years of hardwork.
The vibraphone is such an unbelievably gorgeous instrument. I'm a saxophonist and bassist but if I could go back in time and do it all over, I'd pick vibes for sure.
The rubber bands were such a terrific idea for getting consistent metal marble hits on the vibe keys! I am curious how to make it work across the entire keybed; I think you would need to prop up the band between each key so the tension is the same on the rubber band for each key. That might hurt the ability to control volume but it should act like the nut and bridge contact points on a guitar, giving you consistent tension, especially considering you'll always be dropping the marble right in the middle of the two contact points. That should also help send the bouncing marbles where you want them to go!
I think it's best to keep the obsession with tightness to the mechanism then the notes themselves. Not even the best ears will notice 3 milliseconds, but 3 milliseconds can turn into 30 in as complicated of a mechanism what you're making. I think it's a great mechanism and perfectly follows the KISS method (Keep It Stupid Simple). Keep doing what you do Martin, see you next Wednesday!
FYI less than 6ms is considered real-time for audio processing. At 7-8 ms you will start to hear "Doubling" and it will sound like two instruments/vocalists at the same time.
You got the two S words backwards in the abbreviation. "Simple" comes first.
I think this is one of my all time favorite projects to watch. I've been watching this channel since the very beginning build of the MM1 and with everything that's happened so far I actually got a little bit emotional since we're back on sound design. I absolutely love that you never gave up! We can do this!
I've been with your channel for the last 5 years. I've seen all of the moments of the grinder, the welder, the "Good enough". For the last 2 years I have never missed a video. I was always subscribed but never commented or liked a single video. This one, somehow hit differently. I'm sitting here emotionally broken. When you said it's personal it resonated a lot. What I'm trying to say is that I've seen you come so far and your videos are just extraordinary. I'm probably not alone who watches them and feels a sort of emotional healing effect, calming. Don't give up on the dream. It's never worth giving up on it. Thank you. Thank you that you share this with all of us.
+.
Videos where you go into these tangents, like creating whole new digital instruments is why I keep coming back. The end product will be cool but far more enjoyable is the process of designing, discovering, building, failing, trying, striving, improvising, and on and on. Loved watching your journey over these years.
These 3D prototypes are so beautiful in their own way. It's just so nice seeing devices that are designed from scratch for their purpose.
Surely they don't look as whimsical as the MMX did, this is still soo beautiful in their own way!
I don't think that delays will be an issue, but there are two things that I think you should consider:
1) The different heights and rubber thicknesses will change how the marbles bounce, which could lead to more floor marbles if you don't account for that
2) This adds complexity which means that there are now more ways that the machine could break or fail
I think that the rubber pads of the MMX are a better idea, since they are simpler and less likely to cause problems.
Absolutely. Not only direct complexity (X parts to adjust the height thickness), but also indirect (Y parts) for adjusting the marble reception/recycling. And if they should be adjusted individually, you multiply that by the number of channels, and still need to factor in the space needed for the parts in the machine...
Being able to control how the vibraphone notes sound is worthy of some additional complexity, in my opinion. But l hope Martin doesn’t get too far ahead of himself and sticks to his list of requirements.
Biggest issue is getting the same tension on the rubberband across all notes. It will be near impossible. Especially if you need to do it quick during a consert.
@@SuperCuriousFox It'd certainly sound good if he can make it work, but he's planning to add essentially a spring in front of each of the keys and is thinking of not only changing how much each spring compresses (the height above the key), but also the stiffness of the spring (the thickness of the rubber). This will make catching the marbles difficult. I'm sure that there's a way to make it work, but I want to make sure that Martin at least knows that it is something that he'll have to take into account if he decides to go with this.
The complexity of the height adjustment mechanism, of any changes that he'll have to make to the marble catching funnels, and of anything that he'd add to tie this all in with the programming wheel would add additional ways that the machine could fail. All of that can be solved, but it would take time and effort to build and test all those possible solutions. If he decides to go ahead with it, it could end up very cool, but it's important for him to know what he's getting into.
I think that going with the rubber pads makes more sense since it is almost as good, is already thoroughly tested, and will allow him to focus on other possible issues with bringing the marble machine to life. If it's worth the time and effort is all up to Martin, I just wanted to make sure he's considered what this change will entail.
What was the point of making suoer thin marble gates, if the rubber flaps would take up a ton of space? For this to work, each note would need its own rubberband and fork, because if you stertch one over the whole thing, the middle would be so much looser, than towards the edges. The work cannot be smaller than a certain amount, otherwise, there would not be enough rubber to stretch. We definitely have some space-management issues here.
Darth is a genius too! Those adjustable rubberbands make an incredible difference and as with everything you do, it's tunable... respect! 🤗
So glad I subscribed to this channel after the first marble machine, watching this engineering journey is incredible
I 100% agree
I think the ability of a person to vary the timing on the soft notes and the loud notes actually is what puts the emotion into music.
Keep it up Martin you are an inspiration. Watching your method is amazing.
That is a gross oversimplification
The person has the ability to vary the timing because they know where the actual beats fall. Thus, the machine should also know precisely where the beats fall.
Keep the rubberband at the same distance from the marble drop and raise the entire vibraphone.
Then hard notes will be slightly earlier! 🎉
Yes this is what I was thinking, and you can bring both closer if you want it to land exactly the same time
Might even make the mounting of the band simpler and easier to swap out thicker/thinner bands as it could more likely be just 1 structure instead of 1 for each note
@@joepatbob yeah, I got the impression that the single note band was just for testing and he is planning on 1 band for the entire vibraphone
or move the drop point with the rubber band
@@luketurner314but then you have different resistances for the center and edges of the band
This channel is the reason I keep coming back to youtube, honestly my favourite channel, always so intersting, entertaining, with beautiful music, and I never want this journey to really end, this journey is so much more fulfilling than the end product. I was so sad when the MMX stopped, kept looking for new videos, trying to find out if you would come back, but I am so so so excited to see how you are tackling the new machine, this is really the best channel out there!!
If you use the rubber, you also need to consider that you would have to put a holder between every notes and not just at the end, because the bounce would be deferent then and if you drop 2 marbles at the same time, the rubber would less cushion the fall
I imagine he could just change the orientation of the rubber bands to be in line with the keys, and have an individual rubber band for each one.
Or he could put the clamps on either end and attach 'support' fingers between each note. Essentially isolating the bands around each note.
But how does he get the same tension across all notes?
Do you really need to change softness? Isn’t that just adding a ton of complexity?
Really love the music added into these videos. I can imagine with some of the other, more engineering based videos that its hard to justify playing something but it makes the video that bit more special whenever you do
ayo tride
Martin, your journey has been a huge inspiration to me, and I am grateful for it regardless of the outcome. Watching your videos has changed how I work as a software engineer.
We are cheering you on and learning along with you.
I feel like a process like this can become addictive. If you ever manage to completely finish the MM3 (or MM4, MM5 ... MMn), at that point, if I was in your position, I don't think I would be able to stop just there. I wouldn't be surprised at all if you tried to come up with a completely new idea and started a new journey by doing something just like this, but different.
On a side note, I'm loving these videos, each day I'm eager to check on UA-cam to see if a new Wintergatan video pops up.
Loving this, Martin. Doesn't matter if the project succeeds or fails, the content, documenting, development and reasearch of the entire project has been amazing. Been following since the early days of the MMX and I can't believe how far you already got, simply amazing.
Keep it up! you have all our support and love! ♥
What a cool process! Also loved the sound of that Valse d'Amelie clip 😍
oh hey you watch martin as well, i just saw your 111 instruments video yesterday lol
After all the years and setbacks, all the prototypes and angle grinding, I STILL BELIEVE WITH MY WHOLE HEART!!!
I eagerly await the world tour.
Life before death, strength before weakness, journey before destination.
I love your music, Martin, and I'm honored to bear witness to your struggle.
Your old solution is extremely simple and works well enough. Adding all this complexity with variable heights seems like unnecessary scope creep.
Design requirements are dumb
Not to mention it’s adding rubber still, which will dry out and has a higher chance of breaking
The compromise would be to make sure the design is modular enough so the machine can get an upgrade to the rubber bands later if needs be. The pads seem absolutely fine though!
He explicitly mentions downsides with the old method at 3:40
The new is also more future-proof, it's far easier to change, swap and adjust as needed. In the old it's a pain to do any change so you're kinda stuck with whatever it ends up as
@@Koushakur yep, the rubber bands add features and (maybe?) sound slightly better but is it worth the extra complexity?
That is amazing!!
Simply incredible sound and those 3ms truly are nothing compared to the beautiiful mallet variation
Yo Martin, long time follower here. Science and specifics of music are great , but can we for a moment just enjoy how beautiful it is to hear you play the glockenspiel. Like shheees-kebab, the tunes you play with the winter right around the corner, they really put me in a cozy warm place. If you find yourself procrastinating on the glockenspiel, hit the record button and throw out a quick single :)
Can't spell Martin, without 'art'
"I thought the difference would be bigger" *Sick Xylophone Solo*
It’s amazing how much this also explains why piano hammers and keys are designed the way they are.
Is it against the rules to use a hammer with a marble head? lol
I've followed the Marble Machin projects quite closely throughout my life, marveling at the mystical sound and wonderous machine that you composed a song for, designed a machine to play, and successfully ran it.
I can't wait to see who wins round 3.
Rooting for you!!!
I'm so glad that you found a perspective again. Last year I kinda thought this project comes to an end soon, but this machine feels a bit closer with every video and it's going to be a mechanical masterpiece!
I still think the simplicity of the original rubber pads is the ideal compromise. Less variety, but also not just less parts but more importantly less moving parts.
Right? I saw the title and thought "Didn't he solve this problem already??"
For real. I imagine the original solution also has a more consistent bounce than the bands.
@@CrimsonQuandary I'll be honest I've kind of given up on Martin ever finishing another marble machine. He's gotten too caught in the weeds of perfectionism and "tight music". No musician is as tight as he wants this machine to be.
@@dylan1kenobi That's a really valid point.
Rick Beato has spoken a bit about that idea too, specifically quantizing recorded tracks to make them "perfect". Then took some old great tunes, quantized them, and it took the life out of them.
It makes more sense when you're making a machine like this that will play multiple instruments. But for sure there's something to be said about striving for perfection and how much is too much.
I've been watching the whole journey from the start of the MMX until now, the marble machine is basically a part of my childhood and I'm very happy to see you stay strong and even working towards a third marble machine. Video by video the project is looking more and more feasible and I hope you'll always be there to inspire us, and hopefully keep bringing the world of mechanical engineering and beautiful music together.
Thank you for everything you are doing!
Greetings from Mexico! dude I understand your struggle, I have being following your content for two years now... It amazing! I love it, I don't like to see you fail, is just mesmerizing watching you and all the people that has help, try to solve nonexistent problems (I say this because your machine worked, from the very first version, It just was not perfect). Maybe try and change the way you see this "Martin vs the machine" to "Martin's Machine" it is an extension of you as a musician transiting in this world, if you succeed, It will always be Martin's Machine. Love your content, and I think this is what your doing right now, amazing content. In the mean time, you work on Wintergatan III that I know it will be an amazing machine!
I watch this channel primarily for the engineering and I forget occasionally that you're also a very accomplished musician.
I've loved watching how far this has come and all the steps behind it all. Keep it up, loving every bit that comes of this.
2:08 Almost a copyright strike! LOL well played sir!
holy shit, i always love this channel, both for the music and how entracing the proccess is
I personally like the rubber pads from the MMX. It’s simpler, and it works. Is there a slight dampening of the notes/reverb? Yes, but it’s not enough for me to make that big of a difference.
With the timing, I personally feel like the more perfect it gets, the more robotic it will feel. Of course you don’t want the timing so off it sounds bad, but there should be a “good enough” point. The centrifugal governor is pretty cool though, so idk.
I feel that he needs to watch out introducing more difficulties where things could go wrong.
More customization also leads to more complications.
Adjustible hights and rubber thicknesses leading to inconsistent delays which is exactly the thing he wants to be perfect.
Sometimes you need to accept when something is good enough and focus on the parts that need priority, it might have a few downsides to use the pads, but at least it's simple and had been proven to work.
PollyBonanzas - I agree with you that making the timing "perfect" may make it sound robotic. Do excellent musicians have perfect timing? I think he should do a short study of live musicians he admires and see what he learns. He may find, as he said in this video, that timing varies with dynamics, and he may find that there is also minor random variation that makes the music feel more human and real.
Jimmy Diresta says, "If it looks straight, it IS straight." One needn't always bring out the measuring tape and micrometer. Martin should maybe follow this maxim: If it sounds good, it IS good -- with no need for obsessive measurement striving for robotic perfection.
I agree. The rubber bands are very clever and cool, and I even agree that formalizing a tiny delay on softer notes could be viewed as a feature of the machine rather than a bug (although I disagree with you that "robotic" is a bad thing), but they do not belong on this machine. It's scope creep and the mechanical complexity, extra materials and failure points are not worth the gain of a slightly "truer" reverb from the vibraphone vs. simple rubber pads.
Martin has an extremely hard time weighing various compromises. He keeps discovering flaws and wants to eliminate *all* of them without realizing that solutions to one flaw often introduces other flaws. He needs to pick a flaw to be happy with, rather than obsess over the perfect solution. And he needs to pick which flaw to accept based on his design guidelines: simple, fewer mechanical parts, serviceable, consistent performance.
Or: He could change his design requirements to support doing this! That's fine, too, but he needs to recognize that's what he's doing, and be deliberate about it.
Hi there, MMX restauration team member here.
One huge advantage of these rubber bands is that you can easily replace the band if it goes bad. With the pads you need to pour a very specific amount of latex rubber onto the vibraphone bars and let it cure for several days, and even then you can get the thickness wrong if the liquid latex did not have the correct dillution.
Storing the bars incorrectly can damage the pads, hitting them with a lot of marbles can hurt the pads, ... Overall this rubber band thing sounds like a good idea.
@@jbert42 Thank you for this added context! I had no idea the rubber pads were that complex. Forgive my ignorance but what requires that? A "lesser" solution would probably require more frequent replacements, but it would also allow for easier swaps. It might even be worth testing something commercially available from 3M, like rubber or even felt pads to see how it sounds.
Your journey on this project is an inspiration, as well as educational, and entertaining.
Martin I really think the manufacturing of these rubber bands for each and every note on the vibraphone is gonna be horrendous. The part count is gonna skyrocket if you decide that each note needs to have manually adjustable bands. It’s gonna take up a significant part of your assembly time, and then at the start of each concert you’ll need someone to “tune” the bands because they’ll have a different amount of tension depending on humidity, temperature, and how many times they’ve been used.
Nah, starting point is just a straight band, he just had to slice up some therabands.
Nothing complicated here.
i feel like much more likely/reasonable would be a single layer of rubber that’s stretched above the whole vibraphone and can be easily raised/lowered to adjust volume and tone. Individually adjustable bands would be kind of silly because you wouldn’t be able to meaningfully adjust them live during a performance like you could with a single layer above the whole vibraphone
Wouldn't a single band cause things like chords to alter the trajectory of the bounces? Since the stretch and flex would not be constant when multiple balls hit it at the same time, might also change how hard the sound is with chords vs single notes. Would be cool if he could show how big of an issue the pads are, such a simple and easy solution otherwise
I don't think the bare marbles sound bad, it's just a different instrument. I could imagine it for certain use cases.
If you put them together, it sounds like a comedic skeleton character just entered the scene and that's their theme.
@@AnUnfortunatePotatoabsolutely
Even with all its problems and shortcomings, the marble machine X was a beautiful piece of art. You should be proud of it and not see it as a failure.
They're not mutually exclusive.
@@thomasbecker9676 You're missing the point. It's about mentality and perspective.
Things can be a failure in one aspect and not in the other. I don't think he sees the MMX as a failure in the aspect of a piece of art, which is probably why he entrusted it to other people he trusted instead of just burning it down.
@@Squantexactly and failures are not bad things.
@@Squant But it was a failure. You can still be proud that you tried, even if you ultimately fail.
1:23 I LOVE that sound! I really like the direction the marble machine is going with the noises it makes, and are looking forward to listening to songs that incorporate this sound particularly.
Instead of using different thicknesses of rubber band and different heights, you could use one thickness... one hight... but stretch it to change the amount of resistance the marble experiences.... You could have a pedal or something on the programming wheel to stretch/loosen the band to get the different sounds.
4:03 Dagens eko, musiker hittar återigen hemlighet bakom sitt hemmabyggda instrument.
Really love your work man, keep it up👍✨
The rubber pads seemed to work great, its also less variables to deal with. Maybe youll find a good answer with the elastic bands
My biggest concern with the rubber banc idea would be rubber bands stretching... over time that will be a new ever changing variable that i think will introduce itself "inconsistently" based on which notes are played more... also how will he space the bands so that he can get one over each note. the more i think about this the more variables i keep thinking og getting added vs just the pads. question i think he will have to answer is more variables or a slightly better sound.
imo less variables follows the trend of this marble machine. (also i really couldnt tell a sound difference in the high vs low bands so my ears may not be a good source lol)
Yes, it sounds like the rubber bands would introduce a lot of new parts.
And the best part is no part...
It also doesn't sound like a deal breaker like the other problems Martin was tackling lately.
Hard to get away from this I think. Is the idea explored by Martin in this video making the design requirements less dumb? I am not so sure.
@@safetymangames925 that's kinda like worrying about stretched strings on a piano tho, ultimately it's a musical instrument and little fiddly adjustable bits that need maintenance are just something that comes with the territory
@@henlostinky273 That's why avoiding as many unnecessary "fiddly adjustable bits" is important.
this had so much energy coming back but in a different way - go martin go!
PS: it was great to meet you and see your struggle but also your inspiration in Rüdesheim - and how you inspire so many other people to develop themselves further and deeper
I think each rubber thickness will have a different amount of delay because each will slow down the marbles by a different amount
a heartfelt and moving boxing match is probably the last thing i'd be expecting from the end of a wintergatan video. I love that i can always come back to this channel and see such interesting findings in what is probably the forefront of automaton made music.
This... is absolutely incredible. Me being musically inclined, I'm just amazed that you're able to find breakthroughs in the testing phase. I'm loving this
That was fascinating, to see the science behind your work. I agree that the bare metal impacts sound a bit harsher, but not discordant; the notes are still true. If anything, I think the bare metal gave more of an eerie tone to the notes, which would be nice for a creepier melody. With the bands, the difference is remarkable. It's so much more pleasant to the ear (apart from the hardest hits, but that's more due to a sensory disorder that I have).
"Come in!" That gave me a nice giggle. :D And, yes, I imagine that a marble on the toe feels as harsh as a marble on bare metal, lol!
And, as I'm typing three months after the video was posted, I'm sure that you've probably progressed nicely by now in your little "personal" battle; but, if not, I'm positive that you'll defeat it! Never give up!
You are not alone, I sow your comment
could you use stretch (different materials have different elasticity and therefore stretch more/less with the same hardness) to reduce the amount of height required on the bands?
EDIT: huh, I see you explored that, I actually think keeping the "human" timing is a good thing!
It depends on what Martin wants to achieve. If he used different materials, then he may have to swap them out per channel depending on the song. If he uses one material but changes the height only, then it could be altered during the song itself. He could also do a combination of both but that adds complexity which may affect reliability.
The only thing about a 3ms delay is this question:
Does it create a delay when the marbles are returned? If there is a delay and the delay propagates through the marble mechanisms, that will throw off the music eventually.
However, if you have a means of bringing marbles to a uniform state before they are dropped, then that can compensate for minor delays.
Last but certainly not least, I noticed that some marbles were going wide when you dropped them onto the test rubber strip. Some experimentation to figure out if the marbles dropped from the machine fall in an expected pattern after bouncing off the rubber plus vibraphone is a worthy test IMHO.
Wow, I didn't know I wanted to hear Yann Tiersen music on the marble machine until you played it. More, please.
This felt much longer than 11 minutes. What you do is amazingly interesting.
I only now realised that I used to learn the HP theme on xylophone in school.
Hälsningar från Berlin
This entire video as he is testing the playing all I can hear in my head is "PIPE DREAM" 😂
And now that song is in my head too :D
Guess I'll go and watch Animusic later
@@Lectrikfro You're welcome 😆
I absolutely love hearing Martin making music. You are so gifted! please bring the band back together because I'm absolutely craving for new music!!
Your journey (or quest if you like) in pursuit of mechanical perfection is far more than you've ever intended it to be. You've gained a lot of new skills, met a lot of fantastic new people and most important of all, by sharing this part of your life with all it's ups and downs, you've inspired me to pursue my own dreams. I actually went back to school and became the engineer that i've always wanted to be.
Tack så mycket Martin!
You are doing amazing work on this one, just remember that sometimes the perfection is in the imperfections or you could say this adds character and feel. 3ms you will definitely get away with for the softer notes.
So glad you're still innovating all these years later. Very happy to have followed this journey since the beginning!
I suddenly have the urge to play Star machine on a vibraphone 😅
I love following along on this incredible journey. It's been 7-8 years of fascination for me, and even though most of this engineering still goes over my head, I can't help but pop in regularly to see what's new.
I'm worried about the cables you use; they are reliable when under tension, but the way some of your mechanisms work, the cables are being PUSHED and those types of cables don't like to be pushed since they buckle a small, but measurable amount, inside of the outer part.
The amount of buckling and ultimately tightness will be directly correlated to the length of the cable - Short cable, less buckling (but still present) long cable, more buckling, even less tightness.
@@Vastafari34 It's not just the cable length, but how it's routed and constrained.
yeah! so they are def. not ideal for tightness ! @@thomasbecker9676
This is just a prototype. I'm sure on the final design such considerations will be taken care of by competent engineers.
@@VulpeculaJoy Okay, that made me chuckle.
Thank you for sharing the journey, haved loved your content since mmx, glad to see you picking over the scars
I think you should ditch "adjustable" in favour of "reliable/unchanging"
Smarter requirements!
You're right, it's no longer about the music, the marbles, or the machine, and it hasn't been about them for quite some time. They are only the method you are using for the journey, a personal journey of self development, self actualization, and self discovery. We're all on this journey, it's called life. We're watching you take your journey, and your steps sometimes even help us in our own journeys, which is why we'll all spending years watching a man make the perfect marble music machine. Thank you for sharing your journey.
Hey Martin, i just love the Music you make. To me it does not matter if your music is made by a Maschine, a Computer or played by hand. But i love listening to it. So please never give that up!
When you played us the metal sample its sounded good! Makes me think of a music box that plays on a broken glass)
It was certainly an interesting sound and was rather pleasant to listen to but it sounded nothing like a vibraphone
Sounds like an amazing sample for some ice/snow-themed music
@@quakxy_dukx You are right. The Machine has to be playing vibraphone. But it is a neat sound for the future. I am very interested in what could be done with it)
I have absolutely nothing to do with music, but I love your videos! Please don't stop working on your next machine! You are extremely smart I'm so impressed with what you do. We all need to hear your next marble machine!!! Thanks stay strong
I love this series, and hope that you can produce a wonderful instrument.
I’m only concerned that you are letting perfect get in the way of good. Your audience will not hear milli-second delays.
We will listen to your performance because you are a human, and you have built a wonderful instrument!
I would love to second this, but in reality some of us would. It's like the uncanny valley of groove. Lateness in itself is not problematic. But just because it is milliseconds doesn't make it not count. Sorry for nitpicking, but some of Martin's reasoning is based in very real musician experience I think, both in life and with the two earlier machines. Building something that is musical is magical. Building something that is not is very very discouraging.
Excellence is is borne from the accumulation of attention to all of the details. We don’t know what we don’t know. I love seeing the effort being paid toward the testing this time around.
2:09 I like the panicked stop to avoid copyright infringement
Martin these past couple months: "there is 0.3ms standard deviation, that's too much!" Martin today: "these are 3ms late, but maybe that's ok?" :D Also, very happy to see Martin's procrastination (samples all the notes :D on keyboard) coming back :D Jokes aside, I really hope you're enjoying this journey as much as we are :)
Found this channel recently. LOVING IT. kEEP GOING MAN!
What if the marbles on the vibraphone channel are unique and covered with a thin coating of rubber, or just solid silicone all together? (Metal core to maintain mass)
this could be a thing however the friction coefficient of coated marbles is different and are also subject do a whole different level of wear and tear compared to full steel marbles.
then there is the matter of how will the different weights of the marbles affect their movement trough the tracks and the drop timings
This will be a big mess and likely clog stuff up everywhere. This option was already explored.
Or maybe a metal core with a hard rubber/ urethane/ silicone coating to have the same mass as the other solid steel marbles.
@@backdrafter1988 Remember, he will likely be doing individual marble channels, so they wont mix with other marbles.
@@Sorestlor YEAH but i think some of us thought he was still doing seperate marble channels for the different instraments. either way, if the option was already rejected then I like this solution. otherwise, I feel like isolating the instrament to just have a consistent sound for each note is better than modularity of a band across the vibraphone... shug not a muscician but i feel like a lot could go into making this an extremely complex piece much like a piano
I do hope this machine gets done within the next 20 years, because boy oh boy then Martin can start composing music for it. Those little improvisations on the keyboard or the piano in this episode were alreadu better than most of the music I've ever heard. I do hope that you put out some of those out as full songs, maybe even before this machine is done. Tthank you for all the wonderful entertainment/engineering/music/everything really.
4:08 XD
Such an interesting process that you have been going through and noticing. The issues that are within the machine are the same issues that musicians have but over time have changed the issue into 'Feeling'. This feeling then makes the piece there own and allows us to determine the musician of the current piece just by the details.
You're going to feature-creep volume control into the programming wheel, aren't you. I would say that's an excellent idea.
Here’s what I love when Martín builds a new marble machine: he tweaks some things that were having problems and makes it better. In fact, with each iteration of the marble machine, he tries novel methods to fix these problems just like an engineer would. Martin is my favorite musical engineer… in fact, he’s the ONLY musical engineer on UA-cam that I know of.
Martin: I need to make a simpler machine, so I will actually finish it.
Also Martin: Hmm maybe I should add cam disks to the programming wheel to control marble impact velocity.
Love it. Please keep going.
In this case, controlling marble impact is about making the instrument more musically expressive, which is very commendable. With that in mind, I seriously cannot understand why some commenters here have complained that this is supposedly "unnecessary complexity". It's not an issue of complexity for the sake of complexity and "ooooooo, what a cool feature, this will attract more viewers". This is an issue of complexity with a real valuable purpose. The inner workings of an acoustic piano are also complex. But it's not complexity for its own sake. There's a huge difference there.
When you demonstrated the very first marble with the rubber sheet I just said 'wow' out loud. Amazing. One forgets how much energy the falling marble has - more than enough to overcome the rubber trampoline in its way!
You’re exactly right!! It’s such a vibe to have the softer notes back on the beat. It would add such a nice human element to the dynamics of the machine.
This was a great episode. Thanks. Keep it up, love the journey and it's a very real one.
This journey of yours is helping with with my business projects. I’ve been defeated many times over the few years after striking out on my own. Seeing you reiterate and try again and again gives me hope that I can find my solutions as well. Keep going Martin! 🎉
This whole series is not only about finding solutions to problems, but finding a methodology/system that best produces solutions. This is a great way to think when you are conducting big projects, teams, organizations, etc. I've learned so much with Martin and everyone here, so glad to be part of this! When the world tour comes to Brazil, I'll be first in line to see MM3 in action!
I absolutely freaking love it when you play some short music in your videos. Please keep that up 🥺
It's a consistent timesync issue so that should theoretically be able to be sorted by the programming wheel.
You'd only be out of luck if the issue was inconsistent.
I have been following your journey since the very first marble machine and I just want to say that over the years I've loved watching you engineer your way to victory. Keep going man🎉🎉🎉
I like the way you look back at your journey of the 'finished' Marple machines leading you to essentials and details not only on technical manners.
this journey is nothing short of one of the most incredible journeys I've ever seen. the lessons learned the trials and tribulations. Your journey to build this machine has been something I look forward to watching consistently and cant wait to see what comes next