In case you don't want to read the video's description, here are some links you might want to check: www.native-instruments.com/en/reaktor-community/reaktor-user-library/all/all/all/300659/ - a collection of waveguide synth modules www.osar.fr/notes/waveguides/ - Pierre Cusa's notes on waveguide synthesis with some impressive javascript demos en.wikipedia.org/wiki/Digital_waveguide_synthesis Wikipedia on waveguide synthesis There's also a plugin called S-YXG100plus-PolyVL which is an 8 voice VL synth for Windows, abandonware by now. I could not get it to run on Windows 10, otherwise it would have made it to this video.
Did you try setting the old plugin to emulate an earlier version of windows and set to run as admin? I get the alesis fusion soundfont tools to work in win10 this way
@@mr_floydst I've noticed that too. A lot of synthesizer technicians are also musicians, with backgrounds in higher mathematics. Historically, musicians did mess with their electronics quite a bit…not often in this much depth though. Kind of a nested circle.
As an owner of EX5 I agree that VL is a great technology . Very expressive, particularly with breath controller. I bought Friktion from Reason last year. It has similar vibe to VL. I agree that over the years Yamaha gave up on cool technologies: VL, FDSP, Fseq . EX5 was underpowered but it was not a good excuse to abandon and focus on ROMplers like Motif series
Thanks for watching! I'm waiting for my specimen of this year's "expressive osmose", that super expressive keyboard based on the "eagan matrix" synth. That looks a bit like it will walk in these steps... let's see. :-)
I am an happy owner of 4 vl70m and 3 Vl1m, I program the sounds by myself with some editor and I can say this synth is very underrate, you can create good acoustic sound emulation and new acoustic sounding instrument but you must program it by yourself because the original patches are not rapresentative of it’s real power. Even the few patches from other tihrd parts producers are not very good, you can do better but it’s a very difficoult work and it take a lot of time.
Thanks for watching! You are right, programming the "models" is very much like FM, but with "sharper" results. You can quickly drive the sound into distortion of the "argh my ears" variety. It takes a lot of time and consideration to create a pleasant sound and even more time to find sounds that resemble actual instruments.
The original designer of the VL1, Toshi Kunimoto, give me the original program for program the VL1, but only works on Mac OS 3 if have you some machine could could work for Tun it I will try to share you. I use to work for Yamaha Mexico and meet him in 2006.
Great stuff - was immediately impressed when I first heard the VL1 and then I saw the price :( It's one reason I implemented flute and plucked string models + a tube resonator in my soft synth 'Hyperion Synth' - the next release will also feature MPE support. I would love to see some new hardware versions of the VL series again - the nearest thing I found that is a bit like that is the Sample Modeling VSTs which sound extremely realistic and naturally expressive.
@@mr_floydst SWAM instruments are also a bit VL like. I had a Korg Z1 which had physical modelling, but sold it once I bought the SWAM trumpet VST. Its much better realism wise, which is'nt surprising considering the Korg Z1 was released 25 years ago!
Hardware is still preferred by some users over software. I have VSTs and DAWs and they serve a purpose. At the same time, I have a couple of hardware synthesizers, a workstation digital piano, a digital stage piano, and a MPC. I’d rather use hardware when it’s time to play. Software gets the job done but as a hobbyist, hardware gives me more personal satisfaction and enjoyment. You’re a multi instrumentalist. Must be nice to have those skills.
Yes, you're right! Using a mouse or a touch screen adds a layer of abstraction to the interaction with your instrument. Having a touchable user interface (buttons, knobs, faders and keys) always is superior.
Hey Floyd, I managed to get all of the original floppy disks for the EX5 from Yamaha. There is one internal preset that I love (Love all of them) but it's called "Made in the USA" Beautiful DX7 Electric piano. Anyway, all of the disks are now stored on my computer and I have access to them anytime I need now. Yamaha sold them to me for 8.76 a piece. If you ever need the factory presets, I can send you an image of them.
They were brave enough to develop these technology, but it just wasn't in fashion. People wanted straight forward analog modelling synths with lots of knobs and sliders. Synths such as the Nord Lead, Roland JP8000 and Access Virus were much more successful, and we still see their descendants today.
Yes they really should bring it back. It has so much potential. It really is splendidly dynamic, there aren't words. I am the happy owner of four VL PLG cards, perhaps I should get another two someday.
Thanks for watching! With todays CPU power, you could easily build your own VP-1 knockoff, with unlimited polyphony. But it's the experience of the hardware maker and the attention to detail that makes or breaks such an instrument. There's incredible attention to detail in most of the VL voices. I'd really like to have that in a new instrument.
@@mr_floydst Digital design for a hardware FM synthesizer (tone generator) is like…you can do 96KHz sample rate 24-bit 1,024 operators at some 166MHz. It's not possible to do that on a computer right now-even with a 100GHz 500-core processor, the amount of data contention would be huge, you'd never be able to physically access memory fast enough. Resolving this on a hardware design is a matter of carefully pipelining the thing (1 FM operator = 1 clock cycle) and using separate bits of SRAM (with their own addressing and data buses) so you can access dozens of pieces of information all at once. Even GPUs use a massively parallel architecture with huge, huge cache lines (integrated graphics faces a memory bottleneck). With FPGAs and good design tools (and things like Amaranth to make design much easier), you could do this easily today. In 1990 you could easily do high-polyphony VP-1 kind of thing, but you'd need a dedicated hardware pipeline (like the FM chips Yamaha used to make) instead of a processor running instruction code, FPGA was expensive, and ASIC process is always expensive. CPU power is the wrong way to go for a lot of this stuff, although you're right that this is a relatively trivial computation. Higher polyphony will hit memory bottlenecks in any case.
@@johnmoser3594 Thanks for your insight! Hm. But I did see some JavaScript demos, so I thought the whole process is comparably "cheap"? Also, you can setup up VST plugins routed like the diagram in this video and get some fascinating results (without the attention to detail Yamaha showed in their implementation, though)
@@mr_floydst The process is kind of cheap, yeah. It's a matter of scaling it. Yamaha's FM-X platform has 512 FM operators and so can produce 512 tones (additive channels are 1 tone per operator) with like 128 polyphony (4 ops per channel, even if one channel is 4 ops not modulating each other). Consider the simple recursive delay. You need to store key-on state so you know whether you're starting with excitation or just holding sustain, or if you're in release. Your filter needs its settings (poles, zeroes), likely an IIR filter with 3 poles and 2 zeroes and with 4 delayed samples (in overcomplicated terms, z^-1 boxes)-look up Direct Form or Biquad, second order is most common, b0-b2 and a0-a1 plus the four z^-1 boxes. You need one sample per sample delay, but that's just a ring buffer. You also need your feedback multiplier. In all, for 16-bit audio samples, you will need 8 bytes per filter of z^-1, plus let's call it 10 bytes for 16-bit poles and zeroes, a total 18 bytes or 144 bits each. 8 bits for your feedback decay is probably fine. DRAM like DDR is typically single-port, so the computer can only read/write; internal SRAM in a hardware implementation has simple dual port usually, so you can read and write simultaneously to two separate addresses. In a computer, you'll need to access memory and fetch around 45 bytes of data for these things, and a cache line is 64 bytes, so this is mostly fine. This leaves enough room to also incorporate a good 10 bits of state for your exciter, but that only gives you 10 bits of phase and frequency, so you're going to need to draw two cache lines most likely, in a sequential burst read, which is fast. You calculate one of these by doing each stage of the math one instruction at a time, with the data pulled into registers, manipulated, and then written back to memory. You'll need a significant amount of memory for your feedback delay, and this can't be read at the same time as your channel state, but it doesn't need to be and so being slow here won't matter…on a computer. It doesn't matter on a computer because the CPU will need to sequentially pull in the data and operate on it step by step, then write it back. It needs to write back state as well-the new z^-1 boxes and the samples in your delay ring buffer, along with the updated exciter phase. If you're using any vibrato, there are more calculations to do here to get all of this done, and more data to read. Each step of the way, you're dealing with slow access to cache (even L1 cache can take a few cycles-stuff is copied into the CPU's load-store queue when transferring between L1 and registers), and even slower access to RAM when there's a cache miss (this can be hundreds of cycles). Each CPU instruction may take a few cycles to execute, there may be pipeline stalls, and all manner of other mishap along the way. In a hardware pipeline, the problem is ab it different. Your hardware will have each individual step of the computation lain out, hard-wired, one at a time. At points where you need to read one of these values, you'll have an internal SRAM dedicated to whatever block of information is being read-for example, an SRAM that contains the first filter's configuration, and a *separate* SRAM containing the second filter's configuration. Exactly one stage of the pipeline reads these configurations; the data from this is used, altered, and passed down the pipeline as needed. A later stage will write back to the same SRAM, which is where that simple dual-port thing comes in. Largely, you have to make sure that each individual piece of data that must be read through the process can be read/written exactly one time at exactly one stage of your entire pipeline for the generation of each single sample. Those data will need another write at a different point in time. Your clock rate must be such that you finish all computations with time to spare, and then have a window of time where the external controller can tell the chip to update configurations, such as making a key-on at a certain frequency, while the pipeline is not doing any calculations and so not accessing these internal SRAM devices. The result is that in hardware, if you have a 166MHz (48000*288*12) chip and it takes you 2,000 clock cycles to perform all your computations, you can run at most 3456-2000 = 1,456 polyphony. Call it 1,024 polyphony, plus 432 cycles during which any instructions to update configuration can be applied (which is not enough to update all configuration! Have a look at the limitations on OPL3 register access: you can't change all the registers between each sample, either). In software, the architectural considerations regarding organization of configuration and state data are at the whims of the PC's architecture (or whatever system is executing your code). You aren't going to be able to random-access your way through that much RAM, and you can't simultaneously read and write; running in parallel requires parallel CPU cores, or some really nice pipelining and speculative execution, which is in general pretty good, but not that good. You have a lot of computations to make, and so you're going to need some scratch memory beyond just your registers. You aren't going to produce 1,024 tones at 48000Hz sample rate in 166MHz-not at all. 48,000Hz sample rate with 100 instructions per computation done sequentially, with perfect pipelining and single-cycle memory access (which cannot be parallelized), is 4.9GHz. Instructions include things like load, store, and all the mathematical operations and program state operations in between; a direct form 1 second-order IIR filter is 9 loads, 4 stores, 5 multiplications, and 3 adds. You'll have better luck with this than you will with software FM synthesis, but it's still pretty brutal in software versus hardware. This brings me back to my original point: I talked about 1,024 polyphony, but the figures I gave there also assume you have a sample delay of exactly 1 sample or 0.020ms. If 1ms of delay is allowable, you can actually have a very long pipeline that nevertheless only accesses each of your SRAM banks at one read and one write point in the process. The pipeline might not finish with the first sample until you're well into producing 50 samples, but it'll finish. At a slower clock rate, say 16Mhz (OPL3 was 14MHz), you can do 32- or 64-voice polyphony like this pretty trivially. CPUs today can do a good bit of polyphony on things like FM synthesis and piano physical modeling, but raw processing power has nothing on a well-designed, dedicated hardware pipeline-and that was accessible in 1988.
@@johnmoser3594 Thank you very much for your in-depth explanation, this totally makes sense to me. I wish I could pin this; one of the most detailed and well-argumented and knowledgable comments I've ever gotten on a video. It also happens to be one of the most enlightening comments on the hot topic "why do we need digital synths when we got DAWs". Hm. I could even imagine making this the topic of a video, ... it's really an "aha!" moment for me. :-)
Hi Floyd, enjoyed the video. In particular the simulation of VL using delays on the Peak. Perhaps it might be of interest to you to do a deeper dive on the Peak relative to the intriguing VL simulation technique you demonstrated. Maybe a Peak/VL episode exploring how to leverage the Peaks capabilites to create "basic" EX5 like sounds. Gratefully.
It's a pity more synths don't enable breath control. From my experiments with my Roland AE-30, the envelopes seem to get in the way because the designers couldn't imagine any controller except for a keyboard! 🙄 My Yamaha DX27 and Behringer Neutron are the most compatible synths in my collection. 😎
@@synthrodriguez7518 there's a certain irony in that the reason I turned to the Aerophone is because I'm too impatient to wait for my keyboard skills to improve!
Thanks for this interesting video. It almost looks a bit like physical modelling with all those parameters. And it also sounds very good. I agree it's a bit of a shame that Yamaha only uses FM and Sampling in their newer Workstations (despite them still being great of course).
Thanks for watching! It is physical modelling. And it was groundbreaking in 1994. Everyone talked about this, but the keyboards were so expensive no one could afford it (that changed when Korg released the Prophecy monophonic synth).
Hi Floyd, love your song with the EWI and MIDI guitar. I have two questions about the EWI, what model is it, and did you play a wind instrument before buying your EWI, or did you learn to play ON the EWI?
It's the Akai EWI USB. I learned the recorder as a kid, and then leveled up to tenor flute, as all kids do ;-) With that knowledge, playing the EWI is relatively straightforward.
@@mr_floydst Haha here in the 1980s Australia we had to learn plastic recorder in high school. It had to be out of tune to be authentic, imagine 25-30 out of tune recorders- audio hell! 🔥
Changing the subject, I remember you seemed a bit disheartened last year. I notice that for this video your ratio of views to subscribers is a lot better than what many channels get so you must be doing something right! 👍
Thanks! Well, channel growth was really bad until the end of march, when I did that QY100 video, which for some reason gets recommended to everyone and his parrot, too. ;-) Seeing those numbers grow is without doubt motivating to make and try new stuff.
@@mr_floydst I must say that todays synths can do pretty convincing emulation of acoustic instruments. Got a Jupiter 50 and the SuperNatural stuff is quite nice.
Great !! Do you plan to own Yamaha montage? . Also it has endless possibilities. It will be a great most wanted instrument in the future like sy99. Two of these will be great couple of synth
Hi! If I had a budget and the space for I high end workstation, I'd rather buy the Korg Kronos, as it's more geared towards sound tinkerers (I think). But as I have set up a rule for myself to only spend money from hardware sales and money earned from UA-cam into new instruments, that's just not a thing at this moment. :-)
Thanks for watching! You're right. But Yamaha's VL is limited to two voices (or one voice on EX5), and you can't edit the "models" itself. Putting the software on today's hardware could be interesting, even as a VST plugin. This year, we'll get some new modeling synthesizers (the Osmose and the Omega), let's see how they fare.
@@mr_floydst if so i would like a resynthesis option to superimpose a sample onto the vl parameters , i had fun in the 80s with a program emagic x-alayser wich could read a Midi sampledump into the Atari and transfer them to dx7 parameters , pretty crazy stuf came out
@@cnfuzz While it's not exactly what you're describing in the second part, the sample and VL thing is EX5's "FDSP" engine. It takes a sample and applies that recursive delay which creates the VL sounds to it. I never heard of that x-analyzer. That sounds super interesting, I'll need to research. :-)
I remember VL. It didn't really seem to catch on but it was revolutionary in the day. Do you think the advent of sample libraries had something to do with it?
Hi, thanks for watching! I think cheap storage space (RAM and big hard drives) were the final nail in the coffin. But the Main reason was that the VL instruments were just too expensive and the general audience didnt see the benefit compared to romplers. But CPU performance has increased hundredfold since then..
@@mr_floydst Being able to design your own hybrid instruments was interesting. But isn't VL just software like all digital synths? It wouldn't be so expensive to create one would it?
@@IanWaugh It's all software. There even was a plugin for Windows in the mid 90s which could do 8 voices of VL, but it's abandonware now. I guess you could build your own VL synth using a RasPi and a browser and some knobs, but I'd be willing to pay for a good hardware user interface.
I've recently acquired an EX5R, but don't know much about how to change things, so far. Is there a way to get all of the Br presets to play with breath control (CC2)? I found a Yamaha FAQ which does it temporarily for one patch. I'm looking for a way to do it globally, so they all play that way. I understand that I can edit one then save it to the Internal banks, but is there a way to just override how all of the Preset banks Br ones play?
I'm afraid there is no quick way to do that. (You could of course write an App that does this, or use a patch editor). The only way is copying all the sounds to the internal banks, and then go through all of them one by one, and save them in place.
@@mr_floydst Okay. Thanks. Should the LED next to the Song button be lit all that time? I can't find a way to make it go off, and I have no clear idea what it means. I've found one faulty button, so far, and I'm wondering if I can't turn that LED off because there's another one. The Voice button puts me in Voice mode, but that Song LED still stays on. I've been trying to load VL voices from VL-Wizard, which works fine with my VL70m, but no luck, so far. It shows the patch name for the sound/voice I've sent, but still sounds like a piano. Maybe it's time to give up on trying what I wanted to try, until I'm a lot more familiar with it, and just try out the piano sounds...
@@mr_floydst I got a bit further with playing VL70m sounds, saved with VL-Wizard. It needs a configuration sequence sending to it (partly sysex), followed by the Yamaha FAQ on how to make VL voices respond to Breath (Voice, Edit, F7, F8, F3, confirm with +/Yes). Saving them on the EX5R is a different problem though, and needs another workaround - I have to manually edit the saved sound to point to the right VL voice, otherwise they all sound like trumpets. After all that, my brain needs rebooting, but at least it is possible. It didn't take long to save and edit the built in ones to enable Breath, because it turns out there aren't all that many of them. Thanks again.
A software version of the Yamaha VP1....that would be it ....I don't understand why Physical modelling of the type VL1 hasn't returned....with the processing power we have today....😒 Same for FDSP synthesis as offered on my beloved EX5.
@@mr_floydst Are we talking about the same Yamaha VP-1? ua-cam.com/video/B2TA9KI18XE/v-deo.html I think the software is a emulation maybe of the Yamaha -VL1... PS.= yep, I just checked...the software synth you mentioned is a software emu of the VL70m....not the ridiculous expensive and highly limited (and pretty unknown) VP-1....but I appreciate it. I have already VL synth capabilities on my EX5.
@@mr_floydst Definitely...I'm just surprised that Yamaha seems to be completely unaware of what individuals like us are looking for...the new Yama he a stuff bored me....AWM2 based Masterkeyboards for the Band keyboard player.....but nothing for the electronics music crowd that wants (for example) a modern Fs1r with Knobs, a modernized VL physical modeling synth, FDSP 2.0 , maybe a beastly return of the An1x.... I sometimes don't understand Yamaha....
@@Raydensheraj At least we'll have 2 new modeling hardware synths, the Aodyo Anyma Omega and the Expressive E Osmose, soon. (I will have the latter on my channel some day)
@@ruscular I think that's a thing I had to rehearse a lot. ;) I know you can buy those straps and the guitar belt for the Reface series, but I never considered those.
@@mr_floydst People are willing to pay more for the older Yamaha keytar SHS 200 than do for the newer one with the highly blasted Jam mode! Yamaha needs help in marketing the Keytar and they went backward on that dept. The Reface straps are not popular and flimsy and awkwardly not like a keytar like holding the pitch wheel.
I really miss a polyphonic VL synth... Pipe organs could be emulated in a more expressive way than the real counterparts, which lack air pressure control.
If you're looking for a hardware synth, there's the Aodyo Anyma Phi (which is 3 voice paraphonic) and their new model, the Omega (which seems to be not quite ready for market release yet), that's a super impressive and intuitive hardware "VL" synth. The best "VL" synth overall is Reason Objekt, I have a video on that one. It's a Reason Plugin, though.
Yes, I'd prefer to avoid computers, although I know about their potential. I am glad to know about the Omega, which might be what i am looking for. Thank you!
Oh, wow, it's you! Love your videos! Yep, I forgot to link those. You can find the patches here: - My EX5 soundset updated with the instruments shown in this video: gumroad.com/l/OyFng - The MPC One keygroup program plus samples: gumroad.com/l/SYBYG
I agree with you Floyd more and more companies are now moving to physical modeling and not Yamaha who has the knowledge what ? Aodyo is malingre the Anyma Phy (hardware Synth) Respiro (software Synth) and Yamaha ??? Thanks Fred
Yamaha have probably forgotten how to do this stuff, the original engineers have likely moved on or retired by now. IMO they should partner with SWAM and put a beefy i7 Intel chip in a hardware synth!
4:37 How do you get to those settings? I can't read the screen on your EX5, but I assumed the manual would explain it... It's completely useless, in that respect. I don't see anything at all about settings for the VL instruments, just an overview article, in very vague terms with no specifics about how to set anything. Are the settings equivalent to the ones on the VL70m, just hidden in a different menu, or does it not have the same details? I can edit all the parameter of an "Int" on the Vl70m, but I'm a bit baffled by the EX5R menus.
Hi, thanks for watching! When in EDIT mode, you can use the patch selection buttons 1-8 on the right side of the keyboard to quickly switch screens. They're labeled with the screen name they will open ("Common", "Amp", "Pitch"... and so on)
@@mr_floydst I didn't know they were also patch selection buttons. I found a list of VL "controllers" pretty well hidden in the separate Data List document (126-146)... So, does it all have to be done via [F7] CTRL [F8] SET? I see VL Pressure in that list but no VL Pressure Curve, so it seems it's not the same as a VL70m... unless there's some other way to change the curve, without needing a controller for it? For the VL70m, the pressure settings are like this: Edit -> CONTROL -> CONTROL EDIT 13 onwards: Prs CC No.=02 Prs CtrlDpt=+127 Prs Curve =-08 There are similar settings for Embouchure, Growl, Damping, Toungueing, and so on. If it has to all be set through the mod matrix (SET), there doesn't seem to be enough lines to set more than a few of them at once. Do we just have to choose a subset of them? I'd been told VL-Wizard, which I've used to edit VL70m voices, and as a sysex librarian for it, works with the EX5/EX5R except for EQ, but it's looking like that's just not possible because they're so different. The one thing that does work is saving the Custom elements to the EX5R, though I then have to tie them to a patch by hand.
While VL is technically impressing, normal sampling does a pretty good job at imitating real instruments too (lets say you had sampled a real guitar on the MPC One instead). And you doesn’t have to be an engineer to play the real instruments, surely that’s why Yamaha still makes those. Not saying that I don’t want new VL instruments (a Reface VL1 would be cool!), but rather for making more ”untraditional” sounds to be used with different MPE controllers, like an ”affordable Continuum” (even though you get the Eeagan Matrix with the Mini version too.)
Thanks for watching! Maybe it's my fault 'cause I stuck to "acoustic" stuff in this video, but this tech can produce some rather "out there" sounds you just don't hear on other synths. (The "Mad Tube" is only one example, there are other synth and "hm what's that but it sounds cool" sounds in there). It really depends on the parameters you dial in. And yes, I'm looking forward to the "osmose" synth which is said to be released later this year (I'm one of the backers). It's using the Eagan Matrix and I can't wait to get my hands on that. :-)
@@mr_floydst The problem is to balance komplexity to user friendliness. It would be nice with some kind of ”MIDI-standard for MPE” that auto-mapped the available controllers (of course with the possibility for furher configuration.) What are your opinions on the Eegan Matrix vs the Equator 2, regarding sound, flexibility and user friendliness? Myself, I have no experience with neither. And do I dare to ask if you think it’s possible to install either of them on a small ”stand-alone device”? If I’m correct, both the Continuum and Seaboard has built-in synthesizers?
Thank you! It's the "haven't talked to a native speaker in 30 years and can't be bothered to look up the correct pronounciation on google each time" accent. :-) If you can understand what I'm saying it's mission accomplished for me. :-)
In case you don't want to read the video's description, here are some links you might want to check:
www.native-instruments.com/en/reaktor-community/reaktor-user-library/all/all/all/300659/ - a collection of waveguide synth modules
www.osar.fr/notes/waveguides/ - Pierre Cusa's notes on waveguide synthesis with some impressive javascript demos
en.wikipedia.org/wiki/Digital_waveguide_synthesis Wikipedia on waveguide synthesis
There's also a plugin called S-YXG100plus-PolyVL which is an 8 voice VL synth for Windows, abandonware by now. I could not get it to run on Windows 10, otherwise it would have made it to this video.
Did you try setting the old plugin to emulate an earlier version of windows and set to run as admin? I get the alesis fusion soundfont tools to work in win10 this way
Its very rare to have a technical person also be a great musician. Thanks for the VL explanation.
Thank you very much. I don't know if I'm a great musician, but I observed the synth crowd and the IT crowd venn diagram resembles a circle. ;-)
@@mr_floydst I've noticed that too. A lot of synthesizer technicians are also musicians, with backgrounds in higher mathematics. Historically, musicians did mess with their electronics quite a bit…not often in this much depth though. Kind of a nested circle.
Check out Benn Jordan too. Science? Check. Music? Check. But yes, Floyd is cool isn't he?
As an owner of EX5 I agree that VL is a great technology . Very expressive, particularly with breath controller. I bought Friktion from Reason last year. It has similar vibe to VL. I agree that over the years Yamaha gave up on cool technologies: VL, FDSP, Fseq . EX5 was underpowered but it was not a good excuse to abandon and focus on ROMplers like Motif series
Thanks for watching! I'm waiting for my specimen of this year's "expressive osmose", that super expressive keyboard based on the "eagan matrix" synth. That looks a bit like it will walk in these steps... let's see. :-)
Very Powerful....can you imagine this engine updated for 2022
Now that would be something! Yamaha's attention to detail here was astounding.
They really should. I mean they did the very same with the SY99 when they brought out the Montage. This would be a killer of a synth
Dear Yamaha,
do it. please.
bring it back.
Kind regards,
Everyone.
I'd buy that. :)
I Love my EX5. It is an amazing machine!!
Thank you so much, please make more videos about Digital Waveguide Synthesis techniques. I'm very interested on this topic.
Thanks for watching! I might Do some more on this topic in the future.
I have both!!! Was really complex and amazing synthesis, hybrid acoustic tones is the best. Presence, harmonics, all!!
Congrats! I agree, Yamaha's implementation of that algorithm was complex and well-thought out.
Please, Yamaha!! Great video Floyd!
Thanks! :)
I agree, it needs to be brought back- immediately. Once again, wonderful work Mr Floyd!
Thank you very much! :)
I am an happy owner of 4 vl70m and 3 Vl1m, I program the sounds by myself with some editor and I can say this synth is very underrate, you can create good acoustic sound emulation and new acoustic sounding instrument but you must program it by yourself because the original patches are not rapresentative of it’s real power. Even the few patches from other tihrd parts producers are not very good, you can do better but it’s a very difficoult work and it take a lot of time.
Thanks for watching! You are right, programming the "models" is very much like FM, but with "sharper" results. You can quickly drive the sound into distortion of the "argh my ears" variety. It takes a lot of time and consideration to create a pleasant sound and even more time to find sounds that resemble actual instruments.
The original designer of the VL1, Toshi Kunimoto, give me the original program for program the VL1, but only works on Mac OS 3 if have you some machine could could work for Tun it I will try to share you. I use to work for Yamaha Mexico and meet him in 2006.
Great stuff - was immediately impressed when I first heard the VL1 and then I saw the price :( It's one reason I implemented flute and plucked string models + a tube resonator in my soft synth 'Hyperion Synth' - the next release will also feature MPE support. I would love to see some new hardware versions of the VL series again - the nearest thing I found that is a bit like that is the Sample Modeling VSTs which sound extremely realistic and naturally expressive.
Waaaaiiit. Please send me the link and infos :-) (UA-cam will mercylessly remove comments with links but I should be able to pick them up)
@@mr_floydst SWAM instruments are also a bit VL like. I had a Korg Z1 which had physical modelling, but sold it once I bought the SWAM trumpet VST. Its much better realism wise, which is'nt surprising considering the Korg Z1 was released 25 years ago!
@@lundsweden How is Swam compared to the z1 for non realistic sounds?
@@kicksNCY Haha thats the weak side of it. The Z1 was a synth you could get whacky sounds out of. SWAM is straight up vanilla realism!
yeah its a shame this tech has not been pushed any further, it does sound really good, great video dude!
Thanks! :) There are some plugins around which use the tech.
Hardware is still preferred by some users over software. I have VSTs and DAWs and they serve a purpose. At the same time, I have a couple of hardware synthesizers, a workstation digital piano, a digital stage piano, and a MPC. I’d rather use hardware when it’s time to play. Software gets the job done but as a hobbyist, hardware gives me more personal satisfaction and enjoyment.
You’re a multi instrumentalist. Must be nice to have those skills.
Yes, you're right! Using a mouse or a touch screen adds a layer of abstraction to the interaction with your instrument. Having a touchable user interface (buttons, knobs, faders and keys) always is superior.
Hey Floyd, I managed to get all of the original floppy disks for the EX5 from Yamaha. There is one internal preset that I love (Love all of them) but it's called "Made in the USA" Beautiful DX7 Electric piano. Anyway, all of the disks are now stored on my computer and I have access to them anytime I need now. Yamaha sold them to me for 8.76 a piece. If you ever need the factory presets, I can send you an image of them.
Thanks for the info!
Yamaha has thrown away some great technology over the years.
They have! No doubt the current Yamaha synths are very useful, but they're clearly aimed at "players" more than "sound tinkerers".
They were brave enough to develop these technology, but it just wasn't in fashion. People wanted straight forward analog modelling synths with lots of knobs and sliders. Synths such as the Nord Lead, Roland JP8000 and Access Virus were much more successful, and we still see their descendants today.
Love the VL side of EX5 wish to buy another EX5 I miss it. Ive done some fun stuff with my new Peak same way as yourself :) Great video as always
I agree! I'll never sell my EX5 (unless a reasonable update comes along, which is highly unlikely). ;-)
What a great explanation so clear thank you
You're very welcome!
Yes they really should bring it back. It has so much potential. It really is splendidly dynamic, there aren't words. I am the happy owner of four VL PLG cards, perhaps I should get another two someday.
Thanks for watching! With todays CPU power, you could easily build your own VP-1 knockoff, with unlimited polyphony. But it's the experience of the hardware maker and the attention to detail that makes or breaks such an instrument. There's incredible attention to detail in most of the VL voices. I'd really like to have that in a new instrument.
@@mr_floydst Digital design for a hardware FM synthesizer (tone generator) is like…you can do 96KHz sample rate 24-bit 1,024 operators at some 166MHz. It's not possible to do that on a computer right now-even with a 100GHz 500-core processor, the amount of data contention would be huge, you'd never be able to physically access memory fast enough. Resolving this on a hardware design is a matter of carefully pipelining the thing (1 FM operator = 1 clock cycle) and using separate bits of SRAM (with their own addressing and data buses) so you can access dozens of pieces of information all at once. Even GPUs use a massively parallel architecture with huge, huge cache lines (integrated graphics faces a memory bottleneck).
With FPGAs and good design tools (and things like Amaranth to make design much easier), you could do this easily today. In 1990 you could easily do high-polyphony VP-1 kind of thing, but you'd need a dedicated hardware pipeline (like the FM chips Yamaha used to make) instead of a processor running instruction code, FPGA was expensive, and ASIC process is always expensive.
CPU power is the wrong way to go for a lot of this stuff, although you're right that this is a relatively trivial computation. Higher polyphony will hit memory bottlenecks in any case.
@@johnmoser3594 Thanks for your insight! Hm. But I did see some JavaScript demos, so I thought the whole process is comparably "cheap"? Also, you can setup up VST plugins routed like the diagram in this video and get some fascinating results (without the attention to detail Yamaha showed in their implementation, though)
@@mr_floydst The process is kind of cheap, yeah. It's a matter of scaling it. Yamaha's FM-X platform has 512 FM operators and so can produce 512 tones (additive channels are 1 tone per operator) with like 128 polyphony (4 ops per channel, even if one channel is 4 ops not modulating each other).
Consider the simple recursive delay. You need to store key-on state so you know whether you're starting with excitation or just holding sustain, or if you're in release. Your filter needs its settings (poles, zeroes), likely an IIR filter with 3 poles and 2 zeroes and with 4 delayed samples (in overcomplicated terms, z^-1 boxes)-look up Direct Form or Biquad, second order is most common, b0-b2 and a0-a1 plus the four z^-1 boxes. You need one sample per sample delay, but that's just a ring buffer. You also need your feedback multiplier.
In all, for 16-bit audio samples, you will need 8 bytes per filter of z^-1, plus let's call it 10 bytes for 16-bit poles and zeroes, a total 18 bytes or 144 bits each. 8 bits for your feedback decay is probably fine. DRAM like DDR is typically single-port, so the computer can only read/write; internal SRAM in a hardware implementation has simple dual port usually, so you can read and write simultaneously to two separate addresses.
In a computer, you'll need to access memory and fetch around 45 bytes of data for these things, and a cache line is 64 bytes, so this is mostly fine. This leaves enough room to also incorporate a good 10 bits of state for your exciter, but that only gives you 10 bits of phase and frequency, so you're going to need to draw two cache lines most likely, in a sequential burst read, which is fast. You calculate one of these by doing each stage of the math one instruction at a time, with the data pulled into registers, manipulated, and then written back to memory. You'll need a significant amount of memory for your feedback delay, and this can't be read at the same time as your channel state, but it doesn't need to be and so being slow here won't matter…on a computer.
It doesn't matter on a computer because the CPU will need to sequentially pull in the data and operate on it step by step, then write it back. It needs to write back state as well-the new z^-1 boxes and the samples in your delay ring buffer, along with the updated exciter phase. If you're using any vibrato, there are more calculations to do here to get all of this done, and more data to read. Each step of the way, you're dealing with slow access to cache (even L1 cache can take a few cycles-stuff is copied into the CPU's load-store queue when transferring between L1 and registers), and even slower access to RAM when there's a cache miss (this can be hundreds of cycles). Each CPU instruction may take a few cycles to execute, there may be pipeline stalls, and all manner of other mishap along the way.
In a hardware pipeline, the problem is ab it different. Your hardware will have each individual step of the computation lain out, hard-wired, one at a time. At points where you need to read one of these values, you'll have an internal SRAM dedicated to whatever block of information is being read-for example, an SRAM that contains the first filter's configuration, and a *separate* SRAM containing the second filter's configuration. Exactly one stage of the pipeline reads these configurations; the data from this is used, altered, and passed down the pipeline as needed. A later stage will write back to the same SRAM, which is where that simple dual-port thing comes in. Largely, you have to make sure that each individual piece of data that must be read through the process can be read/written exactly one time at exactly one stage of your entire pipeline for the generation of each single sample.
Those data will need another write at a different point in time. Your clock rate must be such that you finish all computations with time to spare, and then have a window of time where the external controller can tell the chip to update configurations, such as making a key-on at a certain frequency, while the pipeline is not doing any calculations and so not accessing these internal SRAM devices.
The result is that in hardware, if you have a 166MHz (48000*288*12) chip and it takes you 2,000 clock cycles to perform all your computations, you can run at most 3456-2000 = 1,456 polyphony. Call it 1,024 polyphony, plus 432 cycles during which any instructions to update configuration can be applied (which is not enough to update all configuration! Have a look at the limitations on OPL3 register access: you can't change all the registers between each sample, either).
In software, the architectural considerations regarding organization of configuration and state data are at the whims of the PC's architecture (or whatever system is executing your code). You aren't going to be able to random-access your way through that much RAM, and you can't simultaneously read and write; running in parallel requires parallel CPU cores, or some really nice pipelining and speculative execution, which is in general pretty good, but not that good. You have a lot of computations to make, and so you're going to need some scratch memory beyond just your registers. You aren't going to produce 1,024 tones at 48000Hz sample rate in 166MHz-not at all. 48,000Hz sample rate with 100 instructions per computation done sequentially, with perfect pipelining and single-cycle memory access (which cannot be parallelized), is 4.9GHz. Instructions include things like load, store, and all the mathematical operations and program state operations in between; a direct form 1 second-order IIR filter is 9 loads, 4 stores, 5 multiplications, and 3 adds.
You'll have better luck with this than you will with software FM synthesis, but it's still pretty brutal in software versus hardware. This brings me back to my original point: I talked about 1,024 polyphony, but the figures I gave there also assume you have a sample delay of exactly 1 sample or 0.020ms. If 1ms of delay is allowable, you can actually have a very long pipeline that nevertheless only accesses each of your SRAM banks at one read and one write point in the process. The pipeline might not finish with the first sample until you're well into producing 50 samples, but it'll finish. At a slower clock rate, say 16Mhz (OPL3 was 14MHz), you can do 32- or 64-voice polyphony like this pretty trivially.
CPUs today can do a good bit of polyphony on things like FM synthesis and piano physical modeling, but raw processing power has nothing on a well-designed, dedicated hardware pipeline-and that was accessible in 1988.
@@johnmoser3594 Thank you very much for your in-depth explanation, this totally makes sense to me. I wish I could pin this; one of the most detailed and well-argumented and knowledgable comments I've ever gotten on a video. It also happens to be one of the most enlightening comments on the hot topic "why do we need digital synths when we got DAWs".
Hm. I could even imagine making this the topic of a video, ... it's really an "aha!" moment for me. :-)
Hi Floyd, enjoyed the video. In particular the simulation of VL using delays on the Peak. Perhaps it might be of interest to you to do a deeper dive on the Peak relative to the intriguing VL simulation technique you demonstrated. Maybe a Peak/VL episode exploring how to leverage the Peaks capabilites to create "basic" EX5 like sounds. Gratefully.
Perhaps I'll do that in my next live stream. :)
This or FS1R... so good.
Thanks for wonderful video and music.
Thanks for watching!
I own a VL1. It has a slightly different engine and sounds SuperB
Thanks for watching! That's great.
Read my mind with the MPC ONE, I've been making it make polyphonic sample banks from all my mono synths
Nice thing is Akai uses standard WAV files instead of some proprietary format, so you can use those sets anywhere else.
Clearly we need a Reface EX! 😎
Absolutely! (a Reface VL, to be more specific) :)
It's a pity more synths don't enable breath control. From my experiments with my Roland AE-30, the envelopes seem to get in the way because the designers couldn't imagine any controller except for a keyboard! 🙄 My Yamaha DX27 and Behringer Neutron are the most compatible synths in my collection. 😎
You're right, a lot of synths don't bother. which is unfortunate! BC can add so much expression and a totally new angle to sound creation.
No one has the time to learn how to play an instrument anymore, much less learn the nuances of a breath controller.
@@synthrodriguez7518 there's a certain irony in that the reason I turned to the Aerophone is because I'm too impatient to wait for my keyboard skills to improve!
Thanks for this interesting video. It almost looks a bit like physical modelling with all those parameters. And it also sounds very good. I agree it's a bit of a shame that Yamaha only uses FM and Sampling in their newer Workstations (despite them still being great of course).
Thanks for watching! It is physical modelling. And it was groundbreaking in 1994. Everyone talked about this, but the keyboards were so expensive no one could afford it (that changed when Korg released the Prophecy monophonic synth).
@@mr_floydst Ah I see. Thanks clearing things up.
Wünderbar, Alex!
Dankö. :-)
Hi Floyd, love your song with the EWI and MIDI guitar. I have two questions about the EWI, what model is it, and did you play a wind instrument before buying your EWI, or did you learn to play ON the EWI?
It's the Akai EWI USB. I learned the recorder as a kid, and then leveled up to tenor flute, as all kids do ;-) With that knowledge, playing the EWI is relatively straightforward.
@@mr_floydst Haha here in the 1980s Australia we had to learn plastic recorder in high school. It had to be out of tune to be authentic, imagine 25-30 out of tune recorders- audio hell! 🔥
Changing the subject, I remember you seemed a bit disheartened last year. I notice that for this video your ratio of views to subscribers is a lot better than what many channels get so you must be doing something right! 👍
Thanks! Well, channel growth was really bad until the end of march, when I did that QY100 video, which for some reason gets recommended to everyone and his parrot, too. ;-) Seeing those numbers grow is without doubt motivating to make and try new stuff.
@@mr_floydst I don't have a parrot but my friend's Norwegian Blue is keeping an eye on eBay for a QY 😀
Remember VL from the mags like SOS. Only seen one in my life. Just like a Waldorf Wave.
It's a rare synth indeed. There are some VST plugins that use digital waveguide synthesis, but nothing's quite like the VL.
@@mr_floydst I must say that todays synths can do pretty convincing emulation of acoustic instruments. Got a Jupiter 50 and the SuperNatural stuff is quite nice.
@@michelvondenhoff9673 You're right! I completely forgot about SuperNatural. :-)
@@mr_floydst I have checked on the EX5, seriously impressive synth still today 😀
Great !! Do you plan to own Yamaha montage? . Also it has endless possibilities. It will be a great most wanted instrument in the future like sy99. Two of these will be great couple of synth
Hi! If I had a budget and the space for I high end workstation, I'd rather buy the Korg Kronos, as it's more geared towards sound tinkerers (I think). But as I have set up a rule for myself to only spend money from hardware sales and money earned from UA-cam into new instruments, that's just not a thing at this moment. :-)
I can't really imagine any yamaha product to bring back since most of their original stuf still works
Thanks for watching! You're right. But Yamaha's VL is limited to two voices (or one voice on EX5), and you can't edit the "models" itself. Putting the software on today's hardware could be interesting, even as a VST plugin.
This year, we'll get some new modeling synthesizers (the Osmose and the Omega), let's see how they fare.
@@mr_floydst if so i would like a resynthesis option to superimpose a sample onto the vl parameters , i had fun in the 80s with a program emagic x-alayser wich could read a Midi sampledump into the Atari and transfer them to dx7 parameters , pretty crazy stuf came out
@@cnfuzz While it's not exactly what you're describing in the second part, the sample and VL thing is EX5's "FDSP" engine. It takes a sample and applies that recursive delay which creates the VL sounds to it.
I never heard of that x-analyzer. That sounds super interesting, I'll need to research. :-)
I remember VL. It didn't really seem to catch on but it was revolutionary in the day. Do you think the advent of sample libraries had something to do with it?
Hi, thanks for watching! I think cheap storage space (RAM and big hard drives) were the final nail in the coffin. But the Main reason was that the VL instruments were just too expensive and the general audience didnt see the benefit compared to romplers. But CPU performance has increased hundredfold since then..
@@mr_floydst Being able to design your own hybrid instruments was interesting. But isn't VL just software like all digital synths? It wouldn't be so expensive to create one would it?
@@IanWaugh It's all software. There even was a plugin for Windows in the mid 90s which could do 8 voices of VL, but it's abandonware now. I guess you could build your own VL synth using a RasPi and a browser and some knobs, but I'd be willing to pay for a good hardware user interface.
@@mr_floydst Yes, you can't beat a good hardware interface. Maybe someone will make one...
@@mr_floydst hi ! Sounds beautiful-I expect you've answered this elsewhere so apologies but is there a way to boost or upgrade the CPU/OS in the ex5?
I've recently acquired an EX5R, but don't know much about how to change things, so far. Is there a way to get all of the Br presets to play with breath control (CC2)? I found a Yamaha FAQ which does it temporarily for one patch. I'm looking for a way to do it globally, so they all play that way.
I understand that I can edit one then save it to the Internal banks, but is there a way to just override how all of the Preset banks Br ones play?
I'm afraid there is no quick way to do that. (You could of course write an App that does this, or use a patch editor).
The only way is copying all the sounds to the internal banks, and then go through all of them one by one, and save them in place.
@@mr_floydst Okay. Thanks.
Should the LED next to the Song button be lit all that time? I can't find a way to make it go off, and I have no clear idea what it means. I've found one faulty button, so far, and I'm wondering if I can't turn that LED off because there's another one. The Voice button puts me in Voice mode, but that Song LED still stays on.
I've been trying to load VL voices from VL-Wizard, which works fine with my VL70m, but no luck, so far. It shows the patch name for the sound/voice I've sent, but still sounds like a piano.
Maybe it's time to give up on trying what I wanted to try, until I'm a lot more familiar with it, and just try out the piano sounds...
@@mr_floydst I got a bit further with playing VL70m sounds, saved with VL-Wizard. It needs a configuration sequence sending to it (partly sysex), followed by the Yamaha FAQ on how to make VL voices respond to Breath (Voice, Edit, F7, F8, F3, confirm with +/Yes).
Saving them on the EX5R is a different problem though, and needs another workaround - I have to manually edit the saved sound to point to the right VL voice, otherwise they all sound like trumpets.
After all that, my brain needs rebooting, but at least it is possible.
It didn't take long to save and edit the built in ones to enable Breath, because it turns out there aren't all that many of them.
Thanks again.
A software version of the Yamaha VP1....that would be it ....I don't understand why Physical modelling of the type VL1 hasn't returned....with the processing power we have today....😒
Same for FDSP synthesis as offered on my beloved EX5.
There is a software version of the VP - search for YAMAHA S-YXG100 PVL
It's abandonware and it might or might not work.
@@mr_floydst Are we talking about the same Yamaha VP-1?
ua-cam.com/video/B2TA9KI18XE/v-deo.html
I think the software is a emulation maybe of the Yamaha -VL1...
PS.= yep, I just checked...the software synth you mentioned is a software emu of the VL70m....not the ridiculous expensive and highly limited (and pretty unknown) VP-1....but I appreciate it. I have already VL synth capabilities on my EX5.
@@Raydensheraj You're right! Not the same. But it's interesting this plugin existed. :-)
@@mr_floydst Definitely...I'm just surprised that Yamaha seems to be completely unaware of what individuals like us are looking for...the new Yama he a stuff bored me....AWM2 based Masterkeyboards for the Band keyboard player.....but nothing for the electronics music crowd that wants (for example) a modern Fs1r with Knobs, a modernized VL physical modeling synth, FDSP 2.0 , maybe a beastly return of the An1x....
I sometimes don't understand Yamaha....
@@Raydensheraj At least we'll have 2 new modeling hardware synths, the Aodyo Anyma Omega and the Expressive E Osmose, soon. (I will have the latter on my channel some day)
I wish the Yamaha Reface series had breath control function as well!
Thanks for watching! I agree. That would have been great.
@@mr_floydst I put in a idea request to make a keytar from the reface engine! Like the body of the SHS-200. 49 keys.
@@mr_floydst Have you ever consider doing a keytar analysis?
@@ruscular I think that's a thing I had to rehearse a lot. ;) I know you can buy those straps and the guitar belt for the Reface series, but I never considered those.
@@mr_floydst People are willing to pay more for the older Yamaha keytar SHS 200 than do for the newer one with the highly blasted Jam mode! Yamaha needs help in marketing the Keytar and they went backward on that dept. The Reface straps are not popular and flimsy and awkwardly not like a keytar like holding the pitch wheel.
Help me my performance mode on my ex5 says init perform I can't find alot of my patches
Hello, I already gave you an answer on another video. :-) You need to download the factory disk set and reload the performances.
@@mr_floydst thanks man
I really miss a polyphonic VL synth... Pipe organs could be emulated in a more expressive way than the real counterparts, which lack air pressure control.
If you're looking for a hardware synth, there's the Aodyo Anyma Phi (which is 3 voice paraphonic) and their new model, the Omega (which seems to be not quite ready for market release yet), that's a super impressive and intuitive hardware "VL" synth.
The best "VL" synth overall is Reason Objekt, I have a video on that one. It's a Reason Plugin, though.
Yes, I'd prefer to avoid computers, although I know about their potential. I am glad to know about the Omega, which might be what i am looking for. Thank you!
Could you up;pad the patches you created?
Oh, wow, it's you! Love your videos!
Yep, I forgot to link those. You can find the patches here:
- My EX5 soundset updated with the instruments shown in this video: gumroad.com/l/OyFng
- The MPC One keygroup program plus samples: gumroad.com/l/SYBYG
@@mr_floydst love your stuff man. Ex5 still my favorite. I would like to contact you for a project.
You're welcome - best send me a mail to fs at floydsteinberg dot com
I agree with you Floyd more and more companies are now moving to physical modeling and not Yamaha who has the knowledge what ?
Aodyo is malingre the Anyma Phy (hardware Synth) Respiro (software Synth) and Yamaha ???
Thanks Fred
Thanks for watching! I agree, at the moment, their portfolio is rather conservative.
Yamaha have probably forgotten how to do this stuff, the original engineers have likely moved on or retired by now. IMO they should partner with SWAM and put a beefy i7 Intel chip in a hardware synth!
4:37 How do you get to those settings? I can't read the screen on your EX5, but I assumed the manual would explain it... It's completely useless, in that respect. I don't see anything at all about settings for the VL instruments, just an overview article, in very vague terms with no specifics about how to set anything.
Are the settings equivalent to the ones on the VL70m, just hidden in a different menu, or does it not have the same details? I can edit all the parameter of an "Int" on the Vl70m, but I'm a bit baffled by the EX5R menus.
Hi, thanks for watching!
When in EDIT mode, you can use the patch selection buttons 1-8 on the right side of the keyboard to quickly switch screens. They're labeled with the screen name they will open ("Common", "Amp", "Pitch"... and so on)
@@mr_floydst I didn't know they were also patch selection buttons.
I found a list of VL "controllers" pretty well hidden in the separate Data List document (126-146)... So, does it all have to be done via [F7] CTRL [F8] SET?
I see VL Pressure in that list but no VL Pressure Curve, so it seems it's not the same as a VL70m... unless there's some other way to change the curve, without needing a controller for it?
For the VL70m, the pressure settings are like this:
Edit -> CONTROL -> CONTROL EDIT 13 onwards:
Prs CC No.=02
Prs CtrlDpt=+127
Prs Curve =-08
There are similar settings for Embouchure, Growl, Damping, Toungueing, and so on. If it has to all be set through the mod matrix (SET), there doesn't seem to be enough lines to set more than a few of them at once. Do we just have to choose a subset of them?
I'd been told VL-Wizard, which I've used to edit VL70m voices, and as a sysex librarian for it, works with the EX5/EX5R except for EQ, but it's looking like that's just not possible because they're so different. The one thing that does work is saving the Custom elements to the EX5R, though I then have to tie them to a patch by hand.
I think VL synthesis still exists in their high end Electone models
Ah, interesting! I need to do some research. :-) Thanks for pointing that out.
While VL is technically impressing, normal sampling does a pretty good job at imitating real instruments too (lets say you had sampled a real guitar on the MPC One instead). And you doesn’t have to be an engineer to play the real instruments, surely that’s why Yamaha still makes those. Not saying that I don’t want new VL instruments (a Reface VL1 would be cool!), but rather for making more ”untraditional” sounds to be used with different MPE controllers, like an ”affordable Continuum” (even though you get the Eeagan Matrix with the Mini version too.)
Thanks for watching! Maybe it's my fault 'cause I stuck to "acoustic" stuff in this video, but this tech can produce some rather "out there" sounds you just don't hear on other synths. (The "Mad Tube" is only one example, there are other synth and "hm what's that but it sounds cool" sounds in there). It really depends on the parameters you dial in. And yes, I'm looking forward to the "osmose" synth which is said to be released later this year (I'm one of the backers). It's using the Eagan Matrix and I can't wait to get my hands on that. :-)
@@mr_floydst The problem is to balance komplexity to user friendliness. It would be nice with some kind of ”MIDI-standard for MPE” that auto-mapped the available controllers (of course with the possibility for furher configuration.) What are your opinions on the Eegan Matrix vs the Equator 2, regarding sound, flexibility and user friendliness? Myself, I have no experience with neither. And do I dare to ask if you think it’s possible to install either of them on a small ”stand-alone device”? If I’m correct, both the Continuum and Seaboard has built-in synthesizers?
What accent is this sound like you’re doing it on purpose. Nice video tho.
Thank you!
It's the "haven't talked to a native speaker in 30 years and can't be bothered to look up the correct pronounciation on google each time" accent. :-)
If you can understand what I'm saying it's mission accomplished for me. :-)