Physical Modeling Has Gotten REALLY Good 🤯
Vložit
- čas přidán 13. 08. 2023
- HOT TAKE 🔥 Say goodbye to 200GB violins and complicated keyswitches. Say hello to playing your phony violins expressively and emotionally.
Playlist of physical modeling past streams: • Weeklystream78: Sampl...
💗 Support this channel and join an amazing community: / bennjordan
👀 Stalk me on social media for more frequent updates: linktr.ee/BennJordan
🔴 Subscribe To My Streaming Channel. I stream weekly! • Lambient
⚡Those lovely custom acoustic panels in the background: psyacoustics.com - Věda a technologie
![Dubstep Is REALLY HARD To Make [feat. Mr. Bill]](http://i.ytimg.com/vi/vMRUUnQz5P4/mqdefault.jpg)
The sax flips give it away that it's fake...but I play saxophone. Smooth the attack between those notes specifically and you might get me. Smooth like someone halfway separating the notes instead of definitely separating the notes.
Yep, sax has long been one of the most challenging instruments to to model or synthesize. I think us sax players still have a few more years of relevance :)
Yes, the saxophone for example was not convincing at all. Don't get me wrong, this would have been a miracle a few years ago but I still vote for the sampled stuff (for a while) if I want to use something like that.
@@80s_kid To be fair, even most sampled sax and other reed libraries are notably poor at fooling any trained musician of them. The current saving grace for woodwind and brass vsts in general is how little of the population is familiar enough to be put off by unrealistic sounding ones.
@@ChadwellJ I think even the casual listener will realise something is off though. They may think that there is some effect or other on a real sax at best
I think that trumpet players can hear that a trumpet is fake, violinists can identify a fake violin, etc. The thing I've always wondered about is percussive instruments. Can a vibes player tell the vibes are sampled, when there are multiple vekocity layers, round robins, etc.? I doubt it.
on an eventful note, synful orchestra has recently become donationware and it's still quite amazing despite its age
Just downloaded it. A friend of mine used it back in the day and did some good stuff with it. It has a bit of a learning curve, but you can get some pretty impressive result on it with very little cpu on contemporary machines.
Thanks for that Information. Did not know it at all. 👍
Thank you man.
The creator of Synful is a super generous and helpful cat.
Wow, that stuff sounds awesome... I love how with zero tweaks you get varied sounding violin pizzicati even when spamming a note with no velocity variation at all
"and most importantly, stuff that can give you some fresh inspiration when you're tired of the same old VST plugins"
Banger quote. Been really suffering from "same song syndrome" lately cause every time I start a project I fall back on the same sounds and it almost makes producing unfun at times. Thank you for the inspiration benn :)
Hi Exyl!
Omg Exyl!
Hey Benn. I discovered your channel a few months ago and in that timeframe you have become my favorite channel. The topics you choose en how you choose to tackle them is amazing! Especially the video about why there arent more women in music production. That video opened my eyes and is so important. Thank you
Great video Benn! Which of these plugins would you recommend for exploring sounds at the border between organic sounds and synths?
Aye Thanks! Expressivee Imagine and Plasmonic would probably be the most fun of the bunch IMO.
Ooh, also if you have reason, Objekt is Phenomenal, and I’ve yet to find a physical modeling synth that lets you have the same control over the synth that Reason’s Objekt does.
go to bed roomie
All I can say is Swam instruments, truly next level moddeled brass reeds and strings.
omnisphere
I don't think swam is there yet honestly, it was almost there but some note transitions aren't quite right, and theres a weird buzziness and nasality to it that really builds up when you stack multiple instances like in your demo. You should also check out the physical modeling synth included with logic pro, sculpture. It'd be interesting to see your thoughts on that one.
I totally agree with you but I believe SWAM is the best if we'll look at it's ease of play for the result we're getting.
SWAM is insanely configurable tbh, so you could probably fix the transitions by making sure you set the right settings for the instrument you’re using.
My two cents: controller obviously matters a lot when talking MPE, I love SWAM but I use it with Seaboard gear (mostly). I wouldn’t really have the patience to fake all the subtleties by using CCs tied to stripes or knobs or anything similar - I feel like Roli’ gear really allows SWAM to shine.
Although to be completely fair, I haven’t used many other alternatives so it’s not like I can assure you that the experience wouldn’t be better if I did, as your comment states.
However, I can firmly say I noticed the same issues you did HERE and think it just could have been done better even while still using SWAM, and I would imagine it’s quite related to the controller used (but I can’t really say this with a 100% guarantee). Naturally just as stated on the video, one still has to “learn the subtleties of each instrument” too. But I’m sure you’re already accounting for all of this anyway - but it’s still worth bringing up :)
not to mention the timbre doesn't quite sound right, overall they sound very 2-deminsional, so when combined to form a symphony the entire thing has that auto-tuned effect.
Not from lack of pitching offset/modulation either, just the timbre of the instruments. While sampling can sound pretty realistic at times if the "programmer" knows what they're doing.
Stacking single Instruments doesn't make an ensemble. That doesn't work with samples either. I asked them and they said they would release an ensemble in the next months
Sorry to say, *SWAM* is still *FARRRR* from perfect. It sounds like a Toy or Synth trying it's best to sounds accoustic.
Pianoteq is almost there in term of sounds, response and quality.
Hope Pianoteq make another Orchestra, Horn, and Woodwind sounds.
I always thought physical modeling was the way to go. you have way more control over each aspect of a sound, which allows you to weave them together in novel ways, and you can easily switch between the amount of voices you get depending on how you couple or decouple parameters.
Physical modelling is just a fancy eq
The question is, do you think drawing several hundred automation curves for a minute of music is a good way to spend time? Because you'll draw a lot of them regardless if you sample libraries or p.m., but a good sample library will at least give you realistic phrases like scale runs, legatos, trills, etc. Not to mention that good sample libraries will have 10-ish round robins, so you have humanization built in already without needing to rely on S&H LFOs and similar.
@@orbitfoldnah it's fancy wavetables
@@seedmole it's fancy carpus strong to be accurate
@@Milan____ This is always an issue that physical modeling has to butt up against. If I have near infinite control over my sound with an uncountable amount of parameters and automation to fill out, then at what point do I end up saving time by just... learning to play the violin? It seems like in some regard you're just swapping out the medium the instrument is being played from.
Traditional sample libraries can offload the performance aspect to a professional musician. I'm not sure if I'm capable of making a violin sound as good as Joshua Bell does, even if I had hundreds of hours to carve out every aspect of the sound. He dedicated his life to discovering the sauce, physical modeling asks that of me too.
I took a class where we had to program our synths from the beginning by using coding and nodes and I wanted to fucking die
😂
This new wave has me hoping for sequels to the physical modeling plugins in FL Studio. Sakura 2 would be so sick.
It's a pity that Imageline never seems to follow up their instruments though. It would be good to see Harmor 2 as well, but that will probably never happen.
@@adamsmith7058 to be fair harmor is harmless 2
That would be cool.
Funny, I study music and I just wrote my bachelors' thesis about sample libraries in the process of scoring film music. Physical modeling hence is the next step in this evolution, and I love that you dedicate a video about this!
Maybe but I don't have time to wait for it... How long has it been around? Still sounds like an old Casio to me, I mean some of it sounds really good though
When I see most of your videos now, I just realise that you where and still are so worth to be my (and probably many's) inspiration. The Flashbulb inspired me when I was digging into IDM and working on my own first tunes somewhere in the 2000's and this stuff you release here always reminds me what music is about, digging deeper and doing what you love. Music never ends and thank you Benn, you make the world a better place.
@6:28 I can't be the only person who loved this little melody starting with the piano? I still enjoyed its progression through to adding cello and other instruments, though sure by the end it sort of meandered off to nowhere, but I still think there is a banger of a dreamy chill track in there somewhere.
Yeah, I thought that piece was pretty amazing for what it's supposed to be. Definitely felt something there, unlike Benn lol
Maybe it’s just that it was raining all day here, but I was really feeling the vibe of that sketch. When Benn shot it down I was like, whaaa? Now I want to hear a collection of Benn's rejected tracks, lol.
Yeah here I was thinking it sounded really pretty and Benn drops a “it makes me feel nothing”
I kind of feel like physical modeling will be the new TB303, it's really bad at sounding like wind instruments, but can do some interesting sounds on it's own.
Just look at the prophecy and z1. There's a few published tracks that used them
Like synthesizers in general, which were initially supposed to mostly mimic existing instruments.
If you Tweak Pianoteq to extreme you can get some cool sounds, especially the electric piano.
Acoustic Samples Woodwinds is sample modelling and the most recent release in this category, blows any woodwinds library out of the water. Is on sale right now too.
Hey Benn! I am so excited about this one Just made a video on this exact topic (physical modeling) but with a focus on MIDI Guitar software (and MPE). I can't tell you the amount of inspiration I've drawn from your videos. Thanks a million. 🤝
Thanks for mentioning Plasmonic. I’ve been using it since its inception and have made lots of crazy sounds, occasionally even ones that sound like a real instrument. The sound design design possibilities keep me coming back for more. I consider it the most MPE friendly soft synth since you can easily assign MPE parameters to almost anything you choose to modulate. Combined with even a basic gadget to manipulate the macros, you have unmatched (so far) expressive space. I was a bit worried when you were using it to make typical VST sounds on your streaming channel but I see that you grok what it is best used for.
making a basic synth led me to physical modeling and boy how I wish I could get more material to study like this video on 0:50 to 2:00. The basic algorithms and diagrams don't come easy when you're a beginner on the subject. Great video as always Benn!
45 hours? But I want to watch now!
Sampling is obsolete now ladies and gentlemen. Benn has done declared it dead
"don't lie bro. if you heard this on the radio or something you would 100% assume it was recorded music"
I'm not lying bro, the first thing that came to my mind was "that sounds like a MIDI file played with a good GM library" 😀 ...I guess I miss the dirt of the recording. Add some noise, and I might find it convincing.
Agreed
Yeah you could probably get closer by running it through a speaker and then into a microphone, along with some other qualities
@@badbeardbill9956 yepp, good idea.
@@badbeardbill9956That's what I thought a while ago. To record any synth sound, then re record it through the speaker and you might fool most of your audience
His point I guess is to not care about realism, and sure, that is fine, and frankly I find the whole let's replace orchestras thing pretty messed up, especially coming from someone on the left.
2 minutes in and the video already has the clearest and most concise explanation and demonstration of Karplus-strong synthesis I've seen - awesome!
Been watching modular videos for years. This one stands out in its perfect mix of simplicity and informative content. It’s the first one that makes me want to drop everything and open up VCV. 👏👏
I really like physical modeling for sound design, Madrona's Kaivo is one of my favorites just because the structure kinda feels like a modular rig and is pretty inviting for experimentation, it can make some realistic sounds but I think it shines more making sounds that sampling or subtractive can't do. Even the good old Prophecy can make some good stuff that at least sounds different
as a massive fan of pianoteq, who has been trying to religiously convert all of my friends, this video vindicates me in so many ways. I do 100% of my live practicing and performing with pianoteq at this point because I consider it so so much better than all of the other options, especially when you get into live playing. I can definitely get some sampled sounds to sound awesome if I do a lot of work in post, but for live performance pianoteq is seriously in a league of its own.
Agreed - I'm a huge fan of pianoteq, too. It's by far my most used VST.
@@nickmasters8474 also agree. I've many sampled pianos and a roland library. However, pianoteq seems to be my goto anymore. Organteq is a beast as well. Love em for sure!
I aint paying $100 for a vst.
@@sus-kupp go be a poor third worlder somewhere else.
Thanks for a great and timely video, Ben!
Yes, I think there's still room for sampled instrument libraries alongside instrument modeling. You choose and use the tools (whatever they are) to fulfill your intention. That's it.
FYI you can use Reason as a VST in any daw. When you buy or subscribe, you get access to both the standalone daw, and the Reason rack plugin. I understand though if you're used to the paradigm of VST-instruments to want to add them individually to your arsenal. For a brief period the Europa wavetable synth was available as VST plugin from Reason, but I think they decided to make the entire rack be the plugin with all the toys and modularity and rack-flipping fun included.
I think part of the reason why Reason Studios doesn't just sell each Reason device as individual plugins is that all of their devices are way more powerful inside the Reason Rack than they would be as standalone instruments, because Reason's rack has modular routing capabilities that open up some really awesome patching possibilities, also there's the Combinator which allows even more complex patching. That, plus there's a huge amount of devices inside Reason and not all of them would sell very well as individual plugins, so it wouldn't be worth the extra development time to do it that way, especially considering that having that many plugins to maintain would eat up even more development time, and development time is money. IMO, the Reason Rack Plugin is worth the price of admission, considering the amount of plugins you get within the Reason Rack and the modular routing potential that is included in that environment, I'd say it's on par if not even greater than the value of something like Arturia's V Collection. I mean, I love V Collection too, but there's something special about the Reason Rack IMO.
Reason is awesome. I actually do all my primary work inside of it instead of Cubase. Everything is so much fun to play with in Reason and some of these instruments are incredible
Physical modelling has been a thing for a long time already. The first plugin like that I ever saw was EVP73, I think, a rhodes piano. I remember how I was just blown away by the seamless velocity layers. I didn't even understand what the difference between sample-based instruments and physical modelling was, but still immediatly felt that I discovered something special.
Nowadays physical modelling is much more approachable and cpu-friendly. I can only recommend EXC!TE Snare and Cymbal, if you haven't tried those already. As the name suggests they emulate those drums. What makes them special compared to most other physical modelling synths is that while there is a MIDI input, the developer actually encourages you to use it with a Piezo, meaning an impulse-y audio input, to get the full potential of the synths. It's pretty crazy when you actually do that, because you find yourself hitting an object in different ways and just like a real drum the way you hit that object influences each new sound, an experience that you normally don't get from any kind of VST synth, without drawing articulation MIDI or automating parameters.
I think I remember some Sound Blaster cards being advertised as being able to do physical modelling synthesis for MIDI. Can't remember when, though. Could have been as far back as the late 1990s. Might have been the early 2000s.
I’m glad more physical modelling is out there, not for recreation of specific acoustic instruments but more for interesting sound design exploration.
I jumped from mac to pc for several years and the one thing I missed was Logics Sculpture physical modelling synth. Will check out these.
awesome video, as a pianist, who has been away from the kinda music tech world for a couple of years, and i have to say, its actually amazing how far things have come in the physical modelling space, preordered an osmose, cant wait for that badboy to arrive!
As a huge fan of physical modeled/sample modeled instruments I love this video!
Little tip about SWAM strings, they all come with electric presets which allows you to use an IR loader to add your own instrument body. I’ve been experimenting with this concept for a few years and it can definitely help a lot in terms of realism.
Though I think the sound engine itself still needs time to improve in order to truly fool the ears of an experienced violinist.
I don't have SWAM but I've been experimenting in a similar with an electric violin and impulse responses. I've been able to simulate a string quartet in this manner. Really cool stuff.
Benn, very nicely done ! The ‘realism’ thing is curious as it’s less an issue of the state of the modeling algorithms than the state of controllers driving them. No doubt the timbral accuracy has immensely improved with CPU power and model refinements. One reason PianoTeq is more complete in its ‘realism’ than SWAM , etc. to the other engines is Piano/keyboard instruments have extremely limited realtime gestural effects on timbre (initial velocity) compared to woodwinds, strings, etc.
MPE and other controllers greatly help this aspect, but still still miss some critical ability to transfer specific gestures natural to particular acoustic instrument articulations. WIth a controller designed to an instrument idiom, performed by a player that understands it, the results of a PM instrument can be >this< indistinguishable.
Some PM instrument history - Finalizing a lot of the Yamaha VL1 Presets back in the day we brought in L.A. session & orchestra woodwind players to tweak the sounds, having them play their instruments and alternating with the VL1 + WX controller to match timbres and performance responses. We got a large portion to pass the ‘room next door’ test (i.e. removing visual and speaker/audio source visual cues).
Of course, that exact same patch played from a keyboard by someone never taking orchestration or music theory becomes a massive fail 🤪. Modeled sounds need to be learned to play similarly to the real instrument it’s emulating.
Keep up the great work!
Thank you for the vibe! It's so cool and inspiring, and feels very good!
Amazing update! External MIDI sync is huge for me, thank you!!
I've been obsessed with physical modelling since the early 2000's and everytime I get deep into a project that relies heavily on it, I ALWAYS run out of cpu power. Considering how much of these engines actually rely on vector math, it's a good time to start porting this code to the gpu's.
What is your hardware currently that still bogs down? CPUs have become so much more powerful over the last three or four years...
Well GPU acceleration for these isn’t too bad an idea…
@@pixelwash9707 currently about to ditch a 16 core i9 for either an M2 Max or an M3, depending on the date the M3 debuts
@@placebomessiah Thanks for that. Perhaps you are running into software issues, that has certainly affected me. I am running dual 2.7 MHz 26 core Platinum Xeons, and have found myself getting clicks from CPU overload in audio on my DAW from many different instrument plugins, not so much in the SWAM ones, but in some of the soft synths. I found it is very much to do with how the plugins are coded, and how your system is configured, whether they can take advantage of multithreading, I think many simply do not. Bigwig itself assigns a single core to each instrument track, so if I intend to have heavy sounds, I find its best to spread them between Bigwig tracks.
@@pixelwash9707 i think you're underestimating just how many physical models + sidechains + automation I'm torturing this machine with
Really Interested in this Topic, hope physical modeling becomes the new set Standard regarding Virtual Instruments :)
I know people who said the same thing 15 years ago, unfortunately it's unlikely to happen. It's more likely that something fundamentally new will emerge, like instruments that will use AI algorithms or something like that. Synthesizer developers on physical modeling have long had no barriers to making better instruments, it's not about processor power or anything like that. It's just a pretty limited technology if you look at it objectively. Yes, they can sound good, really good. But there hasn't been any particular leap in their quality in the last few decades, and the technology itself was invented a long, long time ago. So there doesn't seem to be any reason why this particular technology will be new standard of virtual instruments.
I remember kontakt with Stradivari violin and Gofriller Cello from Garritan that they used micro samples trying to math their way by assembling all the pieces to basically physical model strings. It was impressive 10 years ago. How far we have come.
I found this very useful and inspiring. I agree that physical modelling provides something really different and engaging. Your demonstration of resonance at the start opened a creative door in my mind that I never even knew existed. Sub earned.
So I've been making orchestral music for close to a decade now (professionally for 5ish), and I have constantly thought that there has been something wrong with my computer. Even though I have a pretty beefy machine and monitor my components, I am always surprised at how hard any of my orchestral libraries are to run when using a full symphony orchestra with all available articulations. Thank you Benn, you have saved my sanity.
Side Note: I do not understand why every manufacturer tries to use only a couple cpu cores when using these vst libraries. I always assumed it had something to do with how vst's are programmed, but it is so annoying when you see a handful of cores spiking yet have a dozen other cores just sitting pretty.
Some math algorithms cannot be made parallel, so you can only do some sort of divide and conquer approach to parallelize it. Sometimes when an algorithm has been developed to make it into a parallel version is basically a whole new development and the guys who write "real" code are busy with "new" stuff that make money more vs updating existing code.
Also sometimes they just don't care.
I love physical modeling stuff, I got into them through stuff like Pure Data and Csound as those two communities are chock full of physical modeling based stuff once you poke around a little. I'm really glad to see it go more and more mainstream. It's able to be highly expressive to a level a lot of normal synth and sample libraries simply can't be with out tones of automation killing the joy of actually playing for me.
I've done some playing around with PureData, but not for music. I'm more interested in making sound effects. It is, however, about as flexible as you could ever have something, mostly because it's an outright programming language. The only real limit is your ability to analyse and implement.
Wow you're so right about these cinematic style instruments. I wonder if we'll ever reach this for things like drums
I'm just a hobbyist who has always liked using Reason, so Objekt has been a fun addition to their line of instruments. I haven't dug deeply into it, but at first blush I feel like it'll be a good creative choice.
I'm in academia relating to physical modelling & deep learning - I can tell you things are gonna get crazy over the next few years. Native instruments especially are going hard on AI combined with physical modelling. We should see a huge increase in the versatility of physically modelled instruments and acoustic spaces. Exciting!
NI?
@@thelabby9998 I'm not personally with NI but there's some excellent AI research being pumped out by them right now that I've become quite familiar with
I think major software developers are really falling behind on physical modelling, MPE is sill just an afterthought or not implemented at all (Objekt and Chromaphone). I own Plasmonic and Modulus and obviously Plasmonic absolutely kills at modulations and fine tuning the response but it's still somewhat lacking in resonator modelling. The Modulus is on the other side of spectrum, it's dead simple but you could really hear how different sound frequencies collide in a modelled piece of metal plate or whatnot.
thanks for sharing these ideas. always gives me ideas what i can try to build! like i never knew the base of karplus-strong was so simple or looked into it and now i can make my own lol.
Great episode and topic Benn. I am exclusively on iPadOS with an M1 chip, and it is this leap of power and of course tax-back money at a fortuitous moment (!), that enabled me to try many more complex instruments simultaneously, or simply better ones. SWAM Trumpet really is a knockout, and I got caught out with the realism of limits on the range, too, suddenly it’s like my midi controller keyboard stopped working a couple of octaves up!
Apple have neatly gotten around this for anybody, by having an instrument in GarageBand for example, they call ‘brass ensemble’ which I now realise, has the main purpose probably, of permitting a very wide range of octaves, by virtue of having three brass instruments (or more?) that can between them cover genuinely 3 or 4 octaves. It’s basically a French Horn, a trombone, and a trumpet? Something like that. Thus, this instrument option gives you a brass sound, that is almost synth-like in that it has a big octave range, morphing from one instrument to another, as you move out of the other instrument’s register/range of pitches. It’s a great one to grab for any-old brass sound, due to that, then you can fine-tune later, for the actual range you need that section to be in.
Agree about the breathing gaps, at first I just kept going with a solo, and realised what any competent person probably already knew, that the spaces are pretty important! I use ‘Heavy Brass’ and SWAM and various sampled instruments/sample players in iOS too. SWAM flute’s a knockout using even just ONE of the expression controls, along with tiny nuances of pitch alteration a la Ian Anderson’s playing style for example. Getting it teetering right on the edges of the instrument’s main capability, is why I love certain flute recordings, I realise that a bit like drum breaks, there’s a surprisingly large amount of the time, the songs I find interesting, have instruments where their rough edges of breath or playing, is the best bit about them. Flute being a prime example, breaths and other embellishments, steal the show for me. A professional flautist though, playing a piece, these are probably avoided like the plague.
I’m not going to be a flute player, ever. But I can imitate the thing being played, with these amazing new modellers. Should I admit to recording fake breaths, and tracking these to fit in with particularly ‘explosive’ points in the flute part of a song?! Quite fun to record, I couldn’t do it for laughing…I just imitated what I’d heard done on Tull albums a shed-load of times, and laid that very quietly over the actual flute parts in my song. It’s terrible really, to realise the thing you might be really, really best at, is faking stuff. Deceit. But the result was pretty persuasive, which reminds me of a saying I heard once: ‘there’s no criticising success’. If it works, and it’s powerfully-real sounding, that’s what you were presumably after, so does it really matter, if you ‘fool’ the audience about how it’s arrived at?
That’s a vexed question I’ve decided, a moral dilemma for musicians of our time, because it is true I think, that bass guitar sounds like they are being pounded with a supersonic thumb joint, really do seem better, when you see it made for real that way and not sampled and plinked on a MIDI controller key. Or BELIEVE that’s what’s being done. Just imho. I couldn’t wrestle that one to the ground. It is deceitful to make the listener think the actual instrument is being played instead of a synthetically-generated version, or even a sampled instrument, by various tricks. But an electric guitar doesn’t sound like it’s portrayed on many albums, except once it’s carefully got the levels controlled and excesses reined-in all over the place.
Vocals, are the well-known most-faked instrument probably. There is outcry about it, of course, all that pitch correction etc. But again, there’s no criticising success - if the tricks used work, and the end result is somewhere near what was envisioned, or at least extremely pleasing to hear, does it matter it’s pretty faked-up. Like I say, I leave it as an open file that one, because I can’t fully come down on one side of the argument or the other. Like I said, great topic and well-explored, thanks Benn. Take care all.
I think physical modeling can make for the coolest percussion sounds. Being a percussionist who also produces, this warms my heart. I can't record stuff in my room, but I can make physical modeled percussion.
Try our snare and cymbal!
Haven’t used Reason in years, but their new physical modeling synth (the the string synth one to lesser degree) has me considering updating and using it again. Also this video totally made me shed another tear for the loss of AAS Tassman 4. Never could get as interesting of sounds out of the rest of their stuff. If a Tasman 5 came out I’d be pulling out the credit card 100%. Even if it offered nothing new. Just would love to use it in my current work computers.
Tassman was amazing for its time.
@@jasperdevries1726 so many times looking back on old cues I’ll have a patch I want to reuse only to discover I used tassman for it. It had its quirks but it sure made for some unique and great sounding patches.
they have really great Players (midi effect devices) too!
Good video! Well yes physically modeling is getting better and better, although there are still some missings. For example think about orchestral sections: they still sound way more believable with sampled instruments…just listen to VSL demos to have an idea.
Yesss new Benn Jordan video 😲!
Loved this video, very thought provoking. Yet by the end of it I can't help feeling this is just a fools errand, that ultimately the problem isn't sampling or modelling but input/interfaces required to recreate the sounds, in which case why not just use the real instrument? Obviously there are good reasons for not using an instrument, least of all being cost, talent/ability, recording etc. Ultimately I find myself considering AI/machine learning in some form is going to be needed to bridge the gap, you play the notes/melody, give a few hints as to how you want it to sound, then let AI control all the parameters, all the dull, boring, busy work to make it sound like a real instrument being played.
Yep, I think mathematical modeling + ML-controlled parameters can lead to physical modeled synths that have the ease of use of sample libraries.
It's definitely this. When I was first getting into synths and production in the mid-2000's, I wanted a physical modelling synth like the Yamaha VL1. Never got one, but I realized much later that I wanted something that wasn't a piano interface, also. There's a tension between a "muscle memory" approach to music where you can converse in it - which is what you spend all those years training scales and harmony for - and a "programmed" method where you write down what's supposed to happen and someone or something else interprets it. When the two are put together well, the musical process can flow, but a lot of what's happened in music tech is just to extend the programming further and further. Every time you put another knob on the synth it becomes more of a program. And making the keyboard more expressive with extra features suggests that the producer become a virtuoso in fake instruments, which gets away from any natural way of interacting with the synth on its own terms, so while MPE gets a lot of hype, it's the kind of tech that creates niche professionals simply because nobody else will bother.
But if you had a setup of "I performed it on this instrument, now perform it on that one", then you start going somewhere really interesting because that intermediate layer of configuration is just human language. And this has previously been done in the sense of "audio to MIDI", or "song styles" in arranger software, but then the result isn't reminiscent of a performance, so much as an effect like autotune or a template like the orchestral samplers. It really needs the ML aspect to cross over into feeling like a real "automated session player".
Reason comes as a vst/aax/au plugin too. It's maybe the best gear for value in the digital domain.
aaaaaw.... doggie so dang cute(!) & so well-behaved. derailer seems like a real winner to me. gotta go back & catch the rest of the vid. thanks, benn jordan!!
I have a couple SWAM instruments and Pianoteq 8, and these physical modeling plug-ins are very impressive to say the least. I do agree on your point about sampling and physical modeling can exist one to another. Great video.
Hey Benn, you actually can use the reason rack as a VST3 plugin. I use it all the time in Ableton. Next to Object, there's also Friktion (string instrument Physical Synth). I love this combo quite honestly...
Ben show Him using the plugin, his point is He dont think it's right to pay $500 and then buy $100 for add on device. He says Reason device's should be offered as stand alone plugins.
Yes exactly, you need to own Reason to use the rack as a VST. I don't think buying a DAW to then not use it as a DAW, instead only using one small part of it inside another DAW that you prefer, is not right.
@@astralfields that's right of course! On the other hand, I use the rack for quite some more stuff than just Object. It has some other very usable synths and FX I love to use in various occasions. And the Player devices are also very ... very good. And I came from the other side ... from using Reason as main daw, to using Reason as a plug-in in Ableton.
One can of course approach this from an economic perspective and contemplate where you spend your money on a strategic choice of your DAW for your studio. That makes sense. However, If you think of it as the price of Reason as a VST (combo synth/instrument/effect - chain) that can be used ON a track in your DAW of choice and leave the DAW part alone in Reason... the economy is sound because I believe it should be compared to other similar 'sound factories'; let's say UVI Falcon and all its extras. If you are creative and making a living on your music or production work... well then another instrument/synth/effect function choice, is clearly more about what you want to work with and think is useful. If you are in the market for ACTUALLY dip your toes in the market of 'what DAW to choose for my production and I want to be able to expand in any direction'...I would not consider tinkering in Reason DAW... it is great on what its doing 2023, way beyond the history of the typical genre that it has. I would then rather contemplate Ableton, Logic and alike...use what is in them first...THEN ponder about my extended portfolio of instruments, effects and mixing/mastering.
Absynth was my Fav Soft synth of all time. I still use it today. My 2 newest favorite Synths have been CODEX & FLOW MOTION FM the past few years. A lot of times layering all 3. I am thinking of trying that new OBJEKT though, seems pretty cool.
I mostly use these for soundbeds, harmonies & arpegiation & drum layered sounds for my acoustic kit.
One thing you missed out on that can make the SWAM instruments sound even more realistic is how you have control over the bow pressure and bow position, position on the fingerboard where notes will be played, being able to smoothly switch between techniques used in playing etc... Which is way more control than just having dynamic layers + articulation like sample libraries.
Love your sense of humor and how you incorporate it into your editing. And...the dawg.
I'm a drummer and my main focus in most cases is feel, velocity and timing. your composition and the sounds of the instruments sounded awesome. The only thing that made me feel it was a bit off was that you can sense that it all was played by the same person/ input device maybe? It could be just me lol. It's like listening to Joe Satriani Flying in a blue Dream album. He played all the instruments and although all the guitar solo work sounds awesome. The whole accompagniment sounds dead and sterile.
Just to add to something you mention in the video, Objekt isn't exclusive to the Reason DAW. It's exclusive to Reason, that's true but Reason runs as a plugin in any DAW so Objekt is just as available to Live, Logic, Pro Tools, FL, Studio One, etc.
Yeah, I use it exclusively in Live as a vst plugin. Wasn’t sure what he meant by that.
@@gregoryglass4248 I think he corrected himself later but mostly off camera 😅
Man that string plucked sound made me spit my coffee, then the cello pouahahaha ! "We're almost there"
That was a smashing intro to Karplus Strong synthesis. Fair play. First time I understand how it works and it took about 30 seconds of the video
8:14 and on - with the exception of the piano(teq), everything sounds like 90s Roland Sound Canvas... and I've heard SWAM instruments in some other (older) videos before, they sounded much more realistic, so perhaps some dials may have been misdialed...
I'm a huge fan of physically modeled instruments. Pianoteq piqued my interest as a software engineer and perhaps one day I'll get into instrument simulation, but that probably involves a physics degree lol.
The one caveat is memory vs CPU. Yes, you save a ton of memory, but at the cost of a ton of CPU. Pianoteq 8 uses the CPU of basically 8 other sampled libraries.
Im liking to see physical modeling on the iOS side too, because every app is way cheaper than the vst version. And we already have good options there, like SWAM, Tomofon, some icegear ones, etc... Very good video! Thanks.
I'm a sax player and your sax section unisons are really nice.
I’ve checked out SWAM instruments, although good in theory, im yet to hear something that I actually like. Most demos that I’ve heard are terrible, perhaps in the right hands it will be any good.
This is a pretty convincing demonstration, I’d say. Certainly on its way.
Unfortunately I have to agree with @somedood6621's assertion - and also admit to experiencing a bout of cognitive dissonance - after watching this video. I really wanted to accept Benn's conclusion that the SWAM instruments in his video sounded more realistic than the samples he used for comparison. But in all honesty I felt that the latter sounded significantly better to my ears. And this is coming from someone who purchased licenses to SWAM's string and woodwind libraries to use with a Linnstrument. So if anything I was looking for validation from the video. Unfortunately, that was not the case for me. In fairness however, the SWAM instruments that I've played in my home studio sound very good, and more realistic than what I heard in this video. So this simply might be a case that listening to an A/B test second-hand via CZcams that's being streamed on a desktop computer with run of the mill speakers might not do justice to what the SWAM sessions really sounded like in Benn's studio.
For real, my favorite violin and cello plugins are the ones by virharmonic. They're sample based but they sound so good to me, way better than the SWAM ones... It makes me think sampling isn't dead, just done incorrectly most of the time lol
tho I still think that physical modeling (or even AI if done correctly) has an inherent advantage over sampling, but right now most of them sound somewhat fake (except for brass, for brass they sound p good compared to samples lmao)
@@SteveinJersey1234 The performance in this demo was quite flawed, and in the flaws it seemed more realistic to me. Who is going to be a master at playing all of these instruments, even through an MPE controller? There is a ton of nuance to each particular one so I feel like we got the "pretty good" performance across the board as opposed to the "master" performance in any particular one. That doesn't make it sound less realistic to me.
Ableton suite users have access to collision, which seems to be a wrapped version of aas chromaphone with a bit less options. It has and always will be my favorite ableton device
It is such a fun plugin to use. I agree with you that it is severely underrated as far as ableton plugins go.
I will preface this by saying that I love techno (real techno, not any old dance music) but your "boring" piece around 9 mins was quite captivating. I'm really jealous that you can just play that into your DAW. Sounds a lot like William Drake from Cardiacs.
Fun Thing that i also get rid ( not 100%...) of Sample Library based Vstis like Konkakt and get more into PM-Synths, i bought Plasmonic right away because i loved Absynth and also bought Cromaphone 3 few weeks ago but i also still love what you can do with Granular Synts an also some Virtual Analog Stuff like Arturia synths and especially Knifonium, on the other Hand i love to use things like TAL-Sampler because of it's Sound !
Keep doing what you do because your Stuff is one of the most inspiring Content Audio Channels have to offer...
Pianoteq fan here.
I remember back when I used Logic Pro 9, it had a physical modelling plucked string synth, something like a clav/harpsichord type thing. I got some wild sounds out of that when I made the string behavior break with physics. I really miss that thing.
Do you mean Sculpture perhaps? It's still available in Logic Pro X.
@@Zero-Sun probably. I don't even use mac anymore, so that's just long gone by now :p
Whoa the VCV rack demo was cool, I had never seen the basics of physical modelling before.
As a flashbulb Fan and a Reason only user, i love to see that you giving the objekt a try...! ✌
@5:14 Hey Benn, is that my template? :O I tried writing a custom plugin using the native FL Studio API but it's dogshit old and impossible to use for a novice C++ programmer. I hope your attention to MPE has sparked interest in Image-Line writing something themselves because in principle, it's not hard to do (the actual code consists of maybe 100 lines; I wrote it in another plugin that uses groovy but it's a bit too slow with all the data flowing through it).
Ah, can't even mention my e-mail here but I'll provide support on the FL Studio forum if you need it.
EDIT: I see you have changed the title somewhat to be less about modelling > sampling. Still, I think most of the points still stand or are interesting enough on their own merit.
Oh boy, okay - lots to unpack here. Long comment incoming! I certaintly WISH you were right about modelling surpassing sampling, but unfortunately we are still a long time away from that point. YES, modelling has a wider range of expressive capabilities, BUT you always sacrifice tonal quality. Sampling has the opposite problem, where it can sound 100% real (it is based on real recordings after all), but the expressive options are very limited (e.g. to the articulations recorded and quality of legato scripting). So yes, modelling can sound very cool but it simply does not sound like the real thing, especially when doing more exposed and virtuosic lines.
But, I hear you say, what about the examples in the video?? Doesn't that sound real to most people?? Sure, but "most people will think this is real" is not a good standard to judge virtual instruments. If that were the standard then we already reached peak realism 20 years ago. For example, people regularly believe the Skyrim Soundtrack is a real orchestra - but that says very little about the quality of the mockups. It simply tells us that Jeremy Soule wrote to the strengths of his samples (i.e. mostly long pads with a lot of reverb to mask the tells of sampling - I call them 'samplisms'). Most people simply don't hear whether a supposedly live instrument is real or fake, perhaps because they don't care that all that much, or perhaps because virtual instruments have their own expressive qualities that are equally valid as real ones. However, if we are going to judge our results on sampling vs modelling our standard should be fidelity to the real instrument. Can the virtual instrument do everything the real one can and sound indisinguisable from it while doing so?
With that out of the way, here are my main gripes with this video.
1. Why release this video now and not 10 years ago? SWAM is pretty old. If you had released this video to talk about AcousticSamples VWinds and VBrass, then I'd understand it more. But right now, all these examples could have been made years ago and both sampling and modelling made progress since the release of any of the libraries mentioned in this video.
2. None of the modelled examples pass the fidelity test, I'm very sorry to say. They have their own expressive qualities, which is cool and valid, but they don't pass as true representation - ESPECIALLY not the solo demos. That being said, I have heard some modelled demos by more experienced users that were more convincing (Infinite Brass, VBrass - winds are harder to get right, although perhaps I am more perceptive to giveaways in woodwind emulations). Still, even in the official demos of these libraries, or the SWAM Divisimate demos (czcams.com/video/P8YMx3VFII4/video.html), that sterile and lifeless sound is hard to get rid of. It requires a lot of fiddling with the musical parameters (dynamics, expression, vibrato, velocity, pitchbend, etc) and musical performance to get there - almost always with the help of a breath controller. Another aspect is room sound. Modelled sounds are dry as a bone, so for many arrangement they will need to be placed inside a three-dimensional space. This is super important for getting a believeable sound and unfortunately one of the hardest things to get right. For orchestral music you'd generally want concert hall sound, and this is nigh impossible to achieve with even the reverb plugins. Samples on the other hand can and often are captured in concert halls precisely to get a specific baked-in hall sound that is impossible to achieve any other way. I realize this is a more contentious point, and certain technologies such as VSL's MIR and Berlin Studio come pretty close, but overal I think this point holds for the average composer. A full exploration of this topic goes way beyond a youtube comment. It would be a good challenge for yourself, though, to to take, say, a dry string staccato sound and place it in the same room as your wettest
3. Again, I'm sorry to say but the sampled examples at 11:59 is pretty terrible. Setting aside the Sonokinetic Minimal, which is a phrase-based library so it will always sound sort of 'fine-ish', the solo cello ruins any illusion of realism. No one familiar with the instrument could mistake this for a real performance. Part of that is the midi programming here, and part of that is inherent in the Chris Hein Romantic Cello itself (as well as trying to blend it with a much wetter library such as Sonokinetic). Solo string instruments are incredibly hard to capture, and the specific ways in which CH has attempted to do so have not paid off. A complete review of CH and especially the mono-ish sound of their libraries is beyond the scope of this comment, but suffice to say I know basically no composers who use CH instruments in their orchestral works (that stay in-the-box).
3.5. When demonstrating the Chris Hein cello, you say "it's pretty easy to tell that even with the round robin we're playing the same notes over and over again." That is because you are not hearing any round robins. Either this patch does not have round robins or does not have them enabled. You seem to be using legato mode, and legato patches rarely have (true) round robins. IIRC, Chris Hein's approach only includes scripted round robins, if any.
4. Since the demos for both sides didn't represent their respective qualities best, I'm going to assume that this was not an intentionally uncharitable demonstration of the samples side to make your point, but rather a lack of deep familiarity with the instruments required to make them sound best. That might sound a bit mean-spirited, but I don't hold that against you or anyone really. Make music however you like and if a tool sounds good enough to you then it is good enough. You showed some very impressive chops in synth programming, something I admittedly know not that much about in comparison. However, if you are going to make a point that modelling can emulate real instruments in a realistic fashion, and that modelling has surpassed sampling in achieving that goal, then I think I'm free to question that point with my own expertise - even if I am also very far from the best orchestral mockupers. I will post some examples of what I consider to be very good sample-based mockups in a reply to this comment below.
With all this being said, and to end on a more positive note, there are definitely some good points in the video that I agree with:
1. Keyswitches are hell. I hate them more everyday :D Wherever possible, I tend to use performance patches that have a lot of expressive range in a single patch, such as orchestral libraries by Aaron Venture, Musical Samples, Performance Samples, and so on.
2. The disk and ram requirements for larger sample libraries is pretty insane. Yes, you need a pretty beefy computer (at least 64/128gb ram) to run any decent orchestral template with 1000+ tracks for all commonly used articulations per instrument per section. That being said, at least hardware has become affordable enough to run everything on one system nowadays.
3. If you want to achieve realistic and expressive results, you HAVE to know the capabilities and characteristics of the instruments you are trying to emulate. Then, you HAVE to know the best ways to get to the desired results using your virtual tools. This is especially true for modelled instruments, where the default sound is absolutely dead and so many expressive parameters need to be controlled for it go come alive. The fact that players need to breathe is a good, basic example of that. Another more technical example would be that, when trying to achieve fast string runs, you sometimes need to layer in some tremolos underneath spiccato (or slurred staccato) notes, to get that blurred effect. Lack of knowledge of the instruments, both virtual and real, is in my view the key reason why beginner mockups typically sound fake. It results in poor phrasing, poor balance between articulations and other instruments, poor orchestration, etc, and these are the signs that give the game away.
Anyways, TL;DR version is - no, modelled has still quite some way to go and samples are here to stay for a while alas. The examples in this video do nothing to diminish that point in my view. But let me know if you think differently :)
Here are a couple of examples of 100% sample-based orchestral tracks that effectively pass the fidelity test through expert use of the possible expressive capabilities:
Starting with the classics, we have Thomas Bergersen's demos for Hollywood Strings. This is a link to his 'Allegro Agitato':
czcams.com/video/ce-_J41LaS0/video.html
How about samples doing a piece of existing music, so that you have a good reference? Here is Blakus' mockup of the 'Star Wars Main Theme':
czcams.com/video/vDQHqf2NYmU/video.html
What about solo strings?
Here is Andy Blaney - 'Jack in the Box' (One of the most virtuosic solo violin demos, even if the actual instrument is lackluster, note the CC data!):
czcams.com/video/3gvXsaQf0Lo/video.html
But what about _exposed_ soloists? How about a solo flute? Check out the Gabrielle Flute - Demo Screencast by Aaron Sapp/Musical Sampling:
czcams.com/video/A1H1ZYJaOOo/video.html
Setting instruments aside, what about sampling the human voice in choral setting?
Here is Jasper Blunk's Oceania choir:
czcams.com/video/SsrcVoDLq4s/video.html
I don't think even SynthV, as impressive as it is, can do something like that anytime soon.
Anyways, if you or anyone reading this would like to learn more about achieving realistic results, I would recommend checking out some educational resources from Spitfire Audio, the Virtual Orchestration youtube channel (associated with Orchestral Tools and Berkeley iirc). Another good place to ask in-depth questions is the vi-control forum.
An excellent video once again! I think you a bit misjudged the element that Reason is selling its package as a range of plugins you can use as VST plugins in your any DAW of choice - instead of it being a standalone DAW you must tie into for full production (which is also an option, perhaps not most optimal)
What a fantastic video! Well put together and presented. You are a master of your craft. Regarding the real/fake discernment; As soon as I heard the piano I could tell it was emulated. Spending the last 40 years in front of real pianos, the tiny imperfections missing were a give away to me- but it's close. The full orchestra was pretty difficult to tell it was fake, however. It's hard to say if I would or wouldn't have known it was fake/real if I heard it on the radio since I was prepped it was fake. It's often that I think orchestras sound fake in modern recordings because of the known prolifiration of synthetic music nowadays, especially because it's so much more cheap and faster to produce. But....there's just.....something that's missing. And ironically, it's typically the tiny imperfections that your subconcious picks up on. It reminds me of when CD's and digital audio came out. Producers started adding artificial hiss and buzz because it sounded too sterile!
Are there any good physical modelling synths (other than SWAM) for brass instruments? Nothing against SWAM, I just haven't found as much variety as there is for string or impact-based modelling synths
Acoustic Sample's VHorns and Aaron Venture's Infinite Brass are two that come to mind, but they aren't pure modelling - they are a mix of samples and modelling.
Alright the piano and sax in swam sound so good! Cello not so much. But still wow!
Archie sounds great for strings though
Interesting, I think the sax sounds the most fake by far. Doesn't sound bad, just obviously not a real sax. But I play sax so maybe I'm biased
I don't know the sax sounds plasticky to me, too much attack and buzz. I listen to a lot of bebop so it is stuck in my head. The cello sounds horrible imo though
@@SamChaneyProductions That is a big part of it. Players of X instrument will have a much more fine-tuned hear for hearing artifacts in instrument X.
@@snarf1504 100%! i don't use or hear sax a lot to really think about those things. yet i do hear strings a lot and use it.
Hey Benn :) i thought this video was going to be about physical modeling overtaking sampled instrument libraries in fidelity, and it seemed to be going that way until it became a sort of physical modeling plugin review video.
I really enjoy hearing you mess around on the keyboard with different sounds and your own compositions so I'm happy either way, but if you're going to give a head-to-head on sampled vs modeled then I'd suggest bringing up the merits of plugs like Spitfire's British Drama Toolkit, or the Solstice libraries, as well as the recent additions to the Kontakt Factory Library 2.
There's some real magic in the performances captured in those sessions. Coming from a commercial jingle/film score perspective I don't know that modeling will overtake sampling in the foreseeable future. However, that all falls apart when you consider sound design, because the flexibility that synth engines offer (subject to interface accessibility) is much handier when the job calls for some evolving sounds. That is certainly where modeling/synth tools overtake samplers. Of course the best musical example would be some typical kind of electronic music vs some typical kind of orchestrated music. I think it'd be very fun to hear you make some orchestral music with physical model/synth plugs as well as electronic music with the more traditional sampled instruments.
Either way, enjoyed the vid and all the best to you
One option I've found that is pretty versatile is to use AAS' plugin Objeq Delay and just turn the delay line off. Gives you a nice choice of simple resonators that you can tweak and run basically anything through. Great for running samples through, or synths if you want something weird. Biggest complaint with that approach is that there's no easy way to set up key tracking - you can sort of fake it in Bitwig but it would be nice if they'd just implement it directly.
Super agree that Objekt should be a VST/Clap plugin though. The first time I saw that I got my wallet out within about thirty seconds, and then I saw that it was locked into Reason and put my wallet right back down.
Objekt sounds absolutely amazing. But the supposed-to-be-real stuff is cringe:
Swam Cello - sounds like a saxophone
Swam Violin - sounds like a mouth harmonica
Swam Alto Sax - sounds like a trumpet with hiccup
Arhe Violin - sounds not bad, but also like a saxophone
Arche Cello - low notes and 'scratches' are impressive, the rest: hello again, saxophone
(“saxophone” as in 90’s general midi keyboard for TV Movie and Series music, such as Mr. Bean ...)
I have a request. I watched an old video on animals and time perception and CFF. You talked about reptiles having a knob to adjust their CFF.
My question is in regards to our ability to alter our own CFF.
I have studied Martial Arts for 40 years. I have experienced this first hand but never understood what or why it was happening. Ive been able to slow time mid fall or in a fight, never consciously but it still happened non the less. I have avoided severe injury from this phenomenon. I've spoken with othe martial artist and this seems to be a common side effect of training.
Just watch the water come out of the tap
Thanks for your detailed videos as ever, Benn. CZcams has really been crushing the bitrate lately :( if you upload in a 4k container it will preserve it though!
Objekt is sounding amazing! I usually get surprisingly "good" results by sampling small pieces of real sounds (not quite granular but not quite loops either...). You can get some nice random results that can feel "natural" that way. It's not sounding like the Osmose but hey... I have to manage with what I have. Anyway > Great video!
My take on this is: why try so hard to make something that sound like a real instrument when there are so many cool sounds you can create with a synth that sounds really good, albeit not like a real instrument. And with a Linnstrument that is extremely expressive, with possibility to track both initial hit on the pads (to give it unpredictability like in a bowed instrument) on the X and Y axis as well as velocity of initial press and aftertouch, you can play it like it is a more living instrument. A real favourite of mine is the Linnstrument and a Hydrasynth and it's default patch - it's just a ton of fun to play around with!
Well, it’s a more efficient use of memory at the expense of CPU. But also, physical modeling could enable some cool new sounds of its own
I think Roland really nailed Physical modeling with their Super Natural synthesis. I own a Roland Integra-7 and the Super Natural acoustic sounds, more realistic than the PCM based waveforms.
I always thought it is a mix between sampling and physical modeling but I might be wrong
@@mudi2000a No. The Fantom, JV, XV series were all PCM based. The Integra-7, Jupiter 50, Jupiter 80 all have the Physical modeling Super Natural Synthesis. The Integra-7 has two Synthesis PCM and Super Natural. All the PCM sounds are from the old JV-1080, JV-2080, XV-5080 and all 12 SRX expansion boards while the Super Natural Acoustic sounds and Super Natural Synths are entirety Physical modeling. So really the Integra-7 is a reissue XV-5080 plus Jupiter 80 on one rack unit.
@@eman0828 I know all of this. I still bite myself that I did not buy the Jupiter 80 when it was EOL and could be bought very cheap. Still what I heard is that it is not completely physical modeling.
@@mudi2000a Yes is completely Physical modeling. I mean that's why the two Synthesis are sperated in the Integra-7. There as a PCM button on my Integra-7 to switch over to the sampled based Rompler sounds. Really the Integra-7 is both a Rompler and Physical modeling synth. The real difference from the Jupiter 80 as the Integra-7 has more Super Natural acoustic sounds than the Jupiter and the Jupiter has double polyphony geared towards lice performances that has an arpeggater built in. Other than that they aren't all that different besides the Integra-7 having the XV-5080 engine.
the Physical Audio stuff is 🔥 - also their plate and spring reverbs worth checking out
I bought Falcon last week. So far I've simply been exploring presets, but this video reminds me to look into the new physical model oscillators that almost certainly are not used by the old expansions.
All those physical modeling plugins you used sounded thin and really fake. Sample Libraries, when programmed right, can sound like you hired an orchestra or sometimes even soloist. It's more time consuming to program a sample library, but the key is to really study, play, and understand the instrument you are programing. Sample libraries like Albion takes out some of the complications of fully understanding the instruments. That's why you hear so many gaming and commercial music use them, especially in modern compositions where it's mixed with synths and electronic elements, average listeners won't be able to tell even if you used a real orchestra.
On the other hand, a lot of generic commercials with physical modeling just sound off. The tone, the strokes, the thinness, and the lack of the actual environment such as a concert hall just make the music seem dead. Of course this all depends on the type of music and composition you or the client is going for. Any real instrumentalist will be able to spot a physical modeling instrument but might be fooled by a sample library instrument when done right.
indeed it's how you play the note, which is why i bought an Akai EWI Solo wind synth recently.
I've looked, I've listened, I've plucked, plinked, and poked, and yet I still cannot find a plugin or software package that will allow me to write music as if I were musically talented or skilled. I am not. Not even a little. I am, however, forever grateful for those of you who are, thank you.
I’ve always been really interested in physical modeling, because you could try to realistically emulate instruments that aren’t common or don’t exist. I haven’t gotten any plugins yet, though
I agree with you on the shortcomings of sampling. Your video, however, convinces me that we're just going to have to live with it for the foreseeable future.
Thank you for the video. I'd like you to consider the complex issue of heating, overheating, and CPU utilization percentages on a musician's computer when using such products based on real-time synthesis on a massive scale. Regarding the same Kontakt and other instruments and ROM players.
This was an excellent overview thanks
Damn always Benn with the amazing content teasing a video I have to wait two days to watch
The SWAM saxophones have been on my wishlist for years. In my opinion, physical modeling works best for solo/feature/overdub type parts like sax or a violin solo or something. For me personally, when I hear it trying to sound BIG it still has that "synth with tons of reverb" vibe. Of course this will only continue to improve over time, but for the time being I still prefer samples for big orchestral sounds. The Cinematic Studio series have absolutely wonderful sounds and are VERY easy to play expressively. 80% of the time you can get by on the marcato legato patch and people won't notice, its very versatile. I think CS strings, winds, brass etc with some physical modeling soloists or features on top would be the ideal combo.
Sounds really good! I still think Sax is one of the tougher ones to model. Isolated from the mix, I might be suspicious of that first modeled sax, but otherwise, it's stunning how realistic the instruments sound. Good job on even mixing your song, even if you didn't like it. Had a bit of a Gil Evans vibe.
Please, i need more than 10 seconds of that gorgeous vcv patch!!!
Just found your channel and really like it. Did you ever share the reactor "plugins" in the start of the video? They sounded amazing!