Yamalandorg EX MK2 plus nano neo AI powered…..

[Paul Woodward]

Atrocious title I admit, but the ‘quiet’ release of the new Roland boards yesterday (I literally spotted a youtube video by accident) got me thinking about the state of modern boards and if we have reached a point where there is nothing new to innovate, just minor iterations.

Thinking back to the late 70’s when I was reaching the ripe age of 10 and an awareness of keyboard technologies. For next few years decades my jaw constantly dropped with the new technologies being developed. Polyphonic synths, DCO’s, memory slots to save patches, out with analogue, in with digital, sampling, SAS, FM….it just didnt seem to stop. 

 

And now a 7 year old piano gets an EX slapped on and a few new patches, same with Grandstage (add an X this time). King Korg NEO, Wavestate EX, I could go on. Most of these are simply new ‘premium’ sounds (usually pianos) or an expansion of software functions. In at least one case, existing hardware features you already own unlocked by paying!!

Seems to me that the last tech to make it into keyboards (and its likely coming) is AI. Not that I am a big fan, but its in everything at the moment and its not going away. Just a matter of time before one of the big three announces a ‘new’ or ‘innovative’ use of AI in a keyboard, and not just assistive features or random generation of sounds.

Other than that, have we reached a point where there is nothing new in the world of keyboards? Is it just a few extra sounds, more poly, change of interface etc?

If AI is coming, what could it actually do in a keyboard or workstation that would be ‘groundbreaking’ and be this generations ‘tech breakthrough’ like we remember from the DX7, D50, Fairlight etc?

[J.F.N.]

Considering the fact that the board would need permanent network access in order to provide "real" AI, I doubt we will see this become reality.

[Paul Woodward]

1 hour ago, J.F.N. said:

Considering the fact that the board would need permanent network access in order to provide "real" AI, I doubt we will see this become reality.

I own a few chinese 'retro games consoles' that cost around £50 and have IPS screens as well as wifi and bluetooth built in. The rate of development of this technology is incredible with new units (Anbernic, Retroid etc.) being developed all the time. Wifi integration costs pennies, cheap phones have had this for many years, and many Roland's offer this option through a dongle. When its a feature in a device costing a few pounds, do you really think its a stretch to put something similar in a board costing several thousand?

If you integrate an iPad with a board, especially an M class chip, you already have this feature (or will when iPadOS 18 releases soon), Logic already uses AI so why not a keyboard. Imagine what it takes today to release a 'Kronos', and by that I mean a jaw dropping do it all piece of tech that we all saw when it was first released. That was 13 years ago, just 3-4 years after the very first iPhone!!

 

You are assuming AI needs to generate sounds/mixes or whatever all the time, you might just use it to prepare sounds or tracks while at home or in the studio, to generate a new sound set, synthesise a recorded sound.....who knows.

 

Point is, what's left to put into the next flagship board once you have unlimited poly, multiple sound engines, the best keyboard, every controller imaginable and a big screen?

 

You might say we dont need all that anyway, but thats the question really...has keyboard development stagnated with only incremental features and working on weight reduction or price to feature ratio?

[J.F.N.]

6 minutes ago, Paul Woodward said:

I own a few chinese 'games consoles' that cost around £50 and have wifi and bluetooth built in. The rate of development of these types of unit is incredible with new ones being developed all the time and wifi integration costs pennies. Phones have had this for many years, and many Roland's offer this option through a dongle.If you integrate an iPad with a board, especially an M class chip, you already have this (or will when iPadOS 18 releases soon).

You are assuming AI needs to generate sounds/mixes or whatever all the time, you might just use it to prepare sounds or tracks while at home or in the studio, to generate a new sound set, synthesise a recorded sound.....who knows. When its a feature in a device costing a few pounds, do you really think its a stretch to put something similar in a board costing several thousand?

Point is, what's left to put into the next flagship board once you have unlimited poly, multiple sound engines, the best keyboard, every controller imaginable and a big screen?

You might say we dont need all that anyway, but thats the question really...has keyboard development stagnated with only incremental features and working on weight reduction or price to feature ratio?

 

What I mean is, if we are talking about AI onboard, it's not going to happen as that would require a permanent network connection until the day we can fit a supercomputer into a small enough box. But yeah, a companion app solution we will probably see soon, where the AI is used in one way or the other.

[Paul Woodward]

6 minutes ago, J.F.N. said:

 

What I mean is, if we are talking about AI onboard, it's not going to happen as that would require a permanent network connection until the day we can fit a supercomputer into a small enough box. But yeah, a companion app solution we will probably see soon, where the AI is used in one way or the other.

AI is the fastest developing tech in a long time and companies are falling over themselves to integrate it rather than be left behind. Im no AI evangelist, but its crept into my work in education and I have had to adapt to use it in certain situations. Flagship boards cost a lot more than an AI powered phone or tablet so I think at least one of the big companies must be eyeing up how to use it on a pro level before it becomes affordable to make it into consumer (cheaper) level boards.

Can you imagine recording a sound and then describing how you want it to be and AI does the rest? Hum the tune in your head and describe the style, arrangement, sound etc?

Even if its a companion app (as Yamaha are doing) and you then load sounds/performances into your board, its certainly feasible. It just takes a need/market for it and boom....

I think the keyboard will eventually become the 'vessel' as Arturia have done. This way the hardware can be suited to the playing/performing and the software has greater flexibility to 'suit your individual needs'....

[Paul Woodward]

59 minutes ago, J.F.N. said:

 

Until the day we can fit a supercomputer into a small enough box. 

Its already here. I have the Samsung S24 with AI built in. Only reason for being the size it is, is to accomodate the battery and screen, the 'supercomputer' chip is tiny and would fit in a smartwatch.

Im sure they said a similar thing about playing more than one note and being able to save your patches to memory back in the 70's 😉 

 

Remember when a megabyte of storage was science fiction.....now 1.5tb in something the size of a fingernail.

 

 

[J.F.N.]

37 minutes ago, Paul Woodward said:

Its already here. I have the Samsung S24 with AI built in. Only reason for being the size it is, is to accomodate the battery and screen, the 'supercomputer' chip is tiny and would fit in a smartwatch.

Im sure they said a similar thing about playing more than one note and being able to save your patches to memory back in the 70's 😉 

 

Remember when a megabyte of storage was science fiction.....now 1.5tb in something the size of a fingernail.

 

 

Hence my emphasis initially on "real" AI, the buzz word is indeed used in a lot of user centric simplified services (as the case also seems to be with that Samsung phone), however, this doesn't mean it's in fact AI.

[ProfD]

2 hours ago, Paul Woodward said:

Other than that, have we reached a point where there is nothing new in the world of keyboards? Is it just a few extra sounds, more poly, change of interface etc?

Yes...IMO, a little over a decade ago with the release of the Korg Kronos we reached that point where ever since nothing groundbreaking or jaw-dropping has been added to KBs.

59 minutes ago, Paul Woodward said:

Point is, what's left to put into the next flagship board once you have unlimited poly, multiple sound engines, the best keyboard, every controller imaginable and a big screen?

Nothing really.

 

Being optimistic, we could see:

 

1) Faster processor for boot and load times.

2) Better navigation to access sounds and features.

3) AI assistance in sound programming to overcome menu-diving.

 

As I mentioned in the Roland thread, IMO, manufacturer sound offerings have been exhausted. They are recycling the same sounds tweaked with effects.

 

Otherwise, as always, innovation will come from musicians reimagining how to use sounds in a musically useful way.

 

It's been happening in music production with sine waves being used for drum and bass lines and reversed pads for harmonic content. 

 

Manufacturers provide a musical tool. Every now and then, they manage to add a feature or functionality to hardware that makes it easier to play and produce music but rarely is it groundbreaking or jaw-dropping nowadays.😎

[Paul Woodward]

20 minutes ago, J.F.N. said:

 

Hence my emphasis initially on "real" AI, the buzz word is indeed used in a lot of user centric simplified services (as the case also seems to be with that Samsung phone), however, this doesn't mean it's in fact AI.

Interested in what you perceive as 'real' AI. Admittedly supercomputers should be used to gnerate cures for cancer etc. but we are talking realistically about generative AI. Last year is was smart eidting of images (removing your ex from a wedding photo) and generating essays to cheat at school. Now its images, models, full videos and music and many modern phones can do this. After all a decent PC spec is a quad core CPU and 16gb ram....octacore and 16gb ram is a flagship phone now. All of this in a SOC the size of your fingernail. I doubt any keyboard today could make use of that, in a few years the tech will be way more powerful.

 

Back to the original question though, if not AI then what is next?

[J.F.N.]

34 minutes ago, Paul Woodward said:

Interested in what you perceive as 'real' AI. Admittedly supercomputers should be used to gnerate cures for cancer etc. but we are talking realistically about generative AI. Last year is was smart eidting of images (removing your ex from a wedding photo) and generating essays to cheat at school. Now its images, models, full videos and music and many modern phones can do this. After all a decent PC spec is a quad core CPU and 16gb ram....octacore and 16gb ram is a flagship phone now. All of this in a SOC the size of your fingernail. I doubt any keyboard today could make use of that, in a few years the tech will be way more powerful.

 

Back to the original question though, if not AI then what is next?

 

AI in my perception is when you can instruct a machine to assist you with a non predictable or predefined assignment, without the need to provide the abstract it may need to solve this, and instead allow its machine learning system to provide this.

 

"Make me a set of scenes with all the sounds needed to play the keyboard parts of the top 15 songs on Billboard right now." etc.

 

Or any typr of automated tasks where analysis of abstract data in a non predictable way or context is used for AI decision making without any human interaction required.

 

[Radagast]

There are several things to consider.  First is economic conditions.  Not good.  Next is the continued and prolonged love affair with analog synths by so many people.  Not much incentive to see what new vistas can developed in the digital world other than in the plug-in arena or in iOS. Even then there are many softsynths that are modeling analog synths.  As long as the economic conditions suck like they do, companies are not going to venture into new territory with innovative ideas, in hardware.  AI is just the latest annoying use of a buzzword in order to sell things.  A few years ago, it was quantum dots in televisions.  We still don’t have real quantum dot televisions on the market, even though development is ongoing.  I would like to see a hardware synth that has real time additive synthesis/resynthesis with a large display screen with plenty of controllers and has MIDI 2.0, and Poly AT.  

[cphollis]

6 hours ago, J.F.N. said:

Considering the fact that the board would need permanent network access in order to provide "real" AI, I doubt we will see this become reality.

Not quite true.  As I write this, I'm running one of Meta's "llama" LLMs on my laptop under ollama, independent of any backend cloud processing.  It's "real" AI, and its taking up a substantial amount of memory.  I use it as a desktop assistant.  But it's a twin of the cloud version with the same training cutoff date.

 

Thinking ahead 3-5 years, it's not unreasonable to posit a slimmed-down AI assistant running on your DAW or dedicated keyboard. 

 

Maybe acting as a musical director to suggest program voicings and parts? 

Or a real-time audio engineer to give you a great mix all the time? 

How about virtual bandmates who never argue, and do exactly what you ask of them?

Or a professional critique of your playing technique, along with suggested exercises for improvement?

 

I think the tech will show up first on "performer" keyboards, then trickle elsewhere.  I think a new generation of music creation tools is upon us, and it's going to be fun.

 

 

[J.F.N.]

1 minute ago, cphollis said:

Not quite true.  As I write this, I'm running one of Meta's "llama" LLMs on my laptop under ollama, independent of any backend cloud processing.  It's "real" AI, and its taking up a substantial amount of memory.  I use it as a desktop assistant.  But it's a twin of the cloud version with the same training cutoff date.

 

Thinking ahead 3-5 years, it's not unreasonable to posit a slimmed-down AI assistant running on your DAW or dedicated keyboard. 

 

Maybe acting as a musical director to suggest program voicings and parts? 

Or a real-time audio engineer to give you a great mix all the time? 

How about virtual bandmates who never argue, and do exactly what you ask of them?

Or a professional critique of your playing technique, along with suggested exercises for improvement?

 

I think the tech will show up first on "performer" keyboards, then trickle elsewhere.  I think a new generation of music creation tools is upon us, and it's going to be fun.

 

 

 

Yes, you could indeed have a ready trained AI running locally in some context, that is actually a very good point. It would be limited and very specific indeed, which no matter what would be the case if it sits in a hardware machine.

[Reezekeys]

55 minutes ago, cphollis said:

Maybe acting as a musical director to suggest program voicings and parts? 

Or a real-time audio engineer to give you a great mix all the time? 

How about virtual bandmates who never argue, and do exactly what you ask of them?

Or a professional critique of your playing technique, along with suggested exercises for improvement?

 

I think AI is mostly going to inspire companies to market products that let people without talent make music they can post on TikTok or Instagram.

 

Also, why the disappointment when new keyboards don't keep adding fantastic new features? Maybe because people aren't happy with the music they're producing with their current gear and are hoping the "next thing" in keyboard tech will be the "magic bullet" that has them sounding like <insert your fav keyboardist here>. IMO there's way more than enough whiz-bang features in many of the existing keyboards to keep the most talented musician busy for a long time.

 

I'm ready to be surprised and maybe even delighted with what AI can do for me for certain tasks - but I have no illusions that some new keyboard with built-in AI will turn me into Herbie. That would require non-artificial intelligence!

[ElmerJFudd]

What’s the most advanced feature-wise music production keyboard on the market?   I’d say Yamaha Genos 2 or Korg Pa4x.  I’m not interested in either regardless of the amazing things they do with a single key or button press because the goal has always been to play the piano keyboard - melody, bass line, harmony, rhythm and dynamics, etc.  I personally also prefer performing with other musicians (even if they need to be paid).  
 

Those machines sound great and allow one person to sound like many - which is amazing and can be fulfilling and even lucrative (after you’ve paid it off) in the right hands.   
 

Anyway, there are many arranger keyboards out there at many price points.  I imagine those that incorporate AI will be on the high end when first made available.  Conceptually very interesting stuff.  But in practice, I’m happy to have grown up without AI and I hope that generations that come after us will find many reasons to ignore that AI exists - especially regarding subjects and activities they love to do 
 

 

[David Emm]

10 hours ago, Paul Woodward said:

Point is, what's left to put into the next flagship board once you have unlimited poly, multiple sound engines, the best keyboard, every controller imaginable and a big screen?

 

I'm with the crowd who sees the major problems as having been solved. Its now about refinement- which really isn't very sexy- and personal commitment to craft. The D-50 came about in part due to clever workarounds & the expense of memory, etc. at the time. Its hard to make that same lightning suddenly strike in a world of multi-gigabyte memory and color touch screens. There's no current challenge that demands that same level of problem solving and *creativity.* When you can now have a CS-80 or Memorymoog in your computer for $200 or less, you have to look pretty hard for a meaningful issue. 

 

If anything, I see a sensible trend towards 'boards that speak "computer" easily, such as the Roland RD line. Their Cloud is one of the smartest approaches I've seen in a while. I find AI pretty masturbatory so far, although I know that these things find their footing in due course. I'm personally all about responsive key actions and a dash of MPE interest, but if you want to see the future, look at the forest of jacks on the back of the Fantom. Welcome to the wild world of gozintas and gozouttas.       

       

[ElmerJFudd]

34 minutes ago, David Emm said:

 

I'm with the crowd who sees the major problems as having been solved. Its now about refinement- which really isn't very sexy- and personal commitment to craft. The D-50 came about in part due to clever workarounds & the expense of memory, etc. at the time. Its hard to make that same lightning suddenly strike in a world of multi-gigabyte memory and color touch screens. There's no current challenge that demands that same level of problem solving and *creativity.* When you can now have a CS-80 or Memorymoog in your computer for $200 or less, you have to look pretty hard for a meaningful issue. 

 

If anything, I see a sensible trend towards 'boards that speak "computer" easily, such as the Roland RD line. Their Cloud is one of the smartest approaches I've seen in a while. I find AI pretty masturbatory so far, although I know that these things find their footing in due course. I'm personally all about responsive key actions and a dash of MPE interest, but if you want to see the future, look at the forest of jacks on the back of the Fantom. Welcome to the wild world of gozintas and gozouttas.       

       

UI development that makes using the amazing features of these boards easy to use.   I’m interested in any board that sounds great, plays great and is intuitive.  

[AROIOS]

On 7/24/2024 at 12:31 AM, Paul Woodward said:

...have we reached a point where there is nothing new in the world of keyboards?...

We've reached that point 30 years ago. I can't think of ONE keyboard sound that's ground breaking after the early 90's. The incremental progress since then has been mostly little tweaks or longer/larger samples of sounds that existed before.
 

Take EP sounds for example, Rhodes/Wurlies aside, DX7II "Full Tines", RD1000/MKS-20 "EP1/EP2", M1EX "Super EP", 01/W "Dyno Piano", JD800 "Crystal Rhodes" were all unique, beautiful sounds that graced countless records, not to mention Robbie Buchanan's glorious TX816 stacked EP heard throughout the late 80's/early 90's. How many EP sounds after the early 90's were memorable in a similar fashion? Zero.
 

Same thing happened in the guitar world. There has been ZERO progress in guitar tones since the early 90's. Technology only made those sounds more accessible.
 

We talked about elements of music in another thread earlier. Other than "timbre", music pretty much peaked by the late 70's, and "timbre" itself peaked by the 90's, save for some developments in EDM.

[David Emm]

I just re-watched a video on the development of the Fairlight and it has "Heavy Lifting" written all over it. Both the basic technology and the manufacturing methods kept it so expensive, the team ran out of steam. Now, numerous boffins describe an instrument within a Mac/PC and out comes board design, parts sources and even marketing ideas. This is all on a desktop computer that has power multiples for which the Fairlight team would have all but killed. We're so tech-pampered now, I think one of the next advances will be recognizing it more fully. That's where MPE seems likely to come into its own. I'm waiting for a Steve Vai-level player to appear with a double Osmose rig, reshaping the game.

[ProfD]

23 minutes ago, David Emm said:

We're so tech-pampered now, I think one of the next advances will be recognizing it more fully.

Right.  Tech pampered to the point of analysis paralysis. No more deep-diving and exploiting every nook and cranny of the gear. 

 

After reading specs or watching YouTube demos or 15 minutes of hands-on time it's back to dreaming up and wishing for other sounds, features or functionalities in version X.0.

 

No amount of cheese or peanut butter will make it a better mousetrap.😁😎

[stoken6]

13 hours ago, ProfD said:

Tech pampered

I would argue "tech restrained" as well as "pampered". I'm referring to the big text, big buttons, designed for touch rather than mouse, limited configuration, settings buried in a menu with "settings", "options", "account", "other". 

 

Back in the day, a Windows or Mac application would have a big "preferences" menus with all sorts of micro-personalisable options. These days, that's the purview of open-source solutions (look at NotePad++ for example). The "big tech" applications don't give us the control - first example from the top of my head: Youtube doesn't allow you to sort videos oldest-to-newest.

 

Cheers, Mike.

[Paul Woodward]

In some ways I do miss 'squeezing' the most out of tech because that's what you had. I had a Roland W30, Korg DS8 amongst others and I knew them inside out as I kept them for years. I started with a Poly 800 II and learnt it inside out to get the sounds I needed. Remember trying to recreate the sound of breaking glass with white noise before sampling was anything like affordable.

Nowadays I play a workstation or synth, get bored, try another and so on before ever really digging in. Only the feel, sound and nuance of playing a good piano weighted board retains my interest. Everything else sound wise can be done in a DAW, or even on an iPad.

Admittedly, much of that is down to me, but it's all so accessible and cheap now that it feels like the music equivalent of fast food.

[ProfD]

2 hours ago, Paul Woodward said:

In some ways I do miss 'squeezing' the most out of tech because that's what you had.

Those limitations sparked creativity and exploration.

 

2 hours ago, Paul Woodward said:

Nowadays I play a workstation or synth, get bored, try another and so on before ever really digging in.

That's because it already contains a reasonable facsimile of everything. 

 

There is no real reason or incentive to create a sound from scratch in a workstation. 

 

The only thing left to do is play and/or produce music. 

 

2 hours ago, Paul Woodward said:

Only the feel, sound and nuance of playing a good piano weighted board retains my interest. 

As I mentioned in another thread, for most of us pianists and organists, all roads lead back to that instrument/sound.  

 

Synths and workstations are an excellent distraction and/or foray into sound design but a musician ultimately...plays.😎

[Radagast]

AI is a buzzword.  The real transformation will take place if quantum computing ever enters mainstream tech.

[ElmerJFudd]

[AROIOS]

8 hours ago, stoken6 said:

...Back in the day, a Windows or Mac application would have a big "preferences" menus with all sorts of micro-personalisable options. These days, that's the purview of open-source solutions (look at NotePad++ for example). The "big tech" applications don't give us the control - first example from the top of my head: Youtube doesn't allow you to sort videos oldest-to-newest...

There has been a brain disease among UI/UX idio... ehem, engineers, to make PC software interfaces resemble mobile touchscreens: ever-shrinking scroll bar, replacement of pagination with infinite-scroll, fonts only legible on high-density screens... The list goes on and on.

That reminds me of the lesson from Joel Schumacher's "Falling Down": idiots will justify their existence through constant changes, even if they are unnecessary or regressive.

[Paul Woodward]

1 hour ago, ElmerJFudd said:

Hmmmm

https://www.korg.com/uk/products/software/korg_gadget_vr/