Rumors: Yamaha Montage successor

[DmitryKo]

13 hours ago, jerrythek said:

before Yamaha licensed the technology (not purchased it), it was running on mainframe computers. Not running on 57 chips

 

Yes, he actually said that the Yamaha GS1 had around 50 chips, not the Stanford implemenation, and thar's why it was much more expensive than the DX7 ($18,000 according to SOS). This was in an 1985 interview from the Aftertouch Magazine. Thanks for correcting me.

 

13 hours ago, jerrythek said:

Korg doesn’t “have to” use Rasberry Pi to do FM, they have done various forms of FM on the Oasys PCI card, and  the Kronos. But to relate to what I think is your point, in todays world you don’t need to develop custom chips to do synthesis and DSP.

 

It's about the economy of outsourcing a third-party microcomputer vs. replicating the FM algorithms in your standard DSPs used across the entire product range. Yamaha Reface DX/CS were implemented with the SSP2 microcontroller, which is typically used as effect processor and USB audio interface and includes a single-core SH-2 host processor.  Korg had to port over their software-only implementation to the Raspberry Pi, a full blown microcomputer board that hosts a quad-core ARM Cortex-A processor.

 

Korg chose to go full software because their original OASYS project, a virtual modeling synthesizer keyboard implemented with a dozen Motorola 56300 digital signal processors, was prohibitively expensive - the actual keyboard would be priced $15,000. The released version in the form of the OASYS PCI card eith five 56K DSPs didn't sell much at MSRP of US$8,000, considering the limited processing power available for the synth engines (see Sound on Sound reviews of OASYS PCI and OASYS keyboard). Several years before, the Yamaha VP1 didn't move for the same reason of being much too expensive, in addition to its very unconventional sounds. VP1 did use Yamaha proprietary DSPs though which would be cheaper to source, but there were 32 of them in total, two on each of 16 voice boards - so the required processing power was there, but at the MSRP of US$30,000.

 

At least Korg and Yamaha still enjoyed good sales of DX-series and M1/i3 derivates as well as their other product lines, so they could cover the development costs, otherwise these undoubtly innovative projects could have easily sunk them down, just like similar product failures resulted in the demise of Sequential, E-mu, and Ensoniq.

 

12 hours ago, AnotherScott said:

 

Jupiter 8 made a lasting impact artistically, but had nowhere near the market impact of the DX7 or M1. Roland's equivalent to that would have been the D50.

 

12 hours ago, jerrythek said:

And the TR-808 was a failure in its lifetime. It only gathered influence when it could be had for so cheap after its discontinuation that young music makers could afford them and started using it to make dance music/EDM etc. I wouldn’t call that Roland having/making a breakthrough instrument.

 

Yes, Roland was arguably more successful with Roland D-50 and derivative Sound Canvas modules, JV/XV modules, and electronic V-Drums rather than professional digital keyboards / workstations; they also had well regarded VS series hard disk recorders, Boss guitar pedals and effect processors, and VM/M Series digital mixers. And their real break-through "product" would of course be the MIDI protocol.

 

However this even better illustrates the point how current "cult classic" and sought-after synthesizers didn't really sell well in their time, and some like the TR-series even became "thrash bin" parts. A smaller company would have much more trouble if their new product line failed like this - which is what happened to one-hit wonders like Oberheim OB-X/OB-xa,  Prophet 5, and LinnDrumm. Their sales were never goood enough to sustain operations, so the respective companies all went down trying to develop or market their next product, the Linn 9000, the Oberheim Matrix, and the Prophet 3000.

[Docbop]

What no cigarette/blunt lighter or cup holder, I thought this was a Pro board???

[jerrythek]

18 hours ago, DmitryKo said:

The released version in the form of the OASYS PCI card eith five 56K DSPs didn't sell much at MSRP of US$8,000, considering the limited processing power available for the synth engines (see Sound on Sound reviews of OASYS PCI and OASYS keyboard).

The card listed for $2,000. I will offer to you, that a big part of the failure of the card was how I (and others positioned it). We marketed it as an I/O card that had some DSP and synths and effects, and it was viewed as way too expensive as an I/O card. I regret to this day that we did not make a clearer case for it being a suite of synths and effects that could run inside of your computer, and the needed I/O to support them. I'm not saying that would have made it a hit, but we certainly did ourselves no good with the approach that we took. Mea Culpa.

 

But that development was part of the path to developing algorithms that got used in subsequent products, now including the Kronos and Nautilus. And who know what in the future. 

 

18 hours ago, DmitryKo said:

Yes, he actually said that the Yamaha GS1 had around 50 chips, not the Stanford implemenation, and thar's why it was much more expensive than the DX7 ($18,000 according to SOS). This was in an 1985 interview from the Aftertouch Magazine. Thanks for correcting me.

Ah, I see. Good info - thanks.

 

18 hours ago, DmitryKo said:

However this even better illustrates the point how current "cult classic" and sought-after synthesizers didn't really sell well in their time, and some like the TR-series even became "thrash bin" parts. A smaller company would have much more trouble if their new product line failed like this - which is what happened to one-hit wonders like Oberheim OB-X/OB-xa,  Prophet 5, and LinnDrumm. Their sales were never goood enough to sustain operations, so the respective companies all went down trying to develop or market their next product, the Linn 9000, the Oberheim Matrix, and the Prophet 3000.

Part of this is true: Many, many of the now cult classic/beloved products did not sell well in their original time on market. But you are simplifying the timeline of that you purport to be "one-hit wonders". From the OB-X/XA Oberheim had decent success with the OB-8, the Xpander and other models before they got to the Matrix. Sequential had the Pro-One, the Prophet 600, Prophet VS and others before their fall from grace. None of these companies were "one and done".

[DmitryKo]

41 minutes ago, jerrythek said:

The card listed for $2,000.

I stand corrected.

 

41 minutes ago, jerrythek said:

 a big part of the failure of the card was how I (and others positioned it). We marketed it as an I/O card that had some DSP and synths and effects, and it was viewed as way too expensive as an I/O card. I regret to this day that we did not make a clearer case for it being a suite of synths and effects that could run inside of your computer

Yes, it seems like most people learned about this board long after it was discontinued, but better marketing wouldn't change much because GigaStudio and VST 2.0 were released a few years prior.

 

41 minutes ago, jerrythek said:

you are simplifying the timeline of that you purport to be "one-hit wonders".

None of these companies were "one and done".

Maybe I am, but ultimately they had a few very succesful products, and the following models were far less successful, while they overstretched themselves by either increasing complexity and cost, or trying to move into a whole new domain, which drained their resources and income. For all I know, Sequential, E-mu, Ensoniq, and Alesis all went down while trying to develop or market their own hard disk recording systems.

[jerrythek]

8 minutes ago, DmitryKo said:

42 minutes ago, jerrythek said:

 a big part of the failure of the card was how I (and others positioned it). We marketed it as an I/O card that had some DSP and synths and effects, and it was viewed as way too expensive as an I/O card. I regret to this day that we did not make a clearer case for it being a suite of synths and effects that could run inside of your computer

Yes, it seems like most people learned about this board long after it was discontinued, but better marketing wouldn't change much because GigaSampler/GigaStudio was released two years prior.

Well... streaming samplers and DSP synths and effect were two very different parts of the market/user-base, so I don't agree. But anyway, I'm not arguing your larger-picture issues so much, I just don't like to see wrong info sit out on the Net uncorrected. People will come back to posts like these later, re-quote them, and that's how misinformation develops. Carry on!

[DmitryKo]

On 9/26/2023 at 4:22 AM, jerrythek said:

Well... streaming samplers and DSP synths and effect were two very different parts of the market/user-base, so I don't agree.

 

This thread, like similar threads on the other forums, clearly shows that most people wouldn't tell a sample-based instrument from real analog or VA 

 

It would be hard to market the Korg OASYS PCI as a synthesizer for project studio users, when they were already accustomed to using multi-timbral polyphonic General MIDI-compatible synthesizers. There was a lot of choice, starting from cheap SoundBlaster AWE32 and Live! cards, based on sample-based 'wavetable' synthesis provided by EMU8000 and EMU10K chips, with lots of SoundFont sample libraries available including classic E-MU sounds, to ROMpler modules such as Roland Sound Canvas series like the SC55/SC88, Yamaha XG series like MU-100/MU-128/MU-1000 and SW1000XG, and E-Mu Proteus series like Proteus Ultra, Morpheus and Proteus 2000, to professional digital synthesizer keyboards from Korg, Ensoniq, and Yamaha.

 

Very few users in that market niche ever used analog synthesizers because of their limited polyphony and timbrality, so the OASYS PCI wouldn't really fly as a virtual modeling synthesizer when GigaSampler/GigaStudio with sound libraries for all the classic analog synths would cost you an order of magnitude less. A similar Yamaha SW1000XG card, which could also host AN/VL/DX/PF/XG plug-in cards like the MU series, S-30/S-80 and the classic Motif, was not popular either.

 

So I guess Korg was probably right in trying to market the OASYS PCI as a multitrack recording and effects processing card, but this segment was already captured by emerging manufacturers like RME, ESI and MOTU which offered more analog channels with external breakout boxes, and effects processing was starting to be handled by VST and "DirectX" plugins and software mixing engines inside computer-based DAWs, so the few available hardware DSP cards like Yamaha DS2416/AX44/AX16-AT and Creamware Luna/Pulsar were outshadowed by Cubase/Nuendo and Cakewalk/Sonar which supported hundreds of realtime audio tracks.

 

[DmitryKo]

The specs sheet and pictures of the Montage M series have leaked online, so I guess they will be posted everywhere very soon.

 

 

Looking back to my "wild speculation" list, it does seem Yamaha actually did tick at least some of the boxes 😁 In particular, there is a 16-voice AN-X  engine, a total polyphony of 400 voices (128 + 128 for AWM2 user/preset ROM, 128 for FM-X, 16 for AN-X), and double the flash memory capacity with preset ROM of ~10 GBytes (LPC compressed) and user ROM of 3.8 GBytes (uncompressed).

 

It says something about "128 element AWM2 architecture" that enables "high definition emulative sounds like piano, strings, and drums" - I did expect more Elements to enable additional articulations, but 128 seems a lot, so this needs clarification in terms of actual hardware capabilities.

 

Maybe some of these emulations are programmed using additive (speech) synthesis techniques - so the original samples would be analyzed and and split into several Elements, then reconstructed in realtime using wavetable, spectral, formant, and/or vector synthesis controlled by additional Expanded Articulation Control (XA Control) sources that include even more player articulations and physical knobs/switches/ribbon controller actions? 

 

There are no details on AN-X oscillator/filter/EG types and features yet, though emulation of other synth architectures is probably not included. Polyphony configuration implies they still use several tone generator chips that stay on their old semiconductor manufacturing processes.

 

 

The Montage M8X has a "GEX weighted action with Polyphonic Aftertouch" and "enhanced key repetition behavior that is typically found on grand pianos" - this looks like a graded hammer action with individual escapement sensors and pressure sensors for each key.

The Montage M7/M6 have standard FSX action.

 

The monochrome 512 x 64 px quick edit LCD can show quite a lot of editing parameters in the graphical form. 

There is a Portamento button and Portamento Time adjustment knob along the pitch bend wheel.

It looks like some button groups use two different colors depending on the selected function, and two different light intensities to indicate the active button. The ribbon also has a backlight to indicate the 5 separate zones.

 

 

There is some form of 16-track song/pattern sequencer, but no word on MIDI 2.0 capabilities over USB, MIDI Clip File (SMF2) format, or support for Network MIDI 2.0 protocol over Ethernet - I'm cautiosly optimistic that it could be implemented/enabled with later firmware updates, when the new MIDI 2.0 driver stack for Windows 10/11 is ready for a wide release by Spring. 

 

I hope they would also support MIDI CI Profiles, Property Exchange and UMP "Function Blocks" to enable external controllers for polyphonic aftertouch and other per-note controllers, as well as analog-style control surfaces with a bunch of realtime knobs/sliders/buttons and built-in displays/LEDs.

 

 

 

On 9/2/2023 at 3:04 AM, DmitryKo said:

Why didn't anyone speculate on Montage m8X specs yet? Here's my take:

 

• NW-GH natural wood graded hammer action
• AWM-X, FM-X and AN-X synth engines
• AWM-X: 16 elements, includes VRM (virtual resonance modeling) for emulation of piano/clavi*/Rhodes and plucked string instruments, and reed (wind, brass) instruments. Articulation switching with continuous controllers and/or Super Knob 
• FM-X: 8-operator FM synthesis, formant synthesis and vocoder/harmonizer modes
• AN-X: virtual analog/wavetable synthesis, 3 morphing oscillators (40 types, 200 waveforms), 3 morphing LFOs, 2 morphing filters (20 types), 3  modulation envelope generators. Parameters/modes to emulate classic analog synths (Yamaha CS-80, MiniMoog/MemoryMoog, Prophet-5, Oberheim OB-Xa, Roland Jupiter-8, Juno-106, ARP Odyssey)
• Plyphony 320 voices (240 AWM-X + 64 FM-X + 16 AN-X)
• MIDI 2.0 Protocol with per-note articulation etc., MIDI-CI Profiles 
• Analog Devices A2B MIDI 2.0 physical interface over DIN5 connectors
• Phrase sequencer, MIDI 2.0 Clip File format (with key/time signature, chords, lyrics etc.)
• 20 out / 4 in class-compliant USB-Audio interface
• Gigabit Ethernet port, Network MIDI 2.0, 32-channel Ethernet audio
• External 'programmer' control surface, knobs/sliders/switches/displays organized in OSC/LFO/FLT/ENV groups for AN-X parameter tweaking, auto-configuration with MIDI-CI Property Exchange

Nah, who am I kidding... nothing of the above is going to happen. In a typical Yamaha fashion, hardware specs will be marginally better in comparison to the original Montage, though total sample ROM could be increased to 8-10 GBytes (after decompressing) and user ROM to 2-4 GBytes (uncompressed).

 

 

 

 

[Keyboardplayer]

Yamaha Montage M Specifications Leaked...

 

You can read all about it here: https://yamahamusicians.com/yamaha-montage-m-specifications-and-pricing/

 

Only the M8x has polyphonic aftertouch. Not sure what Yamaha's reasoning was in leaving out polyphonic aftertouch on the M7/M6. 400 notes of total polyphony, but again, it's divided up into sections with 128 notes at most and 16 notes for the AN-X analog synth engine. It is supposed to now have a fully-baked Sequencer, and if that's accurate, it would be a bonus. Street prices should be somewhat lower than MSRP depending on where you purchase it from. If you already own the original Montage, you might be happy with what you have. But there are compelling reasons to upgrade, especially to the M8x, which has the polyphonic aftertouch. Not sure how much the M8x weighs since it wasn't disclosed in the specs. Oh, and no tone-wheel organ engine, which makes sense since the Montage M still only has 8 sliders, although Yamaha could have given it nine digital sliders used with the touch screen but, alas, not so? Remember, folks, Korg had nine sliders and nine sound engines on the Kronos/2. Suds Clavier in France thinks Korg will release a Kronos 3 in the near future? Time will tell. PS: I wanted to add that the 128 note polyphony for the Preset and User sections might be able to be combined, giving a total of 256 notes, but again no guarantee. Everything will be out in the open on October 9th if that's the actual launch date.  My sources say it is.

 

 

 

 

[AnotherScott]

44 minutes ago, Keyboardplayer said:

Not sure what Yamaha's reasoning was in leaving out polyphonic aftertouch on the M7/M6.

 

Maybe their poly AT implementation was not adaptable to their non-hammer action (e.g. if it's still the FSX), maybe a non-hammer poly AT action is still on the drawing board. (OTOH, it could also simply be the common factors of cost and product differentiation.)

[UnderGroundGr]

I believe that all new synths/workstations from all companies must have polyphonic aftertouch nowadays...

[AnotherScott]

12 hours ago, DmitryKo said:

It says something about "128 element AWM2 architecture" that enables "high definition emulative sounds like piano, strings, and drums" - I did expect more Elements to enable additional artuculations, but 128 seems a lot, so this needs clarification.

 

Here's my guess... The existing architecture supports "multi-part single instruments" of up to 8 Parts, each with 8 Elements, for a max of 64 elements. One of the knocks on the Montage has been that you can only play up to 8 sounds at once from the keyboard (even though the instrument is 16-part multi-timbral). If they have removed that limitation such that you can play all 16 Parts from the keyboard, with no other change, that would allow you to create single instruments of up to 128 elements.

 

The likely trade-off is that you would not have seamless sound switching available if either the part you were switching from or to had more than 8 Parts. This would be an extension of what they did when they came out with the MODX, which supported the same 8 Parts as the Montage, but did not offer SSS unless the parts you were switching from and to both had no more than 4 Parts.

[DmitryKo]

 

3 hours ago, AnotherScott said:

If <...> you can play all 16 Parts from the keyboard, with no other change, that would allow you to create single instruments of up to 128 elements.

 

The likely trade-off is that you would not have seamless sound switching available if either the part you were switching from or to had more than 8 Parts.

 

It seems plausible. Seamless Sound Switching behaviour would depend on the data connections between the multiple Tone Generator chips and the ONFi flash memory chips - considering the new TG configuration, there could be enough memory bandwidth and internal sound processing blocks to enable SSS even for Performances with 16 Parts. That's why I've edited my post to say that the 128 element quote needs clarification "in terms of actual hardware capabilities".

[Jose EB5AGV]

51 minutes ago, Keyboardplayer said:

If you already own the original Montage, you might be happy with what you have

 

This! 👍🏻

[Nathanael_I]

It's a beautiful, deep, and powerful flagship board.  It can do almost anything that needs to be done in one package.  With the expanded sample memory it can do what the Kronos could - become anything you've got in the studio.  Price and weight are in-line with a premium instrument.  Anyone who commits to it will find what they need, I expect.  As a canvas?  It's a big one!   I'm not close to exhausting my new Stage4, but there's no question this would do anything I need.  The current refresh of the big platforms is quite impressive.  It's never been easier to have great sounds in one keyboard. 

[CyberGene]

A hammer action with poly AT is world's first, errr… after the CS80, that is. Kudos to Yamaha. 

[Anderton]

On 9/25/2023 at 7:37 PM, jerrythek said:

I regret to this day that we did not make a clearer case for it being a suite of synths and effects that could run inside of your computer, and the needed I/O to support them. I'm not saying that would have made it a hit, but we certainly did ourselves no good with the approach that we took. Mea Culpa.

 

Don't be too hard on yourself. That's how Creamware marketed SCOPE, and it didn't do all that great (even though it should have).

[pjd]

2 hours ago, Keyboardplayer said:

Yamaha Montage M Specifications Leaked...

 

You can read all about it here: https://yamahamusicians.com/yamaha-montage-m-specifications-and-pricing/

 

 

I'm taking all of this with a big grain of salt. Some of the numbers don't add up for me, among other observations.

 

We'll know all the facts soon enough. The first thing I look at is the list of waveforms to see what is actually new on the AWM2 side.

 

Silly me -- pj

 

[AnotherScott]

43 minutes ago, CyberGene said:

A hammer action with poly AT is world's first, errr… after the CS80, that is. Kudos to Yamaha. 

 

And then the Kurzweil MIDIboard. 🙂

 

 

[CyberGene]

3 minutes ago, AnotherScott said:

And then the Kurzweil MIDIboard.

Ha, didn’t know about that. Thanks. Seems Kurzweil used to be a very innovative company. 

[jerrythek]

30 minutes ago, Anderton said:

On 9/25/2023 at 8:37 PM, jerrythek said:

I regret to this day that we did not make a clearer case for it being a suite of synths and effects that could run inside of your computer, and the needed I/O to support them. I'm not saying that would have made it a hit, but we certainly did ourselves no good with the approach that we took. Mea Culpa.

 

Don't be too hard on yourself. That's how Creamware marketed SCOPE, and it didn't do all that great (even though it should have).

Yeah, but we are (were?) Korg, dammit!  🙂 We likely could have made more inroads. But it's all water under the brdige... I reflect on my successes and non-successes from time to time. Always good to learn from the past. But I sleep just fine.

 

 

[Mr -G-]

1 hour ago, CyberGene said:

A hammer action with poly AT is world's first, errr… after the CS80, that is. Kudos to Yamaha. 

And after the DX1, too...

[pjd]

35 minutes ago, jerrythek said:

 I reflect on my successes and non-successes from time to time. Always good to learn from the past. But I sleep just fine.

 

 

Good to hear. 😃 I still get crazy work dreams. I'm on a strange campus, I have to teach a class in 10 minutes, no PPT deck and I don't know where the classroom is...

[DmitryKo]

 

4 hours ago, pjd said:

Some of the numbers don't add up for me, among other observations.

 

The polyphony does add up for me. There are probably two TG chips, each connected to a separate 4 GB ONFi flash memory chip - therefore we have 4 GB for preset ROM (LPC compressed samples, so 10 GB "when converted to 16-bit linear format"), and 4 GB for user ROM (uncompressed PCM samples).

 

Assuming one of these TG chips is the same as used in the MODX+, it should be capable of processing 128 AWM2 voices and 128 FM-X voices. Then the second TG chip would produce 128 AWM2 voices and 16 AN-X voices - or, if these are actually two identical TG chips, each could have 64 FM-X voices and 8 AN-X voices.

 

(This gives more credence to the notion that the MODX+ could be capable of hosting the AN-X engine, but I would say the chances for this are slim as Yamaha is not typically known for making such major changes to an existing product).

[UnderGroundGr]

2 hours ago, CyberGene said:

A hammer action with poly AT is world's first, errr… after the CS80, that is. Kudos to Yamaha. 

CS80 had not hammer action keyboard...

[CyberGene]

5 minutes ago, UnderGroundGr said:

CS80 had not hammer action keyboard...

I think I’ve seen it's weighted with heavy weights, so pretty much qualifies as a hammer action since it’s mass, hence inertia, that requires force to accelerate, rather than springs that have no inertia and require only initial force equal to the spring force and no inertial resistance. I might be wrong. 

[CyberGene]

P.S. Hard to find pictures of the action… I found a few with low quality and it seems what I thought are weights might actually be spring wires 🧐

[Paul Woodward]

Hmmmm. https://pianocraft.net/whats-the-difference-between-weighted-keys-graded-hammer-action/#:~:text=Weighted keys and graded hammer,the right one is essential.

[timwat]

2 minutes ago, CyberGene said:

I think I’ve seen it's weighted with heavy weights, so pretty much qualifies as a hammer action since it’s mass, hence inertia, that requires force to accelerate, rather than springs that have no inertia and require only initial force equal to the spring force and no inertial resistance. I might be wrong. 

 

To my memory, the CS80 didn't feel anything like a hammer action keyboard (which, as suggested above, didn't really hit the market until about 5-6 years later, I think - CS80 in 1977 and DX1 in early 80's). It had a quite different throw depth, different resistance sensation under the fingers, and played very much like it's own beast...not at all like a piano, to me. I gigged one for a few golden years back in the early 80's.

 

I don't remember actually playing piano-like hammer action keys until I played the K250 and Prophet T8, both of which felt astounding to me at the time. Sequential's T8 actions I'm told were purchased by New England Digital and used in the Synclavier (never had the opportunity to play one of those, however).

[CyberGene]

2 minutes ago, timwat said:

I gigged one for a few golden years back in the early 80's.

Wow, that must’ve been great. Well, unless you have to move it, I guess 😀 Thanks for the first-hand experience and nice memories 👍🏻

[timwat]

It was a two-keyboard player prog band, and we gigged a CS80, CP70B and Prophet 10, all in Anvil road cases (quite a schlep, even with two dedicated roadies).

 

Was too short a time with some really great musical and personal memories, but like a lot of life those days are all in the rear view mirror now.