They didn't lose anything. Things that stay the same allow for more others things to be different (enhanced/improved).

No FM-X update is the biggest disappointment for me in this synth that I otherwise like very, very much.
Since most of the IdeaScale suggestions on FM-X did show up in the HALion software implementation, my speculation is that it has to do with the tone generation chips.
Probably the SWP70 can't accommodate any of those changes.
It is what it is, funnily enough the Montage is superior to HALion in many ways (modulation and effects in the first place).

[quotePost id=123797]No FM-X update is the biggest disappointment for me in this synth that I otherwise like very, very much.
Since most of the IdeaScale suggestions on FM-X did show up in the HALion software implementation, my speculation is that it has to do with the tone generation chips.
Probably the SWP70 can't accommodate any of those changes.
It is what it is, funnily enough the Montage is superior to HALion in many ways (modulation and effects in the first place).[/quotePost]
I hear ya on this one! 😉
If the new M8x was in the cards for me & I was looking forward to it as my new synth, this would have been a bit of a "let down" for me. And to a much lesser degree, seeing that this new Flagship synth still only has 128 note FM-X polyphony, that too would have been a 'tiny' bit of a bummer too, but mainly just because it still only has the same FM-X power as the MODX+, and with a new Flagship I would have expected a bit more horses. Although in reality, thinking about it, I tend to use between 40%-60% FM-X PARTs in most of my Performances, which probably doesn't come that close to the FM-X 128 note polyphony wall anyway. I'm not even sure if FM-X Polyphony is based on number of notes or on number of total Elements used (times) number of notes played..!? I just know that I've never hit the wall yet & not had a single note drop, even in my Performances that have all 16 PARTs fully sequenced with 60% of them being FM-X PARTs...
I was one of the many who added an Ideascale idea for FM-X improvements...although mine was fairly basic though, as I only asked for more algorithms that utilized operators 1 and 2 in more stacks instead of having them as single carriers with no modulator. And I also Up Voted everyone else's FM-X ideas on Ideascale.

But maybe it's not an SWP70 issue, and instead it's something that Yamaha plan to implement in a future OS update..!?
I definitely wouldn't rule out FM-X enhancements from future updates, as Yamaha did a lot of that after they released the Montage classic, whereby there was no onboard sequencer initially until update 3.0, and their whole roadmap/design model was to add more features in later OS updates...
Maybe it's a timing thing, whereby they only had enough software programming, engineering and sound design development time & money budgeted for all the enhancements that the Montage M currently has, and they deemed those features most important to include up front over FM-X improvements. I don't hear nearly as many people talk about FM-X on forums as much as they do about wanting improvements with Sequencers, AWM2, VA, PolyAT, Effects, kitchen sink, etc...
Perhaps Paul from Sand, Software and Sound can enlighten us on FM-X improvements and the SWP70's..? Do we even know at this point what tone generators are in the Montage M..?

If Yamaha had included a newly enhanced FM-X engine 'up front' as you & I would hope for, and they left out the new Rotary Effect for organs &/or the new AWM2's enhancement of 128 Elements per PART, etc., maybe the majority of synth enthusiasts wouldn't have been as impressed with enhancements to FM-X as they would the other things..!? I mean, as much as I would personally love to see improvements in FM-X over many other things, I suspect the majority don't, and I can see Yamaha deeming their new enhancements of the Organ rotary effect, the AWM2 PART enhancements for things like pianos & strings in a single PART, as well as the AN-X engine, as the most important enhancements for bragging rights in all the new demos, reviews, etc.

Maybe one thing Yamaha have learned a lesson on from the Montage classic & the YC, was that whatever features you decide to include up front/out of the gate in a new hardware synth that you release, is what all the major reviews & especially the comparative reviews with other synths/keyboards will be on YouTube for many years when new customers do searches looking for what the new synth can do and the differences compared to other products. I think Yamaha lost out on the Montage classic to a degree by not including the sequencer in it at the time of release, because the forums were polluted with talk about that as the biggest downside compared to other synths/workstations, Ideascale ideas on this were through the roof, and there really aren't as many major comparative reviews on YouTube (or other), showing off the Montage with OS 3.0 (or newer) against the competition's products. Most of the ones that still show up in searches are with the Montage (pre-OS 3.0) being compared...
Same goes for the YC. Many of the major comparative reviews with competing products have been mostly with the YC prior to update 1.2 (whereby they enhanced the rotary effect the way it should have been from day 1) ... So the damage was done, and now if you search for reviews comparing the Nord, Yamaha YC, Hammond, etc., for organs, too many are pre-update 1.2 on the YC. Having said that, there are a few new comparative videos out now with the YC having update 1.2 or newer, but I think the perception hasn't changed as much for the YC against the Nord or Hammond SK because of the damage from all the initial reviews. It will take time to change people's perceptions on the YC now, especially against the Nord and Hammond, but maybe a YC+ at some point..!?
So put your best foot forward, then add features later in OS updates that aren't quite as important to most people as the ones you include out of the gate! Yamaha needed to get this right up front, and I think for the most part have done this with the Montage M, and left room for adding those improvements later!

So, I think there is hope for the future regarding FM-X improvements on the Montage M 😉

So put your best foot forward, then add features later in OS updates that aren't quite as important to most people as the ones you include out of the gate! Yamaha needed to get this right up front, and I think for the most part have done this with the Montage M, and left room for adding those improvements later!

I think that is a GREAT way to express it.

There are so many tradeoffs that have to be made when deciding what to include.

New sound engine? Or enhance an existing one?

They added a new one, AN-X, and a lot of people on commenting that it ONLY has 16 notes of polyphony. When 16 is actually at the top of the heap for analog engines.

They enhanced an existing one, AWM2, which can now define 128 elements per part. Sure to overall polyphony hasn't changed but you now have a lot more flexibility about how to use elements and parts.

An FM-X part has 8 operators period and they generally are used for a single sound. You don't typically use a single operator to make a sound; you use multiple ops. They aren't split up into 3 ops for this and 5 for that the way AWM2 elements might be since each element is really a complete sound system with a ready-made waveform to support it. An operator is just ONE piece of the base sound for a part.

I shudder to think of the difficulty Yamaha likely had deciding to add the rotary effect that people had been clamoring for.

Think about how different that one effect is from the rest of the architecture. You can only use it for part 1. That is going to get a lot of air time even though only Yamaha knows why that was necessary.

It essentially uses both effects for the part. Think about why that might be necessary. There were some pretty big architectural anomolies but they decided to 'shoe horn' it into the existing architecture anyway. You can bet that there are insiders just waiting to say 'I told you so' if any issues crop up.

The loudest 'complaint' I've heard so far is that the 6 and 7 don't have Poly AT. That 'lack' will probably affect more people's purchased decision than anything else including no FM-X enhancements.

I'm not coming out of 'roadie retirement' to lug that 62 lb (without a hard case) M8X from gig to gig. A 6 or 7 at 32 lb is at least manageable.

But Paul probably put it best in his comment in the GEX thread:

Could Yamaha make 61- and 73-key versions? Yes. I suspect someone ran the numbers including the manufacturing cost of synth-pitch sensor systems and the suits didn't like the numbers. Plus, some suit might have set a Fall 2023 release date. Ship or die. 😮 🙂

There are at least 2 pretty good indicators that this was kicked out the door before they would have liked.

1. There was an OS update available just a few hours after release. Certainly the OS used in the release models had to be set in stone a month or more ago in order to get units ready to ship. So anything they found and fixed after that point was a suitable impetus for an update.

2. The Data List doc is entirely missing both of the MIDI details. This includes all of the tables that show how to do data exports, send program change info and do SysEx.

I'm sure that there was a chorus shouting that the docs weren't done yet.

That 'Ship or die' has a pretty solid ring to it. As a long-time software developer My ears have rung with that same sound many times in the past.

My only reservation is that, as I said, most FM-X feature requests form IdeaScale have been developed and have been implemented.
In HALion.

The code is already done.

Just as they improved on AWM2, it should have been even easier to include FM-X code that was already there.

I'm not expecting any future improvements to FM-X since I suspect they're simply not possible.

I watched quite a number of the launch videos, and they say always the pretty much same thing: "we have this completely new analog engine and the sample engine which has been improved a lot. There's also the FM-X engine which was in the previous Montage."

Even if only for marketing purposes, they would have wanted to be able to say "the FM-X engine, also vastly improved compared to the previous generation".

(Of course, to be proven wrong on this would be the most pleasant surprise).

I still think the M is incredibly awesome and it'll be hard to resist getting one.

Not only FMx but where are the expansion "modular" packs we thought were coming? I was positive at release a few packs would be available.

I don't remember modular packs as ever being advertised or promoted.

Certainly there will be libraries released from the usual suspects and probably one or two promo packs from Yamaha direct. These things typically trickle in slowly.

BTW - deserves a different thread but be careful making assumptions on how things work. I'm sure there are folks that can help correct some inaccuracies in what you've described about Montage M elsewhere. It's not bad to complain, but it would be better to do so from a more informed position.

[quotePost id=123797]
My speculation is that it has to do with the tone generation chips.
Probably the SWP70 can't accommodate any of those changes.
[/quotePost]

That's a succinct way of saying it. 🙂

I've been thinking about a way to describe my mental model for both FM-X and AWM2 implementation. My belief is that both FM-X and AWM2 are hardware pipelines. It goes back to the early history of FM. Yamaha implemented OPs as discrete registers, multiplexors, arithmetic units and a shared sine wave lookup memory. No code. I believe they have been expanding and building on this basic design ever since.

Wish I could prove all of this speculation, but Yamaha do not publish information about internal designs. Wish I had a photograph of a decapsulated SWP70. Should show N discrete, time-multiplexed OP pipelines (with a routine network between OPs), if my speculation is valid.

So, if this is true, Yamaha may need to change transistors and wires, possibly needing a silicon spin, too, for any microcode/controller changes.

The pipes are why FM-X and AWM2 polyphony and latency are rock stable. They're not trying to share some single-threaded processor across synthesis types like KRONOS.

Reface DX is a whole different matter. Reface DX is code running on an SSP2 (SH-2 RISC DSP).

Without seeing actual documentation, I could be totally full of B.S., of course. 😉

Hope this WAG is helpful -- pj

I think much of FM-X's functionality is encapsulated in the SWP70, including the 8 operators, the algorithms, waveforms, skirts etc. However, there are things I think are controlled by the main CPU that modulate the SWP70 via register manipulation, for example the motion sequencing. One enhancement they could have made to FM-X, but I don't think they did, would be to allow the more powerful "part LFO" to modulate FM-X parameters, and not just insert effect parameters. The part LFO is such a powerful feature on the AWM2 side of things, but it's not very useful in FM-X. There's just the "2nd LFO" in FM-X, which is more or less what a DX7 had in 1983. No tempo sync, no user waveforms. It's basically like the element LFOs in AWM2.

I know when I had a Motif XS, the "common LFO" on that synth, which became the "part LFO" on the Montage/MODX, was driven by the main CPU. I don't know if it still is on the newer synths, but I would think anything the motion sequencer, mod wheel, aftertouch or assignable knobs/super knob can address in FM-X could also be addressed by the part LFO.

Another thought... there are supposedly two SWP70s in the Montage M. We already know from the MODX+ that a single SWP70 can provide 128 notes of AWM2 + 128 notes of FM-X polyphony. While the two SWP70s in the Montage M give 256 notes total polyphony on AWM2 (divided between user and preset waveform ROMs), apparently only one of the two chips is providing the total FM-X polyphony. Why not use both to bump the total to 256 notes of poly in FM-X? That's assuming one of them isn't also providing AN-X now.

So there are things that, in theory, could be done to enhance FM-X even with much of it embedded in a custom chip.

[quotePost id=123905]I think much of FM-X's functionality is encapsulated in the SWP70, including the 8 operators, the algorithms, waveforms, skirts etc. However, there are things I think are controlled by the main CPU that modulate the SWP70 via register manipulation, for example the motion sequencing.

Why not use both to bump the total to 256 notes of poly in FM-X? That's assuming one of them isn't also providing AN-X now.

So there are things that, in theory, could be done to enhance FM-X even with much of it embedded in a custom chip.
[/quotePost]

Hi Kevin --

Quite correct on your first point. The Yamaha hardware implementation is register intensive. There is an addressable bus that interconnects the host CPU (ARM+Linux) and the SWP70 tandem. The host CPU can directly address all of those control registers in the SWP70.

Another unverified mental model of mine -- the huge list of addressable parameters at the end of the Data List PDF? Some of it is the actual control register layout and addresses. That's a lot of stuff!

I'm willing to bet my hat that AN-X is handled by either SSP3 or SWX09. I don't believe SWP70 is AN-X capable. I'm willing to eat my hat... 😀

Have a good evening -- pj

Meanwhile, I submitted an idea to IdeaScale that asks for FM as an (insert) effect, allowing one Part to FM another, just like the Ring Modulator effect.
https://yamahasynth.ideascale.com/c/idea/158329