I spent some time looking through the Montage Owner's Manual. I didn't see any significant differences between the Montage's AWM2 voice architecture vs the Motif XF voice architecture. Same eight input waveforms, same way of applying effects, etc.

Hi Michael,
There is very little in the Owner's Manual about the synth architecture, and absolutely nothing about Montage Voice mode (sic) because technically speaking there is not one. The yet to be released Reference Manual will go into details, parameter by parameter, it is then that you will become aware of the difference between what you know as a Motif XF VOICE, and an AWM2 PART in Montage. Obviously there is a similarity between the parameters of an XF VOICE and the structure of a Montage PART that contains an AWM2 based program, the Montage can load and play any Motif XF Voice, Waveform, or Arpeggio data.

What is major league different is what it has above and beyond the XF... Those things concerning Controllers (routing), Effects, Outputs, etc. the entire Motion Control engine... Which starts with the Super Knob, and includes the Motion Sequencer, and the Envelope Follower. (Not to mention a Part could be occupied by an 8 Operator FM-X program).

Assignable Knobs
In the Motif XF you have two Assignable Knobs (AS1/AS2). In a Performance each of the 4 possible Parts can have a unique set of parameters assigned to these knobs... When you call up an XF Performance, if [COMMON] is selected you do not see the AS1 or AS2 knobs available at the COMMON level. If you select one of the 4 Parts, the two AS Knobs appear. Each Part has its own two assignable knobs.

In the Montage you have eight Assignable Knobs (AS1-AS8). In a Keyboard Control Performance (where you can play as many as 8 Parts) each Part can have a unique set parameters assigned to these knobs... When you call up a Montage Performance, there are eight knobs available for the COMMON level. And each of the 8 Parts has its own set of eight knobs!

The COMMON AS1-AS8 are automatically assigned to the Super Knob. Here you can control parameters that affect the entire Performance (like the System wide or Master Effects, etc, any parameters that might be available 'in common' to the Parts.
Each PART's AS1-AS8 can be linked (or not) to the Super Knob, but are controlling something unique to the individual Part. This makes things very flexible... You can group certain things to be controlled with a single gesture, or you can setup to manually control individual functions as you may require.

The Controller Matrix is obviously major league DIFFERENT!

Motion Sequences
Motion Sequences allow the performer new and unique ways to have interactivity and rhythmic motion within their sounds. To describe them simply as automated controllers doesn't really do them justice. But the sky is the limit here. And this is,again, major league DIFFERENT.

Interactivity is a key difference, one Part can modify/influence what is happening with another; even external inputs can be used to modify Parts of a Performance; I feel like I am "dancing about architecture' here (_ Frank Zappa) just how it will impact how each person uses this will be up to them. I see comments about how it would only be of interest to EDM (electronic dance music), well, it's fairly obvious its usefulness there, but I'm finding that's a narrow surface view, because if you use Pads or any sustaining sounds in your music, or if your music has timbre changes or would benefit from instruments that can change articulations, oh my, Motion Control will impact your music. Morphing is motion, timbre change is motion, positioning an instrument so it's sounds like you are up close or moving the listener way off to some distant location, etc.

Envelope Follower
The Envelope Follower allows one Part (the source) to be applied as the shaper to one or more other Parts as followers. Overall, it's really similar to having a room full of oscillators, filters, arpeggiators, amplifiers, LFOs, programmable LFOs, programmable envelope generators, etc.etc., that you 'patch' interactively (except you'd never be able to afford a modular synth with this much stuff).

You will be able to use 100% of your Motif XF knowledge in Montage but that only is the ante in... It is exponentially beyond what is possible in the XF. All the parameters you know are represented, Voices, Waveforms, Arpeggios from the XF are compatible, but obviously there are a slew of new and exciting features and functions that go far, far beyond the 'Voice' architecture of the Motif-series.

PERFORMANCE is the only mode. However there are two types of Performance Parts. Those under real time control of the keyboard and controllers (under "Keyboard Control") and those that available for external/sequence MIDI control.

I'd be interested in hearing someone who actually knows say what might be new in AWM2, something that seems to be implied by the number 2.

AWM2 has been used in Yamaha professional synthesizers throughout the Motif-series. The original Motif (classic) was AWM2. I'm hard pressed to remember the last AWM keyboard... The SY77 (circa 1989) was AWM2

In fact the transition from AWM to AWM2 took place either in '87 or '88... The TX1P (single space acoustic piano module) is the last pro product that I can remember that utilized the original AWM engine. But I'll have to dig into the historical documents for the exact history.

So the "2" is not new... What's new is the Montage's application of Advanced Wave Memory 2 and FM-X with the Motion Control Synthesis Engine. Ive read comments about this only being an SY77 on steroids are to be commended because they recognize some of the terms (AWM2/AFM), but is laughable because just knowing it contains sugar and milk doesn't tell you much about what's really cooking! 🙂

The AWM2 process is Yamaha's proprietary way of storing and retrieving sampled based audio. There are literally hundreds of thousands of keyboards in the field that feature AWM2 over its life span. The difference between AWM and AWM2 goes back almost 30 years when Yamaha started basing synthesizers on digitally recorded data (samples).

According to Phil's other answer, it's still 8 elements and everything is the same as before.

-One thing I forgot to ask is, do we have any new eXpanded Articulation modes?
-I also asked a question about the possibility of Performance VOICE element interactions? (sorry I don't remember the proper terms Phil has used so by VOICE I mean one of the 8 performances). So if I have 8 elements in one slot and 8 in the second, can these elements use for example Random XA mode and treated like a 16 element VOICE?

I hope there will be improvements in the sampler synthesis engine.

By not knowing what the structures are called we are not sure answering the question will help you. Again, there are no Voices as in the XF. The best way to make this translate is to think about a "Part"... In the Motif XF a Voice was placed in a Part when you wanted to address it along with other Parts. 4 Parts or 16 Parts could be assembled as follows:

Four Voices could be played together from the keyboard in a Performance.
Sixteen Voices could be placed in a sixteen Part structure to be played by the sequencer, in a Song/Pattern Mixing.

A PART was a Voice in a situation where it was addressed along with other Voices.

In Montage there are only Performances.
One to eight Parts can be played simultaneously in what is considered a "Keyboard Control" Performance
As many as sixteen Parts can be played by the sequencer or an external DAW.

Each AWM2 PART in Montage is equivalent to what Motif XF folks would know as a Voice placed in a Performance (i.e., a Part).
Each Part is a separate entity, with its own independent assignable addressable parameters.

A "Keyboard Control" 8 Part program in the Montage could have 64 Elements if you select all Parts as Normal AWM2
An AWM2 Drum Part is made up of 73 Elements. So the playable sound in the Montage has potential for many more Elements.

Yes, you could create a Keyboard Control 8 Part program that had all AWM2 Drum Kits (not sure why) but that's possible... I haven't tried it (no reason for it I can think of...) 584 Elements, yikes!

The question is trying to figure if you could have two Parts each with 8 AWM2 Elements set to "wave random" XA Control... I don't see why not.
I have not tried everything. But that seems possible. Each Part would need to have each of its 8 Elements set to "wave random", each would do its random thing, separately.

If you set the XA CONTROL to "wave cycle" where each key press triggers the next Element, the cycle will be contained within the one Part - not across sixteen Elements.

Not sure what you are going after, but between Arpeggiators and the user programmable Motion Sequences new levels of random can be achieved. Motion Sequences can be used to animate sounds in a variety of ways. And since each Part can have its own independent motion sequence the sky's the limit.

I hope there will be improvements in the sampler synthesis engine

The Montage does not have a sampler. It can load and map samples but does not have a sampler. If you take a look at the following video, you will see exactly how custom user sampled data can be prepared for use in the Montage:

Custom Samples to a Motif XF Voice