If the Insertion Effect is being assigned in an AWM2 Normal Part, you route each Element to either Insert A or Insert B, or Thru (neither) on a per Element basis.
From the HOME screen, select a Part
“Edit” > “Effect” > “Routing”
On this screen (signal travels left to right) each of the possible 8 Elements is shown in a vertical column. Each can be directed to “InsA”, “InsB” or “Thru” before passing through the 3-band EQ on its way there. You must set the routing per Element. In between the two Insert Blocks you choose the A>B, B> A or “parallel”.

If the Insertion Effect is being assigned in an AWM2 Drum Part, you can route each of the 73 Elements through the System Effects (Rev, Var Send) or opt to route them, on an individual Drum Key basis, to one of two possible Insertion Effects blocks A or B, or Thru The Insert Effect block can set A>B, B>A or “parallel”.

There is no “parallel” Insert routing within the FM-X engine. All Carriers go through A>B or B>A.

Extra Credit: Example: of Effect Routing at the Element level of the architecture
Recall Factory Preset “Piano & Strings”
A Single Part Performance. 8 Elements.
Elements 1-4 Acoustic Piano
Elements 5-8 Strings

Press [EDIT]
Select “Part 1”
Touch “Effect” > “Routing”
Here you can see Elements 1-4 routed to “InsA” and Elements 5-8 routed to “InsB”.

This allows the piano to have a separate VCM EQ from the string section.
Touch “InsA” in column 2 to EQ the piano
Touch “InsB” in column 2 to EQ the string section

Alternatively, you could route either instrument to its own Effect… put an Chorus Ensemble type effect on just the strings… etc.

There is no “parallel” Insert routing within the FM-X engine. All Carriers go through A>B or B>A.

Doesn't that raises the question of whether what OP said is really true?

when I set a Part's Insert Effects to Parallel routing

As far as I know a part, in its entirety can only be of one type: AWM2, FM-X or DRUM.

What might I be setting incorrectly?

In order to know that we need to know the actual details of your configuration:

1. What type of part is it? AWM2? FM-X? DRUM? See Bad Mister's reply
2. What effects are you using?
3. What is the element routing being used?
4. What other 'effect' settings have you modified from the default?
5. Are you using any System effects?

In other words - please provide the info needed to reproduce the problem.

If you're hearing all of your elements but only hearing one of the two effects then the problem is likely you're not routing any elements to InsB (all routed to InsA). In which case even though you have parallel set up - InsB isn't utilized because no elements are told to go through InsB. The same is true if you swap "InsA" and "InsB" in the previous statements.

@Jason, Exactamundo!

@ Andrew, no it means that they are not working with an FM-X Part since you would not even see the “parallel” setting as an option.

Thank you for all your responses. I was using an AWM2 Normal Part. The routing I hoped to achieve was a delay effect and reverb, but in parallel not series, for all the elements in the part. This does not seem to be an option: the output in the Part can be routed only to effect A, B or Thru, without a Parallel option. With Delay on Ins A and Reverb on Ins B, I can see now why, if I choose Ins A or B for the Elements, one of the insert effects won't be heard. This seems to be an omission in the design, I'm not sure why. There is a different result to having a delay going to reverb, reverb going to delay or - in this instance - reverb and delay being mixed postfade.

Parallel is available but the connectivity of effects to sources (elements) is what is limited. The two effect outputs are what is parallel. Each element can be assigned to InsA or InsB or neither. Not both (which is what you want). To use an analogy this would require a splitter to route one source to both effects. Internally in the keyboard implementing a "splitter" would chew up resources (signal routing, buffering, DSP resources). These are finite resources that would be exceeded if the "splitter" option was available (output element to InsA AND InsB).

Now - if you are using 4 or less elements then you could double up the elements and route half to InsA (the original set) and the remaining half (duplicates) to InsB. Adjusting the levels down since adding two elements together would be louder.

The downside (assuming you do have the elements to spare) is that this would also double the polyphony required. If near the limit then this may tip you over the edge of THAT resource.

[quotePost id=121822]Internally in the keyboard implementing a "splitter" would chew up resources (signal routing, buffering, DSP resources). These are finite resources that would be exceeded if the "splitter" option was available (output element to InsA AND InsB).

Now - if you are using 4 or less elements then you could double up the elements and route half to InsA (the original set) and the remaining half (duplicates) to InsB.

The downside (assuming you do have the elements to spare) is that this would also double the polyphony required. If near the limit then this may tip you over the edge of THAT resource.[/quotePost]

Thanks for that. I'd not realised this parallel routing option would have a significant effect on DSP resources, but I'm sure it would. I had considered duplicating the elements, so I'll work with that solution. Everything is finite.

I had considered duplicating the elements, so I'll work with that solution. Everything is finite.

Or you could duplicate the part and send one part to each effect.

What is it about the serial method (A->B or B->A) that affects the sound you are going for?

[quotePost id=121826]
Or you could duplicate the part and send one part to each effect.

What is it about the serial method (A->B or B->A) that affects the sound you are going for?
[/quotePost]Yes, either would work. Serial sends the delay through the reverb or vice versa, which is not what I was after. Separate mixable paths for each effect has a different sound result.

This is another option for you.

If your priority is to have a Part routed in Parallel to Delay on one Side and Reverb on the other Side, you can use Var = Delay, and Rev = Reverb.

Bear in mind, that means the Var FX for all the other Parts will also be Delay... but you don't have to use it... just set the other Parts' Var Send = 0.

If your Performance is only 1 Part, then it's a no brainer, Var & Rev will do exactly what you want. (NB... set Var to Rev = 0 for true parallel separation).

My thoughts on running Delay Parallel to Reverb? You're free to do things however you choose, but in my experience, it doesn't sound so good. Reverb into Delay or preferably Delay into Reverb nearly always yields excellent results. Give it a whirl and see what you think.

With Var = Delay and Rev = Reverb, you can set them to Delay ==> Reverb by setting Var to Rev High (up to Max) and Var Return Low (down to Zero).

You can get combined partial Parallel and Series results by adjusting Var to Rev and Var Return.

Also... remember there is a Part "Dry Level" Control in the Mixer, as well Var Send, Var Return, Rev Send and Rev Return. So the Mixer screen is a convenient one-stop-shop for balancing everything the way you like it.

I have tried and found this to work. 😀
retro goal

"Parallel" is another misnomer, and makes it seem more capable than it is.

It should be called "Single" or similarly indicative of the actuality, avoiding this confusion and instantly informing the user of what's going on.

The routing I hoped to achieve was a delay effect and reverb, but in parallel not series, for all the elements in the part. This does not seem to be an option: the output in the Part can be routed only to effect A, B or Thru, without a Parallel option. With Delay on Ins A and Reverb on Ins B, I can see now why, if I choose Ins A or B for the Elements, one of the insert effects won't be heard. This seems to be an omission in the design, I'm not sure why. There is a different result to having a delay going to reverb, reverb going to delay or - in this instance - reverb and delay being mixed postfade.

Understanding the routing will help you accomplish the goal.
Misunderstanding or not understanding the routing leads to disappointment/failure. You seem to get it now, but are curious as to ‘why’ - I assure you it is not an omission.

In an attempt to explain and help you understand ‘why’… this routing is not an omission.

‘Why’ — is so that if you use a Part to build multiple instruments you can assign a real time controllable Effect to them and have at least two separate ‘performing effects’ (as I provided in my initial reply where acoustic piano and strings share the one Part yet each, via routing, can have an independent real-time controllable effect). Insert Effects can be assigned real-time parameter control allowing you to ‘perform’ them.

As you’ll see what you are looking for is “up a level” of programming…
Instead of trying to do it at the Element level of the architecture — the “big deal” about this synth engine is the breaking of the 8 Element limit to each instrument … you can use a second, third, fourth, fifth, sixth, seventh or even an eighth Part to construct that one instrument!

The selection of “A>B”, “B>A” or “Parallel” takes place ‘inside the instrument’ (this is routing within the Insertion Effect Block of a single Part) inside the Part. The Insertion Effects are part of the instrument, much like the Oscillators, Filters, Amp EGs, LFO, response to velocity… these are all parameters that make up the instrument itself — these are the ‘elements’ that compose the instrument being emulated.

Once you have constructed the instrument… you plug it into the digital mixer… there are 16 channels on the mixer. It is important to see this as a separate entity. And like plugging an instrument into a mixer how the effects apply are to the entire instrument.

__A Single Part (green) program is utilizing one set of 8 Elements to complete the construction of the instrument and occupies one channel on the mixer.
__While a Multi Part (blue) program is utilizing more than 8 Elements to construct the instrument. As you know, by linking Parts via KBD CTRL you can add Elements — each Normal AWM2 Part adds potentially 8 Elements.
Please note: each Drum AWM2 Part adds potentially 73 Elements.

Single Part programming
Element routing can be to InsA, to InsB or Thru (neither) of the Insertion Effect block items within its Part. This is not a design flaw — it’s a feature. When you use multiple Parts to construct your sound you’ll see that you have more resources and their roles can change (see ‘Multi Part programming’ below)…

Since each AWM2 Element can be a complete synthesizer instrument (complete with oscillator, filter, AEG, LFO, and routing to InsA, InsB, etc. In the example, I referenced where 4 Elements are used to construct an acoustic piano, and 4 Elements are used to construct the string section; each of these ‘complete’ synth constructions can use an Insertion Effect of its own by placing the routing = “parallel”. Piano in InsA, Strings to InsB.

Yes, the output of InsA will contain just the piano.
The output of InsB will contain just the strings.

After passing through the the Insertion Block, the signals are combined in the 2-band EQ, before you leave the instrument construction area and are plugged into the band’s (Performance’s) digital mixer. From this point, it enters the mixer and it’s exactly like plugging your instrument into a mixing console.

The channel of the mixer has two Auxiliary Sends (one to the System Reverb and the other to the System Variation).
__ If you use this Send (Rev Send) to route signal to the Effects (System Effects) both the acoustic piano and the strings will go through that send to the Reverb, together.
__ Likewise, if you dial up the Var Send, both instruments (all Elements) will go through that Variation Send, together.

Now before we leave the subject of building a Single Part instrument, you can configure your mixer’s Send Effect (Rev and Var) so that you can determine if the Var Send and Rev Send are independent, or you are routing the Variation Output to the Reverb before being Output.

From the HOME screen:
Press [EDIT] > “Effects” > “Routing”

The Var can be sent on to the Main Out
The Rev can be sent on to the Main Out
But you also can route the Var to Rev before going to the Out.

This has been part of Yamaha’s routing scheme for as long as I can remember. Reason: when you are using the Variation Effect to do Time Delays, (for example: multiple repeats, echoes) and you are using the Reverb Effect to build a traditional acoustic environment (Room size)… it sounds very different if each of the delayed repeats have Reverb or not.

Consider a space where when you clap your hands the sound creates multiple complete repeats (echoes) — having a “Var to Rev” send amount parameter allows you to recreate the natural situation where each handclap creates a reverberation of its own.

So it matters that they can flow one to the other. Without this parameter the repeated echoes do not sound like they are taking place in the same space. (True, most people, not even most musicians, listen that closely, but check it out… worth the time to learn about what’s happening and what you can do).

If you need to maintain separate routing between components — step up a level to Multi Part programming! (Make sure to master COPY and EXCHANGE you’ll find them invaluable when getting into Multi Part instrument programming).

Multi Part programming
Enter the MONTAGE/MODX/MODX+ engine where you can use multiple Parts to construct your instruments. Say I use 2 Parts to build the Acoustic Piano/Strings combination. Use a complete Part for piano, and a complete Part for the strings.
Of course, I now have 8 Elements for the Piano and 8 Elements for the Strings. I don’t have to use the A>B, B>A, or “Parallel” options within the Part… each Part has its own parallel universe. It’s own complete set of processing resources.

I could use 2 Parts to build the acoustic piano and another 2 Parts to build the orchestral strings. Or I could use 4 Parts to build the acoustic piano and another 4 Parts to build the orchestral strings.

In these scenarios, your options are greater. And please don’t talk to me about “burning resources” — that’s what they are for… these are Music Synthesizers (not General MIDI sequencing devices - although they can do that, as well, by limiting your use of its massive feature set)!

You can take advantage of the Multi Part (blue) Performance programming to maintain your independent (“parallel”) routing scenarios. In order to get samples to do more and to “play” more realistically — you can add more Elements, as a strategy. Knowing how they group (in this case in sets of 8 per Part) and how the groups link together (KBD CTRL) and at what point you want combine them … will get you where you need to go.

If you notice in AWM2 programming very few of the Multi Performances are created simply to layer sounds to make them massive (that’s easy programming) the programmers were hired to be creative, to use the multi-part architecture to add interest and expand the palette of articulations.

Extra Credit: Respect the Element
A single Element is a complete synth. I recently posted a Motif ES sound (converted for MONTAGE/MODX/MODX+) that was thought to be a Performance - it was actually originally a 4 Element Motif ES Voice. An acoustic piano, an electric bass, an electric guitar and percussion w/Arpeggio: Here
Elements are the first level of power…
Parts are the next level of power
Learning how they can be used to construct your sounds is well worth the effort.
Knowing you can use Elements to play specific roles, deliver specific articulations. You can route Elements to specific Effects.
Knowing when to move up a level of programming is key.
Notice that some Multi Performances used to make a single instrument sound, use less that 8 Elements in each Part… this is done so they can treat them differently (apply different routing, different Effects, independent control) from what you describe this seems to be where you ran aground.
Don’t be afraid to explore.

Enjoy.

Hope that helps.