Hello Joe - fair bit there! I very much agree with all the odd-numbered ones (1, 3, 5, 7 ... ) but haven't a clue what you're talking about with all the even-numbered ones (2, 4, 6, and 8). If I had to pick jist one, it would be #3 - and if I could have two, they'd be #3 and #5 ...

The idea to commit the current parameter offsets to a new performance aligns slightly with a request I've made to be able to see (externalize) the internally-held offset+programmed ("current patch settings") value. I was wishing for a way to make the programmed value boxes show the offset'd value by enabling such a mode. Wouldn't be that far of a stretch to commit the offset'd values to a new performance if you were in this mode and pressed [STORE].

========
For the random output generator (curve - or I would add pulse as well) topic:

I've also requested the ability to edit all 128 positions of the user curve by external software or otherwise. There is some hope for this particular line of a request.

I've also requested some computational curves which allow for functions - I didn't ask for "rand()" but I did ask for round() as one - to force the input values which can be floating point (fractional) to a whole number. I'd welcome any number of math features to the curve - some items that allow to add another dimension in addition to the existing lookup table based curves. When you dig in to the implementation - more would need to be spelled out what parameters you want (what exactly you have in mind for param1 and param2).

I'm sure you've looked up the word "random" in the documentation and were not satisfied with what's available. If I wanted a motion sequence or superknob automation to generate completely random outputs - I could do that. The issue is that I cannot mix random and non-random cycles. The random (for step value) applies to all cycles. So if you want cycles 1-4 to do something completely not random, then cycles 5-16 to be random - this is not possible. The best you can do is to make cycles 1-4 to either hold a "0" or hold a "64" (pick one for all cycles 1-4 since unipolar and bipolar also cannot be set for each cycle) and then do the random part.

So ... it's possible to do basically a "pause" ... then random ... then "pause". The other problem with this is that the way I use random for a number generator is that one cycle produces one random output. So having the "pause" section reduces the amount of random values I have in each "burst". The method I use for a random number generator would be better suited for "everything is random". For this, I set the tempo to 300, the multiplier to 50%, the note type to a 32nd note and have all of my cycles set to output a random value.

Now - I output a random value in a cycle by setting the curve type to "hold" and use the pulse setup menu "random" value of 127. I also set the cycle amplitude (in the sequence edit menu) to 127 so I get full swing from 0 to 127 values. With all of these settings and using superknob automation I can see superknob dancing around "like crazy" all random values. I output this (link a global assignable knob to a PART assignable knob then link the part assignable knob ...) to PART pitch - and I have a very S&H sounding random note.

The "problem" is that it's random 100% of the time which may not be what you want. I may be able to play with this more to figure out a way to negate the randomness and generate the equivalent of a "gate" for the control curve outputs. But at the moment, I just have a 100% random generator which is probably not what you want.

Personally, of the standard curves, I find "Sequence" most suitable as a "random" sounding generator since "FM" doesn't provide full-swing output for me.

====
pie in the sky:

The envelope follower is kind of my "go to" to generate values dynamically. Since I can output silent PARTs directly to the envelope followers just for the sake of generating numerical data as inputs to the control curves - I'll modify "rules" for the output of that PART - usually using FM since it's easiest to tweak any given aspect - and then feed those values into some destination PART to do something. It'd be nice to have a more direct way of doing this and not have to build value streams based off of essentially amplitude over time - but at least we have that.

We could almost get there if we could just take two lanes, run them at the same time, and choose to add or multiply the two (negative values for subtract and divide). This would allow for easy gating (multiply by 1 = no gate, by 0 = gate) and opens up some other possibilities.

... basically, I'm with you when it comes to these requests because they're similar to some of the ideas I've floated - but I also have to appreciate any hardware limitations that exist (resources) - the development constraints (manpower/cost) - and the support/documentation/etc that Yamaha would have to take on if there was even more complicated stuff inside Montage.

=======
Basic non-sequencer, non-sampler features for waveform parameters:

I think in your list "normalize" is one of the things I would take off. Since that's more of a sampler feature requiring you to modify the WAV/AIFF data which requires a buffer for the output ... which I don't think is going to get very far. Just values like root note, loop points, direction, etc do not require any sampler features.

I agree (except for normalize) - and have asked for this feature as well. It'd be nice to be able to edit the loop points without a graphic editor, forward/reverse/etc properties. Just basic enter the values type stuff. No crossfading or other things I would use the computer to pre-edit the waveform before loading up. I'd have a notepad with my loop point values and would just enter those things in.

Ability to "scarf" those values from the filename? That'd be nice, but I'm not against entering these things by hand one-by-one as opposed to the MC-hammer 'cant touch this' parameter approach we have today.

... it's in multiple waveforms I think is where the gotcha comes up. Unless just supporting scarfing the settings from the filename: The "engine" to allow for entering this data isn't trivial. It's doable - but I'd also be equally (if not more) fine with a free utility that handles this offline.

... other items you wished for all make sense. Filesystem will always be a thorny part. I'd like to see X7P files which are basically X7B + waveform so you can save (on Montage native - not requiring Montage Connect) an X7P to a USB stick which is your current edit buffer performance plus its associated waveform file. You can also load an X7P which would allow for loading a single performance and it would overwrite only your edit buffer - not the user area. Then you could [STORE] the performance to add to your user area.

I'd mostly use this if I were offline generating my own performance with waveforms -- and there was a way to do this with a free app that created the waveform+performance (X7P file). Then I could generate my own content and load it as a single performance without going through the process shuffling around the data through the library area to get it added to my previous user content. This also deals with some requests out there. Although not utopia - it's a start.

Great ideas above. My top requests:

1. EVERYTHING assignable. Way too many and arbitrary parameters can't be assigned to controls and therefore SuperKnob. The whole selling point of SuperKnob is music in motion and yet a majority of parameters can't be assigned?

2. All 16 parts under keyboard control.

For the "everything" request - I think there may be some limitations here. I too would like to see "everything" as the goal - but also believe there are some limitations that cannot be moved which limits the possible count to a less-than-everything amount. On another thread I suggested making a list of what's missing then possibly going through that list and sorting it by rank. Enough of these sorted lists from different sources would give some idea of what's most important to include which, to me, is more helpful than "everything".

Right now what I'm seeing as one possible limitation is, at very best, 136 extra destination parameters. That's assuming no more than 5 internal/reserved values and also that the internal register is capable of holding two bytes (like system-level destinations). If limited to a byte, then there are 12 open slots (at best). Even though the part-level destinations all fit within a byte (8 bits of usage, 0-255), the MIDI data table shows access to 14 bits of usage (0-16383) although I am conservatively matching the high-end possibility to match the system-level amount of slots (0-251).

Beyond that, just the text for the tables so Montage can display, in the GUI, the selected destination takes up space. Because each new destination for this table is 17 bytes of data. 17x136 = 2.3k worth of data. I'm not sure how much size margin there is in the firmware - but if it's near the limit, then 2.3k may be significant. That's just for a datastore for the display. The code that may be needed to manage this takes up resources. There are other possible technical constraints. There are also non-technical constraints.

Just, as also mentioned in another thread, saying it may not be possible due to resource limitations to add "everything" and encouraging more of a prioritized list type request.

Hi all,
I have first two requests:
1. Using a normal PC-Keyboard to bring up names in the name-field or can alternativly navigate through the screen.
For blind users it will be a good idea. So we can do more as now.
2. The chance to place and organize screens myself to make it easier to know there the buttons are on the touch for blind users.
I have one more thing in the connection with Montage Connect.
I think it is posible that Montage Connect can give blind users informations what I do on the board. Native Instruments shows how easy it is to make things accessible for blind users.
If Yamaha jumps not on the train I only can hope that John Melas new tools for the Montage will have the same accessibility like the Motif series does.
All the best
Rainer

Yes indeed more assignable parameters!

That is a big request of mine too and I forgot to include in my list above.

Regarding making a priority list of parameters to add to designation list: everyone will have different ideas based on their own creative needs/wants/workflow/patch making ideas... but sure, when I have a moment to do so, I'll contribute with my own priorities.

Thanks Jason for spelling out some of your ideas re randomness etc.

Cheers

I didn't suggest everything assignable at once. I suggested that every parameter should be assignable. For example(s):

AEG Decay1, not assignable - can't assign the decay amount to a dial? Every analog synth can do this.
AEG Time/Velocity, not assignable - can't change the velocity sensitivity to a dial? My Roland System-8 has velocity sensitivity as a first class citizen (knob).
Random Pan, not assignable - can't change the randomness of a pan with a dial or SuperKnob? Would make for a very interesting use of SuperKnob/Pedal.
Element LFO Fade In, not assignable - can't change the fade-in amount of the LFO with a dial? LFO fade-in is a knob on every other synth I own. Give me a break.
Ins A, SPX Hall: Reverb Time and Feedback ARE assignable, Reverb Delay and Density are NOT assignable. Who makes this up?

It's artibtrary and poorly chosen frankly. How about all of these params in the code have references, let's just go ahead and make them all assignable. I keep having creative ideas for SuperKnob and keep having them shot down when I find out the related params are not assignable. Not the point of SuperKnob guys.

Hi Aaron I think you maybe misunderstood Jason's reply?.. He and I and I'm sure many others too agree with you.

And yes, I too have ideas shot down because the parameter in question is not assignable. Which is disappointing.

Jason is trying to explain that it may not be possible to be able to have all of the parameters as assignable options.

Joe

Right - merely reiterating that a prioritized list from several users (several because each would have a different "spin" on what's important to them) would give better quality data than "everything". Since there may be hundreds of missing parameters (I haven't tabulated the "gap" yet) and only a few handfuls of slots which are available. Thus - any request for "everything" would be ignored as either too vague or technically unfeasible - but a collection of prioritized lists ("votes" if you will) may be helpful. There is also the possibility that there are no open slots. Even so, the lists can serve to influence future parameter control decision making (follow-on products). Presumably you will later be in the market for the next generations of flagship Yamaha synths.

Previous references to similar (destination related):

Original suggestion to make a prioritized list:
https://yamahasynth.com/forum/mod-control-parameters-other-ideas#reply-14961
More destination:
https://www.yamahasynth.com/forum/montage-may-be-most-powerful-synth-yet#reply-14909

different topic - but related to destinations and sources. idea to conserve destinations:
https://www.yamahasynth.com/forum/about-motion-sequence-automation,-only-the-super-knob-can-be-automated-directly#reply-14675
Some early commentary on possible issue of adding more destinations (not what this thread covers, but eventually you'll need more than 16 destinations if you really "like" the longer list hypothetically delivered). General idea that resources are limited transfers:
https://www.yamahasynth.com/index.php?option=com_easydiscuss&view=post&id=7997&Itemid=851&viewallreplies=1#reply-13097

What I have so far is only FM-X related: 80 parameters that are in the MIDI Parameter Change tables, but not individually assignable as a destination.

https://ethercalc.org/5sredj6zmxue

There is no implication that any of these parameters should be a destination - only information that they are not currently assignable as (individual control) destinations.

As an example - coarse tuning is missing but (to me) not an issue given you can affect the frequency in fine-tune increments along with the scale factor - probably OK (even though something is missing).

There may be other clear examples where a missing destination "does not matter". The list provided is not trying to make that determination. More information is necessary by means of ranking each parameter.

I wouldn't suggest doing this on the link - the link was provided to show where I am thus far with generating the list.