Developer Fund

User Feature Request List

Moderator: Solaris Moderators

david
Posts: 19
Joined: Mon Oct 17, 2016 3:55 pm

Re: Developer Fund

Post by david » Wed Nov 29, 2017 6:37 pm

I am not sure I’d want it to become a five voice. Could live with a six voice :lol:

The challenge of programming to current hardware spec begs the question of an internal processor and storage upgrade. As we all know, there is tremendous power and storage available these days. That is prob a huge project and maybe not feasible based on the current layout and you might not want to open that door. If you could upgrade the chips and keep the current chassis that would be epic, but I’m sure you’ve contemplated this and it’s prob easier said that done.

Octupole
Posts: 16
Joined: Mon Oct 31, 2011 10:47 am

Re: Developer Fund

Post by Octupole » Thu Nov 30, 2017 2:15 am

The reduction of polyphony from 10 to 5 would be problematic for me; I would not be happy with that. I have no idea of DSP programming, but could it eventuallly be made possible to load the oscillator / filter code to the DSP in a dynamic fashion, so that only code actually needed in a sound would be loaded? The Claiva Nord Modular G2 must have a mechanism like that. The complexer the patch you set up, the higher the DSP load and the lower the polyphony.

normanion
Posts: 7
Joined: Sun Jan 29, 2017 9:31 am

Re: Developer Fund

Post by normanion » Thu Nov 30, 2017 4:05 am

John,
John Bowen wrote:
Wed Nov 29, 2017 5:44 pm
1) Adding new oscillator or filter types can very possibly mean a reduction in the polyphony from 10 to 5 voices. Would that be OK?
Is it possible, that programmer misunderstood the intentions and was thinking about adding more oscillators or filters to be available at the same time with those already in place? I'm not a DSP programmer, but my limited programming skills tell me that new types should only take computing resources when they are used, so leaving new types - in my mind - should give same polyphony when using four oscillators and four filters like before. OR... new models are more computing demanding than those already in place?
John Bowen wrote:
Wed Nov 29, 2017 5:44 pm
3) Current analysis says that the arpeggiator and sequence modules would have to be single per MultiSet. I’ve asked about having at least 2 separate arpeggios going, so we will see.
I think I see it. But can it be turned off and on for different layers? For example: pressing chord and three voices are playing simultaneously for one layer and arpeggio is going for another? Since sequence is saved per patch, how would it be handled?
John Bowen wrote:
Wed Nov 29, 2017 5:44 pm
4) Difficult to say at this time if it’s possible to have 4 separate Master FX running, or if we have to do the same thing as with the arpeggiator and sequencer (where there’s only one Master FX setup for the Multi). I know almost everybody else has tons of FX sends and so on - the Solaris is a little different in that not everything is running on the same chip. We could definitely expand the number of effects, though.
Again, if I understood right, when patch is recalled, from which those parts are taken from?
John Bowen wrote:
Wed Nov 29, 2017 5:44 pm
I greatly appreciate all those who have posted their responses here - please continue to do so , and please spread the word to other Solaris owners to come check out this and the other topic about new hardware.
Like with OS update, maybe it's good idea to send word to synth news outlets?

Best Regards,
Norman

John Bowen
Site Admin
Posts: 1655
Joined: Sat Apr 07, 2007 3:00 am

Re: Developer Fund

Post by John Bowen » Thu Nov 30, 2017 10:48 am

Octupole wrote:
Thu Nov 30, 2017 2:15 am
The reduction of polyphony from 10 to 5 would be problematic for me; I would not be happy with that. I have no idea of DSP programming, but could it eventuallly be made possible to load the oscillator / filter code to the DSP in a dynamic fashion, so that only code actually needed in a sound would be loaded? The Claiva Nord Modular G2 must have a mechanism like that. The complexer the patch you set up, the higher the DSP load and the lower the polyphony.
Yes, originally I thought we would have the ability to implement dynamic allocation of the DSP code, but it turned out not to be the case.

chapolin
Posts: 49
Joined: Thu Feb 26, 2015 4:35 pm
Location: France
Contact:

Re: Developer Fund

Post by chapolin » Thu Nov 30, 2017 12:41 pm

Yes, originally I thought we would have the ability to implement dynamic allocation of the DSP code, but it turned out not to be the case.
So, I prefer to have 10 voices than 5 ...

CA3080
Posts: 35
Joined: Sun Nov 25, 2007 11:06 am

Re: Developer Fund

Post by CA3080 » Thu Nov 30, 2017 1:15 pm

John Bowen wrote:
Thu Nov 30, 2017 10:48 am
Octupole wrote:
Thu Nov 30, 2017 2:15 am
The reduction of polyphony from 10 to 5 would be problematic for me; I would not be happy with that. I have no idea of DSP programming, but could it eventuallly be made possible to load the oscillator / filter code to the DSP in a dynamic fashion, so that only code actually needed in a sound would be loaded? The Claiva Nord Modular G2 must have a mechanism like that. The complexer the patch you set up, the higher the DSP load and the lower the polyphony.
Yes, originally I thought we would have the ability to implement dynamic allocation of the DSP code, but it turned out not to be the case.
John, forgive me for arguing about something you rather clearly state is unrealistic. However, this is an important point for me.

I want multitimbral mode so I can play layered and/or split sounds. 10 voices are a bit to few for that to be generally useful. (A reduction to 5 voices is even less useful from that point of view.) I fully agree with Octupole that dynamic DSP configuration is desireable, and understand that it may take a lot of work to change the DSP code to support that. That's why I'm willing to pay a bit more than the $200 originally suggested. On the other hand, if the number of voices can't be increased, multitimbral mode is relatively useless for me, thus making me less interested in funding an OS development that will bring only minor improvements.

John Bowen
Site Admin
Posts: 1655
Joined: Sat Apr 07, 2007 3:00 am

Re: Developer Fund

Post by John Bowen » Thu Nov 30, 2017 1:24 pm

Implementing a Multitimbral mode will not affect the polyphony. Adding a lot of new oscillator or filter types MAY cause loss of polyphony.
We still need to do quite a bit of checking on this.

Additionally, there is a new main board in development which would ostensibly allow addition of more models without loss of polyphony, but this development is going to take a while. Possibly we can also use some of the Developer Funds to help along with the new main board.

Benoist Guitton
Posts: 11
Joined: Mon Nov 27, 2017 3:16 pm

Re: Developer Fund

Post by Benoist Guitton » Fri Dec 01, 2017 1:18 pm

Hello,
I’am definitely in for spending in the 200 to 500 $ range

Since multimbrality shouldn’t impair further developments ( I really didn’t expect this at all ) I’am not against it …

Really lukewarm for more global / final stage FX like reverb, I don’t mind a basic reverb or others FX onboard but nothing more or deeper that would spend too much / would be too cumbersome to edit, there’s plenty of excellent existing FX you can feed Solaris into, no need to reinvent the wheel …

In the meantime to add my 200 $ and 2 cents :

If unfortunately more oscillators and filters were to impair voices number below 10 ( a no-go ) : few but powerful additions and changes, basket to choose from below !

I wish more deeper morphing twistable oscillators, like Cylonix Cyclebox 2 + it’s extension or Shapeshifer ( https://www.youtube.com/watch?v=H5B9BHPHg_o and https://www.youtube.com/watch?v=XkSnnc9tNJg ) and Adern’s Flexor 3 oscillators : basic waveforms with parameters to change geometry and beyond like a trapezoid where you can change slopes on both sides and the width of the top ( from the Flexor 3 manual : FR Trapezoid is a free running oscillator that generates a waveform whose variable trapezoid shape ranges from saw through square to triangle. The variable shape can be controlled and modulated. This oscillator module is Band-Limited to reduce aliasing and is capable of bi-polar FM (frequency modulation)).

So just a shape parameter would be not enough, add 2 more slots, since there ‘s not enough room in menus : merge Glide On / Off with Glide time so it’s either Off or with a value : there’s two slots free.

Maybe create a good set ( 20+ ) of waveforms « clay » ( again Flexor 3’s FR oscillators are good inspiration ) that will offer the maximum grip for modifications, make then different groups where the differences will be about the combos of algos used : example on a given group you’ll have Shape / Fold / WaveMult and other algos on an another group with still the basic waveforms.

Karplus-Strong oscillator

Implementation of phase modulation in addition to LinFM and Pitch ( been discussed in a topic, and proposed above )

For the existing and future filters, it would be nice a have a gain / drive control that alters significantly the response of the filters like on modulars, to make them growl and squeal … etc, I haven’t checked the process : it could be a different resonance profile and instabilies aroud and maybe multipeaks resonance appearing ( ??? ).
The boost parameter on the VCA already does that a bit but I think it could be pushed in intensity and behaviour in the various filters via this drive / gain addition.

More signature filters, also for MM1 filters like BR2+LP1 possibility of separate modulation of levels of BR2 and LP1

Nonlinear filters like this one implemented in Jeskola’s Buzz long ago http://jeskola.net/archive/buzz/a15/Gea ... ilter.html which is quite astounding and unique in possibilities

For the insert FX, a waveshaper would be nice like the WMD Geiger Counter using the all the waveshapes in stock especially the WT in modulation

A phase Lock loop like the dopefer one, that type of stuff

I proposed John to think about many more LFO waveforms that will be mostly basic geometrical ones and combination of them, asymetric … etc
Could be nice if LFO could synced to each others

Mixers could be more ! on page 1 there’s a slot free for various modes of combination / ( bits manipulation ) collision, twisting ( inspiration : cyclebox 2 modes and https://vimeo.com/43356010 ) you get the idea …

For the rotors, if we could modulate amount of inputs, I was thinking as well about a radius between 90 and 270 ° rotating so you’d have more than input 1 if you start at input 1

Navigation ideas inside « blocks »: use of joystick or modwheel + data wheel, for instance you select oscillator 1 then use of modwheel for up – down navigation, data wheel for horizontal ( oscillators 2,3,4 ) of coarse LEDs would indicate position, joystick would be even better as you could explore the whole « block » great on the graphic display !

Use of the LEDs to know what’s playing a part in a preset, if LFO4 does modulate something that doesn’t make it soundwise then the LED is off, same when browsing the Graphic Display pages.

Implementation idea : additions in the menus of the Assigns buttons in toggle mode as show on / show off i.e you have show on Assign 1, press … all LEDs above « sounding items « turn on press a gain and you return to classic edit mode display ; meanwhile if you wish you could edit in show on mode by pressing the item to be edited or just checked.
For instance you press show on : among other things all Mixers LEDS are on, if you have confidence in your short time memory you could check one by one what’s in them and the possible modulations.

Otherwise I really miss a Log slope for EGs especially to add shoulders to the A / D stages so turning clockwise you would have exp --> negative values --> Linear --> positive values --> Log

Delay modulation added in page one of EGs modulation

Top of my head for now …

As for participation in this fund, if every Solaris Owners is contacted via mail + a good trumpeting on facebook / Twitter for owners of resold units, I think we could gather a significant amount of owners ( also coupled with the new hardware ideas ) maybe wait a little more for more suggestions then launch the offensive.

Benoist Guitton
Posts: 11
Joined: Mon Nov 27, 2017 3:16 pm

Re: Developer Fund

Post by Benoist Guitton » Sat Dec 02, 2017 5:46 am

A very important and time saving implementation I didn’t include not to be drown in my previous floody post …

An entry that could be added in Assign menus that would do « clean the mess then defragmentation / more «

During preset creation / modification there are plenty of what I call deadends : the most simple of all is an oscillator turned on that doesn’t go into a mixer and in the end doesn’t participate in a preset’s sound, it’s just turned on and it’s visual pollution because you figure because it’s « on » it has to do something somewhere and you *waste* a lot of time browsing !!

So the first « pass « of the algo would set to off for instance an oscillator and it’s modulations and what it’s modulating that doesn’t help produce sound : clean slate ! because if there’s an uncleaned modulation like a LFO rate mod that plays it’s part when you decide you use that LFO, you want to first start with a steady rate.

The second pass is defragmentation and « push up » : because it will create holes everywhere like only OSC 1 and 4 or 3 and 4 that really play a part in the sound, so all the settings are moved to 1 and 2 and very important : same for modulations, if Mod Source 1 and 2 are empty I doubt you’ll going down to 4 but if this one is active ? so by regrouping you save more precious time !!

At the end of the process, all the LEDs of the active components will stay turn on so you know what parts to modify even when browsing the submenus of the graphic display, also the Enable Part LEDs should indicate the whole sound that a preset is capable of because I noticed many presets where only part 1 is enabled but the other parts also produce sounds so you are potentially missing something …

It would be nice if in the process it would also create function group shortcuts in the Graphic Display !

I’am aware that it may require some processing time that’s quite annoying when you’re in creation mode, but the amount of time you’ll save later is priceless.

Also it would be great to have a user preference for order of appearence in the menus of the Graphic Display and … modulation menus especially the monumental List 1 where I’am not found of some placements ( not to be saved with presets )

I know microtuning is low on the priority list but if it could be coded « easily » … :wink:

jgale
Posts: 49
Joined: Sat Sep 27, 2014 1:38 pm
Location: New York

Re: Developer Fund

Post by jgale » Sat Dec 02, 2017 8:00 am

An entry that could be added in Assign menus that would do « clean the mess then defragmentation / more «

During preset creation / modification there are plenty of what I call deadends : the most simple of all is an oscillator turned on that doesn’t go into a mixer and in the end doesn’t participate in a preset’s sound, it’s just turned on and it’s visual pollution because you figure because it’s « on » it has to do something somewhere and you *waste* a lot of time browsing !!
+1 - what a neat idea! I usually leave these as is and struggle later when editing the sound - or have to clean these up manually if i am making template...

John

John Bowen
Site Admin
Posts: 1655
Joined: Sat Apr 07, 2007 3:00 am

Re: Developer Fund

Post by John Bowen » Sat Dec 02, 2017 9:49 am

I just start from the end point and work backwards if I’m trying to ‘dissect’ someone else’s patch, but in general I don’t need all of the ‘visual defragmentation’ with patches I’ve created, since I already know the structure. I know Benoist and I talked about this some time ago, because in my Scope plugin for Solaris I do have little ‘LEDs’ showing if an oscillator or filter is being used by that particular sound. We discussed using the hardware section select button’s LEDs to indicate these as well, with something like ‘halfway dimming’ them if they are on, and other such devices, but I was never really happy with any of the approaches.

I guess we could do something using the Shift button (since that what it’s generally useful for - adding extra functions that were normally covered). Holding down Shift and the Enable Part button, and have the section LEDs stay on for the bits that are being used in the current preset.
I have no idea of how complex the code would have to be to do this...probably lots of “if...then” statements :-)

david
Posts: 19
Joined: Mon Oct 17, 2016 3:55 pm

Re: Developer Fund

Post by david » Sat Dec 02, 2017 10:05 am

You could also accomplish visual feedback and tracing by building an external editor app.

The bulk of the programming workflow could still be on the Solaris but you could scrub, tweak and check your work on the external app.

This just another angle to look at the issue, I am not sure it’s the optimal solution for everyone.

Studio Electronics recently put out VST apps for their old rack synths and it helps with the setup and routing.

John Bowen
Site Admin
Posts: 1655
Joined: Sat Apr 07, 2007 3:00 am

Re: Developer Fund

Post by John Bowen » Sat Dec 02, 2017 10:15 am

david wrote:
Sat Dec 02, 2017 10:05 am
You could also accomplish visual feedback and tracing by building an external editor app.

The bulk of the programming workflow could still be on the Solaris but you could scrub, tweak and check your work on the external app.

This just another angle to look at the issue, I am not sure it’s the optimal solution for everyone.

Studio Electronics recently put out VST apps for their old rack synths and it helps with the setup and routing.
Yes, this is what I was thinking as well. Would be better to spend time on an external editor, and not burden the keyboard code with this task.

minorguy
Posts: 62
Joined: Fri Dec 05, 2014 12:09 pm
Location: San Jose, CA, USA

Re: Developer Fund

Post by minorguy » Sat Dec 02, 2017 10:37 am

My Solarium editor already does this. It starts at the output and traces backwards to find all modules that are participating in making the sound and it only displays those. And you can click on Enable Part buttons to see which modules participate in each part. It's particularly useful for quickly seeing what controllers are modulating things and where.

What it doesn't do is actually modify the preset to turn off unused modules. So if you are working at the front panel and you want to see, for example, OSC4 changed to be OFF if it's not connected, it doesn't do that right now. But I suppose that could be added.

The code is currently open source. I'll be improving it in the future when I have time. But I can't build it for Mac since I don't own a Mac.

I agree that this sort of thing might be better left to an external program.
Solaris #249

david
Posts: 19
Joined: Mon Oct 17, 2016 3:55 pm

Re: Developer Fund

Post by david » Sat Dec 02, 2017 11:22 am

John,

I would vote that part of the developer fund be used to hire someone to work on the external utilities (if the original developers will open source their code).

It might not be a huge project to combine, polish and finish them, and there is sure to be someone skilled in external librarian/utilities for synths.

Therefore, you might be looking for two programmers- one for DSP and one for the external utilities.

Post Reply