MSoundfactory design thread

[focusrite]

Right, I understand it better now thanks

[dmbaer]

The bottom line? A saved device can't be edited but it can be added to?

By "added to", if you mean creating new device presets, then absolutely it can. A device preset is just a definition of values of device controls (which are the external manifestation of visible multiparameters). You can take any device preset and tweak the device controls, and if you like you can save your tweaked result as a new device preset.

[ZentralmassivSound]

Here is something I made a few days ago. It consists of 4 FM operators that blend with an Oscillator. Each operator has its own LFO that also controls the way that specific operator blends with the Oscillator. I left all modulators under the hood unused for the user to personalize.
There's a 16 step Sequencer that has a filter blend as well as Convolution + MBDistortion effect for big dirty rock sounds, which can get nastier with the customizable Ring Modulator. I went off of the main template in the presets (from which I made my own similar template) so the effects are for the most part to the template standard. I also experimented with a different approach with displaying the EQ.
The presets should demonstrate fun ways this can be used.

Just tested this. Sounds really nice! It quickly kills my CPU though when I play some chords and turn knobs. Seems that some knob positions cause heavy CPU load and others are light on CPU.

[tylerdedera]

Here is something I made a few days ago. It consists of 4 FM operators that blend with an Oscillator. Each operator has its own LFO that also controls the way that specific operator blends with the Oscillator. I left all modulators under the hood unused for the user to personalize.
There's a 16 step Sequencer that has a filter blend as well as Convolution + MBDistortion effect for big dirty rock sounds, which can get nastier with the customizable Ring Modulator. I went off of the main template in the presets (from which I made my own similar template) so the effects are for the most part to the template standard. I also experimented with a different approach with displaying the EQ.
The presets should demonstrate fun ways this can be used.

Just tested this. Sounds really nice! It quickly kills my CPU though when I play some chords and turn knobs. Seems that some knob positions cause heavy CPU load and others are light on CPU.

That’s too bad I wonder if it’s something that can be fixed in the globals? I do have one knob that controls the transform mode of the osc and there a lot of effects. I haven’t had cpu issues with it through Logic Pro though. I have experienced cpu spike with long release times & chords with more than four notes on similar devices I’m working on. Maybe there’s a more cpu efficient way I could be setting up my devices?

[ZentralmassivSound]

That’s too bad I wonder if it’s something that can be fixed in the globals? I do have one knob that controls the transform mode of the osc and there a lot of effects. I haven’t had cpu issues with it through Logic Pro though. I have experienced cpu spike with long release times & chords with more than four notes on similar devices I’m working on. Maybe there’s a more cpu efficient way I could be setting up my devices?

at the moment I can't contribute much because I haven't really dug into MSoundFactory a lot. All I can say is that MSF has quite a CPU load for many presets on my machine (Phenom II X 6).

[David]

Is there any way of adding a new MP or several, in empty slot(s), to an existing device with a bunch of D Presets, in such a way, or with some settings, so that each load of an existing preset does NOT load in with the NEW control ON, thus changing the preset? And, AFAIK, requiring that every preset be replaced one at a time with it turned off to get back to the original sound?


…
The new MPs I added, btw, were some targeted randomizers for a few bar-formatted parameters, my intention being to allow me to randomize very specific aspects of any existing preset, and be able to save the best randomized results as new variations.

What I got instead was actually kind of cool, but not AT ALL what I wanted.

Once added , the Randomizing buttons work perfectly to randomize just the targets, BUT they also get activated with every load of a different preset, so none of the existing ones ever load now the same way twice…!

Kinda neat to know that this is an option, but, WHY is this happening, and HOW best to fix/avoid it? I then added an enable button for the Randomizer group, which works great, EXCEPT, that I still have to go in and replace every existing preset with that switch set to Off… AND when doing that, the old presets of course come in randomized, NOT sounding like they used to! I did change the default for the Enable button to OFF, but that doesn't stop them from loading On.

Sheesh… TIA for any suggestions!

EDIT—Randomizing question answered, here…
viewtopic.php?f=138&t=525918&p=7412392#p7412392

Would still love to get any insight into setting new MPs to stay off when old presets are activated…if that's possible.

Second EDIT: Well, the answer in the other thread about not-grouped randomizers doesn't quite cover some other weird behavior with the triggers I made, so I posted the device+Presets over there, if anyone's interested. It's my favorite thing I've yet made with MSF, fwiw, definitely a sort of mini-synth, yet targeted at a certain type of sound:) Vojtech returned it as too "crazy", so it's a cautionary example if anything.

[MeldaProduction]

Zentralmassivsound: Actually everything is optimized for Intel CPUs, which are complete dominators on the audio processing market. AMD CPUs are more for games really . I wouldn't expect to have a huge effect, but perhaps it does have.

[bmanic]

Hmm.. I'm having second thoughts about the whole "lets build mini-synthesizers instead of presets" thing for the MSF easy-front panel thing.

I just remembered a discussion I had with the Camel Audio guys back in the day. Most users don't really care if they get a "whole new little synth" instead of a preset and variations on presets.. as that's ultimately what the vast majority want to use. I was opposed to this school of thought back then and wanted Camel Audio to implement a much more complex front panel with many more macro knobs and things but I now later have realized my folly and their wisdom.

So, yeah.. my point is: I think the default template for builders is silly. The only part I can agree with a little, but even that goes against the philosophy above, is the Effects tab.

I really do think instead we should be building presets with a few (or many) macro controls.. but only controls that are relevant and useful to the actual sound itself. Same goes for any included effects. There should only be a tab for that particular effect and perhaps only a few knobs that are relevant.

MSF is NOT a "build your own synth" thing. It's far too limited for that. NI Reaktor demonstrates this nicely as most people just load up devices and play built in presets instead of actually using them as new stand-alone synths. Only a tiny little percentage of people do the opposite

Thoughts on this?

[David]

Hmm.. I'm having second thoughts about the whole "lets build mini-synthesizers instead of presets" thing for the MSF easy-front panel thing.

I just remembered a discussion I had with the Camel Audio guys back in the day. Most users don't really care if they get a "whole new little synth" instead of a preset and variations on presets.. as that's ultimately what the vast majority want to use. I was opposed to this school of thought back then and wanted Camel Audio to implement a much more complex front panel with many more macro knobs and things but I now later have realized my folly and their wisdom.

So, yeah.. my point is: I think the default template for builders is silly. The only part I can agree with a little, but even that goes against the philosophy above, is the Effects tab.

I really do think instead we should be building presets with a few (or many) macro controls.. but only controls that are relevant and useful to the actual sound itself. Same goes for any included effects. There should only be a tab for that particular effect and perhaps only a few knobs that are relevant.

MSF is NOT a "build your own synth" thing. It's far too limited for that. NI Reaktor demonstrates this nicely as most people just load up devices and play built in presets instead of actually using them as new stand-alone synths. Only a tiny little percentage of people do the opposite

Thoughts on this?

Yes, definitely:)

The very thing I love most about MSF, which has captured my attention now for FAR longer than any other new synth I've ever explored—which is many!—is that for me, it IS a Build Your Own Synth thing. And that's because it's the only one I've ever found that's perfectly suited to MY particular skill set, time, and inclination, unlike Reaktor or Max, and FAR more capable than synths with otherwise robust Macro possibilities such as Falcon or Halion…

It's not at all "too limited" for my purposes.

The next thing I love most about MSF, is that despite the understandable "keep it simple stupid" marketing decision the dev has made with his "requirements" and guidelines for the up-coming factory Device content, he HASN'T kept MSF itself simple. It's still capable of being used in wildly crazy and complex ways, ready for almost anything any given user wants to do with it.

While I'm fascinated to see what any other users who take on the task of building factory devices for all these faceless "most users" about whom so many people on KVR have so many opinions, I, frankly, don't give a damn about most users, or what they want, or how they use their synths. And I'm both delighted that I don't have to with MSF, and in complete disagreement with any who would try to suggest that MSF ought to be rethought to better suit "most users" or the "vast majority", or even those who have some "we should…" advice for how I use my tools! What I want from these future factory devices, and from this thread, is whatever I can glean or steal for use in MY devices, for my own use. So, presumably, I'm not "most readers" here, either. So what?

I'm also a passionate fan of Reaktor for several decades now, btw, AND still love it. But I completely fail to see anything pertinent to MY use of it, or to my love of the incredible User Library, in any guesses about how "most people" use it, either. Apparently most people ONLY use presets on ANY synth and have done only that ever since synths could be bought. Recall the tales from Yamaha about how 99% of the DX7s that came in for service had NO user-saved presets? Presumably only a tiny little percentage of those users had SYSEX'ed their creations and sent the thing back initialized.

So what? Should Yamaha have removed the option to save user presets as a result?

I hope I've made my position clear:) Thank you for asking.

[MeldaProduction]

bmanic: I think it isn't really limited or anything, it's more like "high level". Instead of playing with micromodules like in Reactor (which in effect is unusable for 99.999% people), here you can actually build your custom synth in a few clicks. Anyways I understand the skepticism about the devices. However you can always just use global presets, problem there. I'd just like to get a bit further than that oldschool approach, which is easy for design, but doesn't really exploit all of the possibilities here.

I like to think that the global presets are just, well, presets of a synth. And devices are like kontakt instruments.

[bmanic]

I like to think that the global presets are just, well, presets of a synth. And devices are like kontakt instruments.

This is where I think you'll be doing yourself a disservice.. and I don't think either of you two understood my point.

Yes, MSF is capable of building a sort of DUPLO synths (if Reaktor is LEGO then these things are DUPLO) but my point was not if it's capable or not. The point was, will users actually use all the power? Is it worth the investment from the creators? In my experience (and I have a lot in this area as I've tinkered with lot of various low level stuff that has later been packaged into various higher level modules) the answer is no.

Contrary to @David here I'm NOT talking of my own preference. I'm talking in general terms and usability. I'm thinking of the musicians, producers and the majority. Why is this important? Because they are the ones that either create or doom any market. There's no real market if only 10 people buy a few banks.

I'm trying to brainstorm here about what is smart and what isn't when it comes to the hopes of a future marketplace for MSF. In my experience, what has been suggested is not the optimal way to go. Actually it looks to me like nobody has really put any real thought into this at all.

In my opinion the "device" layer is a wrong term and the wrong approach. There's only one main synth and that is MSF itself. It could either be the full thing (as in purchased and can be opened with the EDIT button) or a MSF Player type of deal. The next preset layer is what I'd call a "Macro preset", so yeah, kind of like a Kontakt library with macro knobs for the most important parts of the program. In this regard I agree with Vojtech's example.

However, if you look at the good Kontakt libraries that are actually being used and are selling properly, they all have one thing in common. They aren't too cluttered with crap that is irrelevant to the preset in question. Instead the libraries have only things enabled that are relevant to the current sound being created. The point is: These presets aren't all "mini synthesizers" with equal capabilities all the time. They are optimized to be just the right amount of exposed features per preset. There are exceptions of course but I can't think of a single hugely successful Kontakt library which is an actual complex synth in itself.

This is where I think the default template goes wrong. I think there's simply way too much stuff exposed. This causes some real issues from a design perspective, if these exposed parameters are enforced.

These are the problems that I foresee:

1) Each and every setting that has been exposed needs to be properly tweaked for the sound in question to make an illusion of a "high quality preset". Leave most unused modules at default settings and the value of the "device" will immediately be diminished. It will feel generic and not completely thought through. It will feel amateur, for a lack of a better word.

Thus having so many controls exposed will automatically cause a lot of extra work for sound designers who want to create premium content. Alternatively all content just ends up feeling not so premium and just hastily thrown together where almost everything is left at default settings. If this becomes the case, then why even bother? Why not have fewer specific controls but with an illusion of much higher quality content.

2) Will users actually benefit from all of these extra exposed features? If I'm looking to find a nice distorted bass sound.. and I scroll to a preset that is ok but isn't distorted. Do I really start searching for all the places that could help me distort the bass or would I just scroll to the next preset and try to find one that matches closer to what I'm looking for? In my case I'd probably tweak it as I know exactly how to do it. I'm an expert. This is not at all true for the vast majority of people. Thus all the added work in point 1) is thrown down the drain and simply not worth it.

3) I'm pretty sure the vast majority of people who purchase additional sound banks would rather pay for 200 well thought out preset that are concise and match their description instead of 50 half-assed "mini synths" trying to do tons of different things and "giving the user freedom" to tweak further.

I suspect the vast majority of potential customers are buying these things to create music, not to muck about with 50 "mini synth templates".

A nice "macro preset" with a few well designed controls for changing the overall timbre and tonality of the preset, plus maybe a few controls for a sequencer and specific FX. This kind of preset could take perhaps 30 minutes to 2 hours of work. Whereas making one complete "mini synth" layer with tons of exposed parameters and tons of effects, and not make it feel like a silly generic template, would probably take a minimum of 2 to 4 hours of work. Probably more if you need to then populate that "mini synth" with it's own presets.. perhaps 10 or more at least.

My point is: Going the "mini synth" way will take significantly more development time from any sound designer if it is to be done properly. Is this something the users will truly appreciate and use or is it a complete waste of time (and thus money)? I suspect it's a complete waste and an absolutely great way of making sure the MSF marketplace never takes off.

Having said that, there IS enough space for BOTH approaches. One could make single "mini synth devices" that contain for instance 128 preset sounds and sell that single mini synth for lets say the price of a normal sound bank.. like 39$ or so, depending on the amount of content. That would be fine. If somebody wants to make a massively complex pack with 60+ "mini synths" then sure.. by all means. I just hope the current default template never gets enforced. In fact, I hope nothing gets enforced. An exception may be that there must be a few common controllers assigned like Modwheel and Aftertouch. I wouldn't even enforce MPE controls because the majority of the customers will not have a MPE device and thus it should be up to the sound designer if they want to put additional resources and time on implementing full MPE support.

Anyhow, that's my point. If the current default template or anything similar, with tons of exposed controls, is going to be the norm or enforced then I foresee a bleak future for the platform. If the market will stay open and unenforced then I foresee that the market will decide itself what will sell and what will not (I suspect well made sound banks with minimum fuzz and confusion are the ones that are going to sell well).

[bmanic]

double

[Chandlerhimself]

bmanic - I think you're misunderstanding what MSF does and the idea behind devices. The things you're suggesting are how things work now. 3rd party creators don't have to make complex devices or mini synths. The only thing they really need is the global tab. Without the global tab users won't be able to adjust the volume, pitch-bend range, etc.I guess if you wanted to make your own 3rd party presets without that you could, but it seems like you'd get complaints. The guidelines are for people who want to make factory content for MSF, and even then they aren't anywhere near as strict as you seem to think. You could easily make a device with the global tab and then 1 more tab with 2-5 knobs to control the sound if you wanted. It would only take about 5 min to add the knobs, so it really wouldn't take much more time. The current devices aren't representative of the requirements. You could make something more complex or much less complex than those. There is nothing stopping anyone from making a 3 knob device, so if that is easier for you and you prefer that, you should do it.

The suggestions you had at the end are good, but I imagine that's how most people are already doing things. I doubt many people are going to create 20 "mini synths" and try to sell them all together. I'm guessing people will make 1 or maybe 2 devices and then make lots of device presets for them to sell. What people create will depend on the type of person making things and the type of sound they're creating. Comparing it to Kontakt now doesn't make much sense because MSF doesn't have a sampler yet. Most people will tolerate a violin that isn't adjustable, but they won't if its a supersaw.

Another thing to keep in mind is that in the future it will be possible to search all the device presets at the same time. You wouldn't have to look through 8 different devices to find a distorted bass. You could just search for distorted bass and see a list of everything inside of MSF that fits that description regardless of what device its in. This discussion of macros vs mini-synths happened months ago when the first version of MSF was released as beta. From what I saw there is a 50/50 split between those that want to program their own sounds and those that just use presets. Doing things only 1 way will probably limit your market. I suggested using the presets at the top of the synth for simple presets with macros and using devices for mini-synths, but that didn't happen.

There seem to be a lot of different types of people who use synths and samplers, so limiting things to simple presets could end up hurting the 3rd party market.

[Reefius]

There's probably an easy solution to this, but how do I assign the modwheel to do a simple vibrato?

Edit: I think I found it: add a vibrato module and modulate its depth parameter with the mod wheel.

Before I was using an LFO to modulate the pitch of each oscillator and then use the modwheel to modulate the LFO's depth. This kinda worked but it was quite an awkward workflow because I had to apply the settings to each oscillator. The vibrato module is a much easier solution

[Sampleconstruct]

A first attempt at creating a vocal synthesis sounds in MSF using an additive oscillator and a modal filter excited by pink noise.

https://youtu.be/YAamsl22J0Y