PlanetZ forums - scopeusers.com

Hi!

After reading some threads on the subject here on Z, I decided to try out the SilentWay plugin to control Modular instead of MIDI. I have made a simple test to compare this to regular MIDI note triggering in Ableton Live, and published the results here. I apologize in advance for some of the font colors, I'll have to deal with that later.

Since some of this has already been discussed in other threads, I'll just sum it up quickly:

Modular can be sequenced without MIDI, if used with the SilentWay plugins. I have tested the performance difference between Audio/MIDI triggering, and also between two different MVC modules (MVC A and MVC C) and three EG modules (Vintage ADSR types, stock, Flexor and CWm). Results are presented as screenshots from the DAW.
In the end, it seems that Audio triggering is faster than MIDI triggering, and seems to be more stable. Besides better timing, there may be other advantages in using audio signals for modulation, due to its greater resolution and sample accuracy. The trade off is that it takes a minimum of two audio channels to control a single voice, plus another for each "CV" controlled parameter.

Another important and unexpected result was the poor performance of Flexor's EG with MVC C, something one should be well aware of, be it a bug or just character.


I hope you find it interesting or useful. Any questions or comments can be placed here.

Thanks!

T

You should inform Expert Sleepers about your elaboration using their plug ins to control a software DSP modular system,
as they only point out to control hardware modulars. I guess that will boost their work and brings you some cheers.

I really enjoyed to follow your great research and work together the last days.
Thanks to you, tomylee, neutron and all other modular explorers.

I have always wondered why MIDI is handled in such an atrocious way by DAW software/operating systems when audio is much more data intensive and always sample accurate. Certain interfaces which send the data out early by USB and time stamp it get around that problem a bit but still it is ridiculous. (and you cant do that with scope) .

Sending hihat hits to external gear for example, they are never in phase with each other because of the jitter and the apparent loudness can change, especially when using a delay.

Now there is a new problem which compounds the MIDI hodge podge, everyone has a MIDI USB driver and thinks their device is the only one you would ever want to use, once you end up with a few of them things can really get bad.

I would like to see a new standard, It bypasses the crap MIDI subsystems all together, and uses the fat audio stream to send MIDI.

"MOA" = "MIDI Over Audio" It could even still use USB but the DAW would see an audio interface. plugins like silent way would be built in to the DAW, or provided by the interface company, and it would look like a normal MIDI interface to the user. Some interfaces might even be able to still be used with a firmware upgrade.

Many people must have MIDI gear and would love to get the tightness(or better) that old ataris and hardware sequencers have. 44.1 KHz sample accurate = 44 MIDI instructions per MS(and thats only 8 bit). standard MIDI at 31250bps = 4 MIDI instructions per MS.

Chris: thank you, it means a lot to hear this from you!

Neutron: totally agree, I have a fair amount of actual "gear" in the studio that just sits there for ages because it's just too much trouble to get it to work properly with the computer... But, at least for Modular, there is an option now, and I believe there's a lot more that can be achieved even before the new MIDI paradigm shift.

For example, there's a plugin in the ES suite called "SMUX". This is to be used when working over ADAT @96KHz (SMUX), to compensate for the loss of half the audio channels (as you know, this will send 4 channels of 96KHz audio over 8 ADAT ports) Here's what the manual says:

"The Silent Way SMUX plug-in gets around this reduction in channel count by providing a complementary multiplexing scheme, which takes 4 channels of CVs and combines them onto 2 channels of the double-speed audio stream. When the ADAT output is then de- coded by a normal-speed interface (such as the ES-3 mentioned above), the channels are demultiplexed, and so you get your 4 channels of CV back again."

If it can be done for analogues, why not think of a single stream of control, that can be "demultiplexed" (love this word), so that one needn't dedicate so many audio channels for control? Maybe 44.1KHz at 24 bit is overkill for MIDI-like control, but if signals can be modulated and demodulated...

Just imagine, for a start, a S|C VSTi, precalibrated, with no knobs or parameters, just dedicated to ModIV control, generating Pitch and Gate CVs on a single audio output, received and demultiplexed by a dedicated module that feeds the Oscillators and EGs- that alone would already be half the audio port cost. And, since we do not have to conform to the "CV" paradigm, I think that a lot more information could be modulated into an audio stream that Modular understands. As long as it's quick and accurate, at least to Expert Sleepers' standards.

VoiceControl even provides the possibility to mix different signals to each output. I was trying to get Pitch and Gate on a single output, but I'm not quite there yet... Soon...


Much peace to you all!

T

Yeah in that case silent way is still encoding as analog, i ts very wasteful, you could encode all 16 channles 2 times over in the time of one MIDI message, of course the perfect analog option should be there as well. for modular stuff which needs to be pure. MIDI over audio could carry a lot more.

Well I am haqving zero luck so far, but instead of getting frustrated
I just keep double checking and trying something else.

Tau could you check and see if I am missing something..?
I have all MIDI routed to the VI, and that is the MIDI track.
But it also acts as a master out for monitoring purposes.

In the lower left of the VU you'll see a little white rectangle, which is
a MIDI activity VU. I have recorded 16th notes hoping this is what the
sequencer needs to send to keep the plug in activated and working.
I hear zero sound coming from the Modular Patch. I am using MIDI still
to trigger the notes from my controllers, and the audio ins is carry the
SWave stuff.

Jimmy, sorry I missed your post...

It looks like everything is in order, I can see the meter on the VoiceController track sending the CV info (left, Pitch, and right, Gate), and the patch looks OK.

You should assign the VoiceController track outs to an individual pair of ASIO outs (no monitoring is needed, you don't want to hear that), set gain at zero. Treat it like a buss, and send directly to Modular. No other audio should come through these outs.

For troubleshooting, write a few MIDI notes on the track and play them through (could be a MIDI monitoring problem)...

Maybe it's something to do with Reaper, or its routing?

Let me know how it goes. It's working perfectly here, been experimenting with different EGs and Gate modules. There's a problem, not unknown to Modular which is that new notes do not retrigger the Amp EG if the last key is still pressed. It needs a small space between notes in order to trigger, say, a 16th note phrase. But it also happens with MIDI triggering...

I am still fighting with this because I am a moron.
I will get it working next week when my current priorities are set.
Thanks again to everyone.
I can even see where the Analog synths will benefit...

It is very noticable to my ears.
The Portamento is incredible now, and the filter having audio
track it goes much lower than I have ever heard before.
I am very thankful and extremely impressed. Somebody from S|C
needs to see this.
But I really need some polyphony now, and my learning of the 2 x knobs
on the Silent DC track are being troublesome, but Tau,....my Brotha'....
You are now a God Amongst Men, and Tomylee too, but your patience
and tutorial were well written and deeply appreciated...

So it seems polyphony is somewhat of a problem (I can see this being a big problem for a whole drum set), but timing is rock solid. Always a trade off.

My question then- aside from the better filter response, any other improvements or degradation people see?

I'm asking because I don't own the FleXor modules (yet), but was wondering if you find yourselves using these modules even more since they exploit timing improvements using audio-rate control instead of MIDI, especially for clocks. I'd really like to hear how these two (floxor & SilentWay/Volta) work together.

I'm at the cusp of moving beyond learning about modular to begin making some big patches, but before I commit to building them, I might throw in moving to FleXor modules and using software CV control. In my mind, that would help working my growing analog modular into my rig more seamlessly later on. And if I'm going to cross that bridge someday, might as well cross it now.

There's several Flexor Modules I use. Filters, shapers, control smoother and certain envelopes.
Also Silentway gives the portamento a great boost. It sounded like Kontakt before, but now much nicer.
When you use multiple Oscillators a 6db LP or Control Smoother can be inserted pre Osillator and be used for applying seperate amounts of glide to each Oscillator for truly analog ( Oberheim ) sounding portamento where each Osc gets glide amounts, FAT & Wide.

Thanks a bunch, Jimmy. I like the sound of your unison like technique. And I've never heard anything bad about FleXor. I was just curious if the paradigm shift of FleXor vs. standard modules was enough that I should dive in now. Or maybe there'll be a FleXor IV released once the SDK for Scope 5 comes around? Hard to tell....

@j9k I believe Flexor is a worthy addition to your arsenal, it has some great sounding modules (Oscs, FIlters,...) and tools you won't find anywhere else (I'm thinking along the lines of control smoother, ramp divider etc...). You also get faster envelope generators, which is useful in some cases, but lately I've been dropping the "Gate CV" connection, and modulating the VCA directly from the DAW, which means that the EG is defined on the plugin.

For sequencer modules, using a ramp signal is the only way I know to properly sync the DAW and Modular, and that hasn't changed with "CV". In a way it's exactly the same, only instead of a Ramp generator VSTplugin (hint ) you record a ramp signal and dedicate an audio track and output for that. Flexor sequencer modules take this ramp signal and are very accurate, unlike stock modules, which need a clock input. However CWModular's Seq modules also take ramp input, and appear to be just as accurate (haven't "measured" it, though). Maybe you can try these for your ideas and see if the results please you. I'd be happy to recreate your patches here with Flexor modules, so we can see how they differ.

Tau wrote: For sequencer modules, using a ramp signal is the only way I know to properly sync the DAW and Modular, and that hasn't changed with "CV". In a way it's exactly the same, only instead of a Ramp generator VSTplugin (hint ) you record a ramp signal and dedicate an audio track and output for that. [...]

@j9k, @tau, actually you can use silentway lfo, choose saw and use is as a ramp driving anything rampdriven in modIII/IV/flexor, which is nice, the cool thing about that solution: ramp is always in sync with your bpm aslong as you tell silentway lfo plug to be synced with the transport, which works flawlessly. you could even add some groove to the ramp by slightly amplitude modulating it with another lfo (in modular) before it goes into any flexor sequencer etc

to all thanks for the cheers, did not know that this thing has created some waves already...

I tried to convince SC about this, and I had a nice chat but after all I am not bob moog, they would listen to him immediately. I was asking for more adat ports in their hardware (as alot of audio I/Os will be needed in future) and general improvements towards real sync audio control, thus also trying to convince them to make gate signals sync domain, not Async (whatever an async signal really is, i just know you cant trigger devices with gate inputs into audio range! (but I would like to do that nasty thing (aaand: I can do that with an eventide, like triggering a sampler faster and faster until the attack creates an audio range waveform, though in the sampler case it works only to about 800hz, neither there they are meant to do that, probably kyma would do that for me, but I cant buy it all, im not rich, and I have limited amounts of time, as we all.)

Lets all dance around the grave of midi No life can be digitally determined, but its better to determine it with 24bits rather than 7 bit midi cc power we have enough after all with xite

Hi TommyLee, great idea to use the LFO as ramp, I'll be trying that in a minute. This has been a great step up for Modular, and you are the one who started this with your Modular configuration file. Been having loads of fun!

A way forward would be for S|C to come up with a VST plugin that could generate a signal to be demultiplexed inside Modular, so we could have proper poliphony and multiple control signals in a single audio stream. No need to stick to the CV paradigm if it's not in fact analogue, and besides, Silent Way and Volta have got that covered.

I believe there's some interest stirring in this, but S|C users are still few, and maybe even less are using modular, let alone analogue gear. But as soon as you hear the difference, you can't wait to get rid of MIDI altogether S|C have a real winner in their hands with this one!

And now, I'm going to LFO for a while

Cheers!

Tau wrote:A way forward would be for S|C to come up with a VST plugin that could generate a signal to be demultiplexed inside Modular, so we could have proper poliphony and multiple control signals in a single audio stream. No need to stick to the CV paradigm if it's not in fact analogue, and besides, Silent Way and Volta have got that covered.

jep you are right about that, if its really possible without degradations or limitations, I want to have audio to be able to modulare other audio up to 20000hz, ideally, if you put many streams into one audio channel you may get limited, apart from that I was talking about this issue with my tech, I supposed there should be a new standard in analog audio control signals, that could be polyphonically, using for examples 16 frequecies for FM and a demodulator on the other line - it should be doable, but it may add some lateny...well we did not talk about it much, as I stick with the direct thing, one channel for each voice, one for each gate.

Btw something like silentway should have had to come from creamware/SC, since digital voltage is what they are masters of...too bad the companies owning that technologies had somehow bad charma, well, I dont know for sure, maybe just bad marketing, or bad drivers and stability problems, hell I dont know why, but thats another matter - glad things work out now, and I recently bought Modular IV and scope 5.1, was worth it!

thanks all! I also have MAX for Live that I will do some experimenting with (since it has access to much of the inner workings of Live, including the transport, clocks, etc).

But so there's no confusion, I'm just "jk", not j9k, who has developed some modules and things here...

j9k is a true geek master. I came after he moved on it seems, but he was always kind enough to respond privately and help out.
I still can't figure out some of his incredible modules, but they do work in x64 Scope Modular IV shells and II/III.
So I have an endless decade of discoveries ahead of me, and I hope jk becomes another j9k.
Just in case some wish to visit his site.

http://www.google.com/url?sa=t&source=w ... lh2Nvd8lYw

Tau you know I am lame so please continuing sharing your discoveries here, especially with the LFO thingamajig....