Multiple instances of same plugin

Searching the forums, I get the impression that what I’m trying to do is not possible. I’m building the “chord stack” for the main sound in Billie Jean. Each piece of the sound was created on the Yamaha CS-80. I’ve got the Arturia CS-80 and have recreated 3 layers (patches) of the sound, each using a preset I built in the CS-80 plugin. Each will play by itself, but 2 or more combined will not. Attached is a view of the wiring diagram. Is there any way to accomplish this?

How is it not playing?

From your wiring view this should definitely work.

1 Like

Yep, that should work. Is it possible there is an issue with the midi channels?

Does it show midi going through? Maybe use the midi monitor?

The effect is almost 100% audio dropout when 2 or more instances are wired. It’s more of a crackling effect as if the plugins were conflicting with each other. They’re all being triggered by the same MIDI IN block on channel 3. So MIDI channel 3 is triggering all 3 blocks at the same time. MIDI Monitor appears fine, though I’m not sure what a “problem” would look like. Connecting each plugin separately shows MIDI in the Monitor just fine. No CPU resource or RAM issue. If I disconnect any 2 of the 3 wiring paths from MIDI IN to the plugins, the audio output is fine.

I see what’s going on. It’s not a comptuer CPU problem, but a GP CPU problem. With one instance of the CS-80 plugin, the GP CPU is running 74%. With 2 instances of the CS-80 plugin, the GP CPU is pegged at 100%.

I have used multiple CS80s before with no problem. There is no way one instance should be taking 74%

So some of the possibilities are

  • misconfigured or poor audio driver
  • Too high a sample rate
  • Too low a sample buffer size
  • Computer too slow.
1 Like

I don’t understand this. I’m using the ASIO “audio device type” and the Universal Audio Volt “device” in the audio setup. Sample rate is 44100 Hz and the buffer size is 128, but I’ve tried all the way up to 2048. The computer is an HP Zbook 17 G5 which I believe is pretty over-spec’d for GP. I even ran through “The Ultimate Guide to Optimize your Windows PC for the Stage” with no change. Seems like this should be something simple. It doesn’t happen with every plugin, but I believe every time it does happen, it’s with Arturia’s plugins. This is the first time I’ve tried to “layer” the same plugin with multiple instances, but I’ve had quick 100% spikes in the GP CPU with single instances of other Arturia plugins.

Seems that any plugin that hits 40% or higher in GP CPU starts to crackle. It’s not just the CS-80.

My impression is these type of issues are almost never a flaw in GP.

What happens CPU wise if you use that plugin as a stand alone?

Have you made sure you have the latest driver for your audio interface?

2 Likes

Checked drivers. All are current, especially the audio interface driver. I’m not sure I can test the plugin as a standalone as I never installed the standalone versions.

Are you using Analog Lab or the actual synth plugin?

Does this happen with other Arturia synths?

Other synth plugins?

There’s definitely something wrong. My first guess would be the driver. Have you tried the ASIOforAll driver?

Yes, other Arturia synths, but not as bad as the CS-80. Perhaps that one is a resource hog. I’m using actual synth plugins, not Analog Lab. I also noticed that even if no “wires” are connected from the MIDI In to the plugin instances, simply having them in the rackspace increases the CPU. If I switch off that rackspace to a simple piano rackspace, the CPU drops. I guess this should be obvious to me, but the more complex the rackspace, the higher the CPU. But I’ve got another rackspace with three Prophet -5 V plugin instances and a Cherry Audio plugin and the CPU is only at 25%. Just having a simple rackspace with a single instance of the CS-80, jumps the CPU from 7% (blank rackspace) to 17%. Adding 2 more instances of the CS-80 brings the CPU up to 34%. Then, playing them all at once pegs the CPU at 100%. But the crackling begins with any rackspace (playing any rackspace) that brings the CPU up to around 40%.

No, I haven’t tried the ASIOforAll driver. I just have the ASIO driver. If I download and install the ASIOforAll driver, will GP pick it up as an option to select?

Just discovered another plugin that really spikes the CPU: Toybox Thump One. Even worse than the Arturia plugins.

I guess give the specs of the computer. Maybe it’s just less powerful than you think (?).

If your cpu is underpowered, but you have multiple cores, you may benefit from running multiple instances of GP to spread out your cpu usage to other cores. (?)

Exactly! This would be extremely helpful!
Just the model name doesn’t tell the actual specs of the machine.
From what i read, it could be just a i5 at 2.4GHz with 16GB of RAM, which wouldn’t be a real GP-rocket, but it could also be something completely diffrent.
Most of the G5 series don’t have a base frequency over 2.5GHz and also 16GB of RAM could become short, depending of the number and type of sample based plugins that are used.
So, please provide more details to narrow down your issues.

I duplicated your rackspace, but had to use Analog Lab V since I don’t have the full version of CP-80. I selected different presets for each of the 3 instances. I played each of the presets with the other 2 instances bypassed and then played all 3 at the same time. Analog Lab is well known for being CPU-heavy. Idling with all 3 instances active, the CPU meter in GP bounced around between 40% and 45%. When playing with all 3 instances active, GP reported around 55% CPU. The audio was clean, with zero dropouts or faults of any kind.

ThinkPad i9, 32 GB RAM, 7 TB SSD; Windows 11 Pro; PreSonus Studio 1810c; GP5

ASIO driver: 48K at 128

Note that I also had an instance for 4drX’s VirtualKeyboard loaded to drive the 3 instances of Analog Lab, so I didn’t have to uncover my NI S61 Mk3 MIDI controller.

What’s really interesting about that test is that generally Analog Lab uses way more CPU than the individual plugins as a matter of course.

1 Like

I suspect your cpu is not quite so powerful.

But, there are things you can do to reduce CPU strain.

If you can tolerate a 256 buffer size instead of 128, that would help all around (FWIW, I use 256).

The synth itself might give you options to reduce CPU without any real effect. Maybe you can reduce the polyphony (especially if chords are sustaining)?

There might be different “quality” settings. In live use, my impression is the lowest “quality” is fine. That really becomes more of an issue if you are recording.

Of course, you can try multiple instances of GP to spread out CPU hit among different cores.

There are probably other techniques you could use that I am not recalling right now.

Jeff

1 Like