I'm sure you guys know the situation... creating software for different platforms requires multiple computers. Not only do you need a Mac and PC, but you need multiple versions to test different operating systems and software updates. A few years back, I got tired of switching between multiple computers to get the job done, so I created triple boot PC/Hack/Linux machine. It works well. No complaints. Then I decided to upgrade my GPU to an RTX3090 for faster 3D rendering. It's quite a beast! The issue is MacOS does not work with Nvidia chips, so every time I wanted to boot up macos, I had to unplug my GPU. I also started to get annoyed by not being able to use multiple OS at one time. Powering down MacOS to use Windows can quickly get old.
Since I'm relying heavily on virtualization to run my mini datacenter, I took a shot at creating a virtualized workstation. The result?

One computer that simultaneously runs Monterey, BigSur, Win10, Win11, Linux desktop and Ubuntu server!
I'm still in the testing stages, but it's working wonderfully. I've successfully exported plugins from HISE on each platform.
The system is running on Proxmox. Not to be confused with things like VirtualBox or Parallels (which are type 2 hypervisors that run on top of an existing OS), Proxmox is a type 1 hypervisor. That means that your computer's actual hardware is divided up and given to each virtual machine.
Since the allocation of hardware is sectioned off and given to a Virtual Machine, that means you need a computer with a bit of horsepower to use multiple operating systems at the same time, but you may already have what you need. Or worst case, You only run 1 OS at a time with full resources.
By using some utilities like VirtualHere (share usb over ethernet) and Barriers (use one mouse/keyboard to control multiple computers) I'm able to use one set of hardware resources (usb sound card, dongles, etc) and share them with each computer. Each system sees the devices as if they are actually attached to the computer. No layer in between. The only downfall is since it's "directly attached", you can't share your sound card with multiple machines simultaneously.... I'm mean, that sounds like a nightmare anyway! hahahh
Right now, the Mac virtual machines feel like they have a little bit of lag, so I'm going to have to install a GPU for those machines to use.
If you guys are interested, I can post some details on building a machine like this.

It looks like you're trying to hook up a gain reduction meter to a dynamics module.
here ya go! in 13 easy steps
I'm using "right click" auto complete as much as possible.

#1 - Add Dynamics module

#2 - Create a reference to the module

#3 - Create a slider and add reference

#4 - create a timer

#5 - Create function for timer

#6 - create variable to hold a reference of gain reduction

#7 - convert reduction to dB

#8 - set value of slider to gain reduction

#9 - set slider mode to decibel

#10 - Start Timer

#11 - test that reduction meter is working

#12 - set slider image to your strip

#13 - test that strip is working

Merry merry everyone!

I would love to have SideChain inputs available in the bus routing.

https://docs.juce.com/master/tutorial_audio_bus_layouts.html

Finally had a second to make a quick video. I apologize for the poor quality and zero planning... but you get the idea

https://youtu.be/WuqUNGE2tx0

I just ran into this when opening older projects on my new rig. Images and binary are way below 50mb.
Did a little searching and found that this may happen if Visual Studio doesn't use the 64 bit toolset. I think it's some kind of global change between VS2017 and 2019.
Manually setting the architecture in Environmental Variables fixed the problem for me.
https://phoenixnap.com/kb/windows-set-environment-variable

@Christoph-Hart said in Audio Spike when loading a Neural Network node.:

thanks @Dan-Korneff for the contribution!

It's nice to feel useful for once
Got a few more enhancements coming soon...

@Christoph-Hart https://github.com/christophhart/HISE/pull/787

@Christoph-Hart I make a PR today. Also have a fix for the bypass being ignored.

@Christoph-Hart you might want to consider adding a HPF as well. Many models contain a DC offset and filters are built into the NAM player for this reason.

I've added a filter to remove DC offset in the Neural node and it seems to have solved this issue for me. Gonna do more testing this week to be sure.

Stacking two 1-pole (6 dB/oct) high-pass filters is definitely an option, but it doesn’t behave exactly like a single 12 dB/oct design. The slope doubles, but the overall response (phase and the cutoff alignment) will be slightly different.

The SVF can do a true 12 dB/oct response that's mathematically “neutral”.

If you need more control over the resonance, a Sallen–Key filter is another solid choice for a 2-pole high-pass filter.

Maybe an X-Fader would work for you.

HiseSnippet 1179.3ocuW80SiaDDecBl.oW6wIcR807HWEBECgbf5CGPfzF0CH5LEcugVVulrJ16ZYuFHspu2OV8iz8MncF6MI1IAZTp5EjPYm+r6uYleyra5GqX7jDULwp90ih3DqWY6NRpGzY.UHI8NiX8c1WPSz73F4hNcTDMIg6Qrrp9Sn.qMWij84Ke3TZ.Ux3SEQH2nDL9GEgB8To8O9WDAAcod7qEgErt0w8XJYGUfJEvSU6ljHJaH8d9kTzrJ1jellLfX8C1Gr29rVdu2au8N3n1LZa9Q9Teem88Z0p8gG4r+QzVG1l27.h05m6IzpXWMUySHVqcpxaj6.0ix7C3FQh3t.Ntvg3Bmbt3tp.OLDQojNCDAd8GmnRHvtzeZZqZdZ6s1WH7DSjOM8sUlhFS8nXBzpRY3UsD7bJBulEf2BfjUAHsVNjdisKKVDompAwy2X2SBUSeJTmJBkbaIV+kcGEXfTuaHcHuaLrXhCa2tYycZ.+6c+nepjoEJYCk7Rklekb62U+2quY8+ndiYU46uPc3wDqBB3wKTMRMheIG2VlFdGOdmFOPCR4SLDB+x4T6mOmVrjyxi5BFpj8jB8UQb4yQDHlTE7ses2YTMEKDFYfcQ7Xs.gf0Y7Gf1f7xxl1mwSFpUQPivb0Lfsn7RCn5xTHrQyn.xAkpaXwQlHziJ1HNKuZskiW8xz9kEtuwtuPyFrX7VYA3ExZ+eiWSW52ZetuOmomB10r694udsj0MsjvYhG9qLKaz8yNKrY7OsmR7ih4QzX90p9AzQamPCiB3eBv3NMtKPwF5J9M97cKQ4P3TzhsYCnRIOHYUZpVeoyJNuXVg7IUpVHu+BpNV7DzcdYZnKLumw6XPGHypB1akutItF4HtboW1h+F9XT5fqsLJcFqrPa4kb8ip3gYkCy2IV0xy8IYI4a8exgbRPf5wNpvHggwB0fLY8UAihFnjBFJJ2hwH8jPUJbHF3BWMcMUDfTb2zDXng2URWv3r6CspdCONIaiqY2bW3OfxeoxCTUqKkAI1Q8o5AXuCNGBHm73cYSHokvZQdR9EikTi653qPySb.pCwrAdAJVx5SiA2fQ53joJVUKLoJmzhqf1yJ4GcujaPsLZ.4FjTfNkcJn4lfXczzMrexGtSOlTJh1JKh.p0tFsyi+IN9xXeFjVYFj9ZjJUH1LfMu5Plw30wAW7w1rg8GgLNMF3QOhittlFeOWmcHEEXltK4YMMIkWZxE87voLIJe8s2kEnjIPBUU2dR3SdwslTHRvpx3U.XuPHM3d5.zKnOMiru7AWMOBmJTzN2g7GyKOEsDyH0ry2f415y39zz.8bxOIUqBgoAdSHDFdP4vuDY36KPuKZz7bhxax+IhwyPgmi93FpT5AvrorWkNg97V73L78RUkmuDUKOgBX7NEY4JWac7zxkssCLfXwkpkr7.OAXonpSZ8JwRGyEHmpTCCoYyOWsW370XZeHkEqt0LkAC0Mxj.zBYFUZS3Gv.qa3PdX1jWH7z7aYrxa0bNt2p539qpisVUGOXUcr8p536WUGO7e2Q7WNYFyfzVnuo+4YjPKqykT38FYCLH+SeJu.I

@ustk Sorry for the super late reply. Completely missed this one.
@Adam_G I just downloaded the latest AAX SDK and HISE on W11 and everything is working as expected.
Is it possible that you've compiled both the Multi-threaded and Multi-threaded DLL Libraries? That may cause confusion when exporting. I would delete the AAX build folder and compile the Multi-threaded library again. See if that makes any difference.

@Yarost https://docs.hise.audio/scripting/scripting-api/neuralnetwork/index.html

@Christoph-Hart said in Moonbase:

JUCE_USE_BETTER_MACHINE_IDS

https://github.com/juce-framework/JUCE/commit/120f9266dc881c81371109c69b37a0a246accdeb

Here's the fix