Mulab 8 has allergy to VPS Avenger

Official support for: mutools.com
Post Reply New Topic
RELATED
PRODUCTS

Post

mutools wrote: Tue Nov 05, 2019 7:40 pm Replacing "Threads" by "Cores" would be technically wrong for that's not how computers work. "Threads" is the only correct term there. As Dakkra also already explained well earlier in this thread. (Threads vs Cores)
OK, but it's misleading as people might type in the number of threads that is specified in CPU specs...

Post

Btw, do the overloads in Mulab, that occur during tests for instance and sound like very strong distortion, damage the processor, sound card or other components? :?

Post

Not unless your amplifier over-drives your speakers/headphones.

You seem a bit new to the digital audio scene. I recommend you explore the general forums a bit. Lots of good tips for newcomers! :tu:

Getting started forum: viewforum.php?f=74

Computers and system setup forum: viewforum.php?f=16

...and pretty much anything else is worth glancing at: index.php

I'm an addict to this forum (clearly, I post nearly every day). Even though I don't talk on many threads, I read about 20-30 threads a day. It helps me keep up with what's new in the digital audio world.

For the record, I'm no professional. I have gone as far as putting together a simple home studio, but I make no income from my audio work. I just love it, with a passion!

Cheers!
My Setup.
Now goes by Eurydice(Izzy) - she/her :hug:

Post

heks wrote: Sat Oct 26, 2019 10:55 am Not a very scientific test, but quick. Instrument used Lush 101 (it just kills hosts) preset was "Beautiful Pad", a triad play'd over four bars looped.
Mixcraft 8, five instances before sound breaks up.
MuLab 8 , three instances before sound breaks up
Mixcraft ( at this time) seems to be at a similar price to MuLab so more relevant competition than say the big names.
I just downloaded Lush out of curiosity. I can't find that Beautiful Pad preset you mention, but after trying various presets it is clear that Lush's internal MC support also helps a great deal, just like with Repro-5.

Post

dakkra wrote: Tue Nov 05, 2019 9:26 pm Not unless your amplifier over-drives your speakers/headphones.

You seem a bit new to the digital audio scene. I recommend you explore the general forums a bit. Lots of good tips for newcomers! :tu:

Getting started forum: viewforum.php?f=74

Computers and system setup forum: viewforum.php?f=16

...and pretty much anything else is worth glancing at: index.php

I'm an addict to this forum (clearly, I post nearly every day). Even though I don't talk on many threads, I read about 20-30 threads a day. It helps me keep up with what's new in the digital audio world.

For the record, I'm no professional. I have gone as far as putting together a simple home studio, but I make no income from my audio work. I just love it, with a passion!

Cheers!
So, my speakers or headphones are all that may break?
While the sound becomes utterly distorted, it is not loud, though.
I was more worried about my processor burning up or something like that.

No, I am not really new to digital audio. But I avoid the technical side of it where possible as I consider myself more of a musician. As long as everything works the way it should, I don't care how and why it does :)

You seem like a friendly, helpful guy :hug:

Post

Alright, I'm doing the LuSH test now. Using the demo version with preset: "Single Zone/Poly/Pad/BigSection"

The number pertains to the number of instances able to be used without having overloads.

Mulab(8 Threads) + Lush(single-thread): 8 instances
Mulab(8 Threads) + Lush(multi-thread): 8 instances
Mulab(1 Thread) + Lush(multi-thread): 7 instances (switching windows causes errors though, so more like 6-7 instances)
Mulab(1 Thread) + Lush(single-thead): 1 instance

My computer specs are in my signature. This is using the Focusrite ASIO driver for my Scarlett 2i2. 48khz Samplerate with 128 sample buffer.

e-crooner, what CPU do you have? Is it a modern AMD or an older version? AMD didn't do so well with multi-core until the Ryzen series.
My Setup.
Now goes by Eurydice(Izzy) - she/her :hug:

Post

Doing the same test with Bitwig. Bitwig always has multi-core enabled.

Bitwig(individual sandbox) + Lush(single-thread): 8 instances
Bitwig(individual sandbox) + Lush(multi-thread): 9 instances
Bitwig(group sandbox) + Lush(multi-thread): 9-10 instances

My conclusion is that Mulab is not best at multi-threading. However, it's not a straight-forward architecture like Bitwig either.

So, with heavy VST's I can see the issue now. Falcon and MSF (both heavy VST's) however show no preference. Perhaps it's a type of optimization. Could have to deal with SSE or AVX optimizations in the plugins, but I'm not certain nor am I an expert on that. I know some plugin developers go down to the assembler level to optimize their code, which is insane.

Once again though, I'd chalk this up picking the right tool. Currently, Mulab isn't the right tool for instance spawning of heavy VST's. However, as I discovered earlier, Mulab is much better about chains of VSTs when it comes to distribution.
My Setup.
Now goes by Eurydice(Izzy) - she/her :hug:

Post

Interesting that Lush's MC support setting doesn't influence the number of instances when you use Mulab's MC support.
On my system setting Lush's MC support on always yields better results.

What exactly did you do per instance? 1 note? A chord? Every additional note matters with that synth :hihi:

My processor is an AMD Ryzen 5 1500X. Nothing spectacular, but I am happy with it. It is enough for me as long as I stay away from modern plugins.

Post

Dakkra pls explain me the issue you see.

If you are saying that MuLab's mc audio processing is not very compatible with mc vsts then that's what i'm already saying all the time. If you see another issue pls elaborate.

Post

@e-crooner: F Major second inversion.

@mutools: Mulab simply handles less instances of LuSH than Bitwig (getting overloading/distortion sooner). Not an issue per-se, just how it is. I don't know how Mulab handles multi-threading, but Bitwig does it per track. Meaning a track is assigned a thread (and thus all VSTs on that track are on that thread).
My Setup.
Now goes by Eurydice(Izzy) - she/her :hug:

Post

dakkra wrote: Mulab(8 Threads) + Lush(single-thread): 8 instances
Bitwig(individual sandbox) + Lush(single-thread): 8 instances
8 is not less than 8, right?
I'm sorry if i'm missing something.

Post

And if Bitwig is simply track = thread then please also test this: 4 tracks with 2 layered single-threaded vst synths per track. Thanks to MuLab's distributed processing that case should be handled more efficient because it will process the layered synths multi-threaded whereas Bitwig will only use a single thread for the layered synths per track.

Post

Bitwig(individual sandbox) + Lush(multi-thread): 9 instances
Bitwig(group sandbox) + Lush(multi-thread): 9-10 instances

That's what it's in reference to. Bitwig multi-thread + Lush multi-thread allows for more instances (in this case)
And if Bitwig is simply track = thread then please also test this: 4 tracks with 2 layered single-threaded vst synths per track.
Bitwig overload with just one track with 2 layers. Mulab does not, of course.

The discussion here is now about how enabling multi-thread on Lush increases performance in conjunction with Bitwig's multi-thread. This is in contrast to Mulab where it is equatable or worse when Lush is using multi-thread.

For a single-thread VST yes, Mulab is better at distributing. However, Bitwig handles VSTs that do their own multi-thread along with Bitwig's multi-thread in a manner that increases the number of instances capable (9-10 instead of 8 ).
My Setup.
Now goes by Eurydice(Izzy) - she/her :hug:

Post

dakkra wrote:That's what it's in reference to. Bitwig multi-thread + Lush multi-thread allows for more instances (in this case)
Yes i already said on page 1 of this thread that mulab's mc is not very compatible with mc vsts.
I had already taken note of that and maybe will research an extra mc option in some future version.
Bitwig overload with just one track with 2 layers. Mulab does not, of course.
What i expected.
The discussion here is now about how enabling multi-thread on Lush increases performance in conjunction with Bitwig's multi-thread. This is in contrast to Mulab where it is equatable or worse when Lush is using multi-thread.
I think the discussion is about mulabs audio processing compared to other daws. If you focus on the worst case (mc vsts) then indeed mulab is not the best, but at the same time mulab performs better in other cases. Right?

Btw can you pls give me a percentage of how many of the popular vsts use mc nowadays? I think you have a better view on that.

Post

NB From a practical pov: If a user intends to often use cpu-heavy multi-threaded VSTs he can choose to lower MuLab's Num Audio Processor Threads a bit so there are some spare threads for the VSTs, cfr http://www.mutools.com/info/M8/docs/mul ... setup.html

Post Reply

Return to “MUTOOLS”