What's new

I9 9900k - No slave

mscp

Senior Member
Anyone running a large track count (500+) in VEP on i9 9900k without the need of a slave machine? If so, what's your track count at the moment? Reason I ask: I want to ditch ethernet handling (if possible).

Sorry for the silly question. I'm just VERY curious about it. :-p
 
I wish.

I have two of those CPUs on two different satellite computers. they are awesome and have allowed me to cut my buffers back considerably.

If you are writing full (pretend) orchestra with all the articulations, I doubt you can do that, but if you're not, maybe?
 
I'm getting great performance with that CPU using one articulation per track on a single machine (with and without VEP). Seems to me with a more traditional VEP setup the issue would be RAM limits though.
 
I wish.

I have two of those CPUs on two different satellite computers. they are awesome and have allowed me to cut my buffers back considerably.

If you are writing full (pretend) orchestra with all the articulations, I doubt you can do that, but if you're not, maybe?

I'm indeed writing full orchestra mock-ups. I've thought of building a massive, purge-all template VEP project file and host it on my PC, alongside my DAW (localhost connection). I think the CPU alone would be able to handle the pressure but I'm being a tad skeptical about it. The RAM is what I think I'd struggle with in my "dream" setup. lol. I just want to find a way to reduce the size of my setup so the room can have more space. I'm tired of seeing a multitude of cables, synths, etc. I just want to keep what I really need, and sell the rest. Right now, my studio looks like an old man's cave. :rofl:

But then again, with this new trend going on - where library companies are starting to make their own VSTs to host their samples on, purging can become a tad harder.
 
Last edited:
Here's an example of a fairly complex track that I sequenced in Pro Tools using 70+ instances of VEP, each hosting its own articulation/instrument. The computer didn't break a sweat with a 128 buffer in PT and 4x (512) for VEP instruments until I added some effects to the master channel at which point I had to increase the buffer.

 
I sequenced in Pro Tools using 70+ instances of VEP, each hosting its own articulation/instrument.
I heard now from some people doing this, but why? I would really want to know why you think this is better then hosting as few instances as possible. Imo hosting fewer instances reduces complexity and is in theory way more performant.
 
Here's an example of a fairly complex track that I sequenced in Pro Tools using 70+ instances of VEP, each hosting its own articulation/instrument. The computer didn't break a sweat with a 128 buffer in PT and 4x (512) for VEP instruments until I added some effects to the master channel at which point I had to increase the buffer.



I’m at work so I can’t really listen to music right now as ironic as it might be...but I’m considering bouncing tracks off to audio when things start to become tricky to handle in real-time with a buffer size of 128. I just wonder if this new work style would be a major roadblock in certain events. I’m carefully scrutinising what could make me want to go back to an Ethernet config, and justify it with a plausible argument.
 
I heard now from some people doing this, but why? I would really want to know why you think this is better then hosting as few instances as possible. Imo hosting fewer instances reduces complexity and is in theory way more performant.
Instead of using a template I made track presets with one articulation per track and only added what I needed. I don't really have the RAM to have have a massive template loaded at all times and I also like to be able to tweak the Kontakt instruments and save them with the project which isn't really possible with the typical decoupled VEP templates that people use. Despite the solid results I got working this way I've decided to go back to disabled tracks in Cubase and not bother with VEP for now.
 
  • Like
Reactions: Ben
Top Bottom