Software vs Hardware

[muzicxs]

The advantages these network-based solutions like VE Pro have is that they offer the ability to almost instantly open up ANY project. If you work with virtual orchestra and using software like MIR Pro, you know that loading big sections can take ages. Having a server preloaded for all projects is such a timesaver.

[Jim Roseberry]

I don't understand why you are all so fixated on CPU power. I've been getting up on stage with a PC since about 2002, when they were still single core jobbies, clocked in MHz, and it had to provide every sound we used, apart from my vocals. If you are just using something like GP to play keyboards in a band, surely any old consumer grade laptop will work more than well enough, won't it? If all you're after is hardware-like latency, then we're talking about 10ms or thereabouts, which is a doddle with even a cheap audio interface. e.g. My sub-$100 Zoom U24 runs at about 5ms, which gives me round-trip latency of 11.8ms, which is well within the ballpark.

... the band wanted to expand into songs that required more playing vs triggering (Journey, Night Ranger, Kansas, Styx, Foreigner, etc). I had to become more of a traditional keyboard player... and that meant being tethered to a keyboard.

You never thought to go down the Keytar route? I used one when I was performing solo, for about 12 years. I had a long MIDI cable and a headset mic so I could roam about the stage and jump off the stage into the audience if I felt like it. I decided to rule a line under it when we started doing NOVAkILL in 1997 but it served me very well as a solo artist.

I've measured my Nord's latency (MIDI note to when the sound is actually generated).
It's 2ms

I get that not everybody is ultra concerned with latency.
I've been chasing ultra low latency performance my entire career (most of my adult life).
I like tight/responsive timing.
ie: If I'm playing "The Grandeur" (nice piano library included in Komplete), it feels a LOT more immediate playing it at a 32-sample ASIO buffer size... vs 64, 128, or 256.

If I'm going to take out a software based rig (and the time/effort to make it happen), I want it to eclipse my hardware rig... in every way.

I'm probably not a typical example... as "high-performance computers" is my day job.

Hearing me talk about pros/cons of each keyboard, the wife tells me I should build my own.
A software based rig would essentially allow me to achieve that.
I just want it to be a step forward.


Regarding a Keytar:
I've owned a couple.
It would be hard to play two-hard parts for songs like: Carry On Wayward Son, Don't Stop Believing, Sister Christian, etc.
I never got comfortable playing one.

[IvyBirds]

So you're using an instance of Audio Grabber for each plugin inserted in GP?
Sounds similar to VE Pro.

It's similar, but it's freeware and awesome. Many people think its better. The guy that developed it wrote it for himself and then made it available to everyone

This is a good video that shows it in action.




I use it extensively both on stage where I use it with one computer and in my studio where I use it with multiple computers networked together

In the video he is using it without a monitor on the server which shows how seamless it is, and how everything is controlled from your main PC, however I use a monitor for each one

I think if I was going to build a live rig based around multiple PCs they would all be rack mount with a USB C port on the front and I would get one of the smaller USB C tablet monitors that could be swapped between the servers maybe on a pull out drawer

I really like the idea of being in control of how my cores are allocated. Each core is a different synth.

[Jim Roseberry]

Using GP as the main host:
Would I load each VST plugin into a separate instance of the Audio Grabber plugin... or would I load all VST plugins into a single instance of Audio Grabber?

[IvyBirds]

Using GP as the main host:
Would I load each VST plugin into a separate instance of the Audio Grabber plugin... or would I load all VST plugins into a single instance of Audio Grabber?

For each instrument load a separate instance of Audio Gridder "client" which will then look for a "server" which can be local to your machine or across the network

Then load the instrument into the Audio Gridder Client which will load it off the server

You can select if you want the GUI as well or just audio and MIDI

so if you want 5 different instruments/effects you would need 5 instances of the client of Audio Gridder which loads as a standard VST3, AU, or AAX for Protools in whatever host/DAW you wish

You can even load an instance in Logic or Protools and run the VST version in the server either locally or on a different machine with a different OS

[Jim Roseberry]

Using GP as the main host:
Would I load each VST plugin into a separate instance of the Audio Grabber plugin... or would I load all VST plugins into a single instance of Audio Grabber?

For each instrument load a separate instance of Audio Gridder "client" which will then look for a "server" which can be local to your machine or across the network

Then load the instrument into the Audio Gridder Client which will load it off the server

You can select if you want the GUI as well or just audio and MIDI

so if you want 5 different instruments/effects you would need 5 instances of the client of Audio Gridder which loads as a standard VST3, AU, or AAX for Protools in whatever host/DAW you wish

You can even load an instance in Logic or Protools and run the VST version in the server either locally or on a different machine with a different OS

Thanks for the explanation!
Downloading as we speak.

[SamDi]

Humanity is nothing more than a plague on this world...

If you think humanity is nothing more than a plague, then what to worry about? Apart from 20 times overkill by nuclear destruction, it is very unlikely that we are able to destroy nature. Nature is a very fluid and impermanent thing. It will go on - with or without us.
All effort to reduce overpopulation or environment destruction is to preserve our base for humanity existing in a good way. Nature, the universe or whatever doesn't really care we do.

Of course it is and that's exactly the reason. People are, by and large, completely unaware of their biases and totally unwilling to accept that fact. As I always say in these discussions, if you prefer hardware, that's perfectly reasonable but don't try and justify a personal preference by making stuff up to make yourself feel like your preference makes objective sense.

True

[BONES]

as has been pointed out, the host i mainly work with runs on one single core only.

But it's not actually doing anything, it's just serving up instruments and effect, which is where your CPU power is needed.

i run FFT based FX with latencys of 80ms and more. Up to 200ms in fact.

What? FFT is hardly cutting edge, I was using FFT based effects in Cool Edit last Century on a Pentium 90.

That´s for realtime play purposes. Not exactly ideal

So use something else, FFS! You're situating yourself so that you have to spend more money than you need to. That's just dumb and totally unnecessary.

I've measured my Nord's latency (MIDI note to when the sound is actually generated).
It's 2ms

That's much lower than most hardware synths. Scot Solida did a big test of a lot of the synths he owned at the time and found that some of them have latency up to almost 20ms. My bandmate plays drum pads that trigger drum samples on the PC and anything under about 20ms is perfectly fine for live work. Until you tell him what it is, then he'll swear he can feel it. So I just don't tell him.

I get that not everybody is ultra concerned with latency.
I've been chasing ultra low latency performance my entire career (most of my adult life).
I like tight/responsive timing.
ie: If I'm playing "The Grandeur" (nice piano library included in Komplete), it feels a LOT more immediate playing it at a 32-sample ASIO buffer size... vs 64, 128, or 256.

Again, I'll guarantee that you only feel a difference because you know there is one. In 2ms, sound travels less than 10 feet so if it was detectable, you'd hear a time difference between two instrumentalists playing on the same stage. But you don't.

If I'm going to take out a software based rig (and the time/effort to make it happen), I want it to eclipse my hardware rig... in every way.

Why, when the other advantages far outweigh any tiny difference in latency? Any time you spend setting up your software rig will be repaid after just a handful of gigs in how much less effort it is to transport and set up. In fact, if I compare the time I've spent getting my software working over the last 25 years to the time I used to spend doing the same with my hardware, I reckon it's probably only one-tenth of the time. Hardware always required so much more effort than software does. You had to find stands, or build your own, do all the cable work to make it run (audio, MIDI and power cables, with junction boxes, etc.), set up a mixer and effects racks/pedals. It took forever compared to plugging a MIDI controller into a computer and connecting that to an audio I/O device.
[quotew]Regarding a Keytar:
I've owned a couple.
It would be hard to play two-hard parts for songs like: Carry On Wayward Son, Don't Stop Believing, Sister Christian, etc.[/quote]
You don't have to use it for every song. I built a small pedestal stand so I could use mine with both hands when I needed to. It's just like your guitarist changing guitars between songs.

[Funky40]

That's just dumb and totally unnecessary.

ahh, i´m dumb ? .....thank you !

You're situating yourself so that you have to spend more money than you need to.

you care for my money ?
so another thanks is due from my side i guess.

But it's not actually doing anything, it's just serving up instruments and effect, which is where your CPU power is needed.

ohh, a teacher is enlighten me.
another thanks !

pesonally i don´t care where my CPU is going. The only relevant thing is: it IS going !

What? FFT is hardly cutting edge, I was using FFT based effects in Cool Edit last Century on a Pentium 90.

That´s for realtime play purposes. Not exactly ideal

So use something else, FFS!

Great !
bring me into this century: some links would be appreciated for that dumb ass here who could not find the light on his own.

until then: do i use what i found.
AND: i am in fact happy with it


Waiting now for some links, so i can step up my game........

[IvyBirds]

But it's not actually doing anything, it's just serving up instruments and effect, which is where your CPU power is needed.

A single core can only process a limited amount of data at a time at a given speed

That's why chip manufacturers make multi core processors so that that each core can process data on parallel with the others that are also processing data

Modern computers can process a lot of data in a single core, potentially multiple synths at the same time, however there are some plugins that use a lot of CPU power especially when you have a decent amount of polyphony going along with sustain

So when you put those instruments into a dedicated core they get all the CPU power they need without interfering or competing with other plugins. That ensures everything is stable and there are no crackles or other artifacts that happen when a CPU gets overwhelmed

Maybe that's not an issue for you based on the instruments you use and the amount of polyphony you use at a time, but it can be for others

So many users find it helpful to manage the cores you are using so you have complete control over everything

[frag]

Using GP as the main host:
Would I load each VST plugin into a separate instance of the Audio Grabber plugin... or would I load all VST plugins into a single instance of Audio Grabber?

For each instrument load a separate instance of Audio Gridder "client" which will then look for a "server" which can be local to your machine or across the network

Then load the instrument into the Audio Gridder Client which will load it off the server

You can select if you want the GUI as well or just audio and MIDI

so if you want 5 different instruments/effects you would need 5 instances of the client of Audio Gridder which loads as a standard VST3, AU, or AAX for Protools in whatever host/DAW you wish

You can even load an instance in Logic or Protools and run the VST version in the server either locally or on a different machine with a different OS

Thanks for the explanation!
Downloading as we speak.

I can't describe in words how much I love AudioGridder.

Thanks to this plugin, I can try and use some of the latest VST3 plugins on my older desktop.

It runs in virtual machine very well.

[Jim Roseberry]

I've measured my Nord's latency (MIDI note to when the sound is actually generated).
It's 2ms

That's much lower than most hardware synths. Scot Solida did a big test of a lot of the synths he owned at the time and found that some of them have latency up to almost 20ms. My bandmate plays drum pads that trigger drum samples on the PC and anything under about 20ms is perfectly fine for live work. Until you tell him what it is, then he'll swear he can feel it. So I just don't tell him.

Yes, the Nord's response time is fast.
I've measured other keyboards that are significantly slower.
ie: The 3rd Wave is about 7ms.

I get that not everybody is ultra concerned with latency.
I've been chasing ultra low latency performance my entire career (most of my adult life).
I like tight/responsive timing.
ie: If I'm playing "The Grandeur" (nice piano library included in Komplete), it feels a LOT more immediate playing it at a 32-sample ASIO buffer size... vs 64, 128, or 256.

Again, I'll guarantee that you only feel a difference because you know there is one. In 2ms, sound travels less than 10 feet so if it was detectable, you'd hear a time difference between two instrumentalists playing on the same stage. But you don't.

I'm aware that sound travels at 1.13 feet/millisecond.
When I play The Grandeur at a 32-sample ASIO buffer size, it's far more immediate than when playing it at 256-samples. It's not psychological, it's absolutely perceptible.
We're just talking about ASIO latency; The Grandeur has additional latency beyond the ASIO buffer size (scripting and or convolution processing).
So it's not really playing back at 1ms latency.
My guess is that it's probably more like 3-5ms.

When we're playing live, you've also got to factor in other latency sources.
Today, virtual all mixing is done digitally. The best consoles have ~1ms round-trip.
My wireless IEMs have some latency.
If I'm using a wedge monitor, it has onboard DSP (which adds some latency).
The mains have onboard DSP... so they have some latency.
With the Nord setup (2ms response), by the time the sound is monitored, there's probably 3+ms additional latency.

Years back (earlier in DAW development), I was as a tech on some Nashville recording sessions. We tried a few takes monitoring thru the DAW application... but the singers quickly put an end to that. They could perceive the latency... and it was messing with their performance. Today's computer/audio-interface performance wasn't available back then... so we had to monitor straight off the console (avoiding DAW related round-trip latency).

We've got a loud drummer... and sometimes the sound engineer will bring out the plexiglass "wall of shame". It takes the sound a few milliseconds to get around that barrier... and it has (at least for me) thrown off timing of transitions.

I've gone back to using IEMs (trying to save what hearing I've got left).
Prior to that, when working a large stage, there were moments when timing felt uncomfortable (because the distance). Of course, you've also got FOH filling some of that void... which helps.

If I'm going to take out a software based rig (and the time/effort to make it happen), I want it to eclipse my hardware rig... in every way.

Why, when the other advantages far outweigh any tiny difference in latency? Any time you spend setting up your software rig will be repaid after just a handful of gigs in how much less effort it is to transport and set up. In fact, if I compare the time I've spent getting my software working over the last 25 years to the time I used to spend doing the same with my hardware, I reckon it's probably only one-tenth of the time. Hardware always required so much more effort than software does. You had to find stands, or build your own, do all the cable work to make it run (audio, MIDI and power cables, with junction boxes, etc.), set up a mixer and effects racks/pedals. It took forever compared to plugging a MIDI controller into a computer and connecting that to an audio I/O device.

To play the songs we're doing, I need a two tier keyboard rig.
I was making it work with a single keyboard, but with only a split-second to switch sounds... if anything goes wrong on any of those changes... the part is borked.
Even with the best of software rigs, I would still likely take out the pair of Nords (as controllers). They're actually on the smaller/lighter side. The 88 is about 42-lbs and the Compact is 23-lbs

Regarding a Keytar:
I've owned a couple.
It would be hard to play two-hard parts for songs like: Carry On Wayward Son, Don't Stop Believing, Sister Christian, etc.

You don't have to use it for every song. I built a small pedestal stand so I could use mine with both hands when I needed to. It's just like your guitarist changing guitars between songs.

Good idea with the keytar and stand. I've thought about adding one... but just haven't done it. Several local guys are using them. That's partly why I've avoided it.

[Funky40]

Thanks to this plugin, I can try and use some of the latest VST3 plugins on my older desktop.

thats a good idea !

downloaded it !


one short question here:
it only runs on another computer,
or could i run audiogridder also within same Mac from app to app ?


wait, another: it can host whole chains of plugins, not just one single one, from what i grasped on the quick ?

[frag]

Thanks to this plugin, I can try and use some of the latest VST3 plugins on my older desktop.

thats a good idea !

downloaded it !


one short question here:
it only runs on another computer,
or could i run audiogridder also within same Mac from app to app ?


wait, another: it can host whole chains of plugins, not just one single one, from what i grasped on the quick ?

You can run it on same computer.
For example, on Win7 PC, you can install virtual Win11 and AG server.
It's a bit complicated and there are issues, but for testing synths it's fine. If I really needed to do a track with 15 year old sequencer, 32bit VectorSector (2004) and Jup-8000 V VST3 (2025) running in the same project, yes, I could do it...
But for serious work I would just get Ultra 9 285K and use only a few best plugins.

[Funky40]

But for serious work I would just get Ultra 9 285K and use only a few best plugins.

i ´mac !
for what i´m doing and how i´m working is a win machine just nothing.
thanks