Pierre
Are you using an external audio interface?
Just for reference, I use Noire followed by IR1 Waves reverb. This configuration rise “Time Load” in Cantabile at 12% with up to 16% peaks.
Sorry Corky, we overlapped a bit 
Not a prob Paolo. I tend to do that a lot. 
I’m asking myself about nuc intel configuration…or the latest surface pro.
Right now I’m using a surface pro 3 with 8gb ram.
What can it handle at the same time?
Well, 2 instances of omnisphere with 8 patch for each + one kontakt library + blue 3 + a rack with gate and reverb for a microphone.
After that I use a presonnus sound card at 44khz and at 256 samples.
All run without any glitches. BUT I’m at a limit. I have to really organise the effect in omnisphere too ( aux common )
Saving CPU is the most important thing. And you need to save at any point by choosing your effets in omnisphere.
Today I tried an I7 Dell inspirion 8th gen (5 years later)… it seems to be a bit powerfull than my surface i7 3rd gen… but not that very much… my gain is about 20% cpu only for the same project… I’m a bit dissapointed. I was Waiting for more.
What are your expériences? Any better killer config to share? Any infl about the nuc intel i7 10th gen?
And at last, you really need to tweak windows 10 too!!! I USED THE GLITCH FREE MANUAL FROM BRAD. It works so much more after that !
Yes I use a presonus audiobox
Thank you. I read it all!
And … still no idea of what to do.
We should all Share a same project and make comparaisons…with our own gear.
We could Share a cantabile song with few free vsti… all pre configured and put in a dropbox. Make it turn with asio4all at 44kz and 256 samples. So we could test only the processors.
We need People with different pc…
Is it possible to do? Who would be interested in?
What do you think of that. We could in this way make better choices.
Sure I am. I need to know asio4all, at now I used only the asio driver supplied. It’s a good testing way, cause the real speed of a PC is affected not only by processor. Ram, chipset, thermal management, goodness of the driver, system tweaking… lots of variables.
What a intriguing test 
I’m too
I have my desktop computer for the moment and can compare with my future refurbished laptop !
Interesting idea - but I’d recommend using the “Null Audio” driver. We won’t be able to hear anything with it, but it takes the efficiency of the respective drivers out of the equation. Even with asio4all, you still have the driver underneath asio4all in the mix, which may be different from system to system. With Null Audio, that is taken care of.
The problem is that loads are very different: softsynths (Diva etc) use tons of CPU power, but little RAM or hard disk access. Sample-based instruments stress mostly the hard disk and RAM, with far less strain on the CPU.
What would be a “fair” mix of plugins to compare systems? If we just want to run CPUs against each other, then probably softsynths. So maybe just a couple of instances of a free soft synth (Tyrell?) holding a chord via “song->onLoad” bindings?
OK, I knocked together a quick one: simply download Tyrell from here, install it and download the following song file:
Perftest.zip (11.3 KB)
Before loading the song, set the audio engine in Cantabile to “Null Audio”, 256 samples @ 44.1 kHz , then load the song “Perftest”. Let it run for a while to settle performance and report back on time load. Also report your system spec (CPU etc.).
Let’s see the results come in!
Cheers,
Torsten
2 Likes
Just ran it on two of my main machines (figures from Cantabile profiler):
- Live Cube (Core i5 9600K @ 3.7 GHz) runs this at around 31.4 % average, 35% max)
- My studio PC (Core i7 7700K @4.2 GHz) is at 32.1% average, ca. 38-40% max)
Cheers,
Torsten
Torsten
Just to know, both your PCs are desktop, right? Frequencies are too high for a mobile processor. My desktop music PC is dead, so I’ll test only the laptop. I’m on my way…
Trouble: I see five Tyrrel all with unconnected midi in. In binding page i have 15 “missing target”. I connected all midi in to my keyboard, but I don’t know the trigger. A single note? A chord? I am sure there is something very important that I have not understood. @Torsten feel free to blame me, but please enlight this nightly keyboard player. It’s 3:20AM here, My brain is working 5%
Excellent!!!
I make the test! But the cantabile song you post won’t work for everyone… (different folders…) That’s not a problem. let install it the way we want.
I think we should all put 5 instances of Tyrell, prest 1, make a chord with 3 notes and hold it 10 second and see what’s going on. Note the result A
Play the same chord several for 10 seconds but hit it each second. Note it B
then stress it and play fast all the notes possible and note the result C.
We have to set the same audio parameters.
I put agressive mode beacause it really helps to see if the pc is strong or not.
Do you agree?
Very useful @Torsten
My results
Dell Inspiron 5559, Intel Core i7-6500U 2.5 Ghz runs at 50.0% average, 68% max
In general I am ok also for live performing, even with Keyscape and Omnipshere patches. But my setups are fairly simple, as I am in a 5-person band. But apparently I can improve my system almost a factor 2
I am planning to update my laptop system next year. Will get back to you and this forum for best specs to run Cantabile.
1 Like
Surprising my old pro 2 seems to be not that bad if I play normally…
Really put the same audio parameters:)
Have you actually installed Tyrell in the correct plugins directory? The bindings should point to the five Tyrell plugins (first three to Tyrell 1, next 3 to Tyrell 2, and so on). If you have “mising target”, then something may be going wrong loading Tyrell into the song. Can you try if you can manually set the targets of the bindings to the five instances of Tyrell?
What these bindings do is simple send an E-major chord (e-g#-b) in the middle octave to all five Tyrell instances. So you could connect your MIDI input to all five instances and play (and hold) this chord - should give the same result.
Anyone else with these issues?
