Solved Sampling to save CPU

Hello, first, sorry for the english. I want to save cpu on large projects with maschine´s sampling function. I have a group with 8 patterns from reaktor. it sucks a lot of CPU, and i am sampling all the patterns. Next (and I don´t know if I´m doing it well), erase the reaktor plugin when I render the 8 patterns to audio. Then I paint a large midi note (C in my case) to cycle in each pattern. But when I change for example patter 1 to pattern 8 in the middle there is no sound.

I know that this is normal, that if I paint a note along the pattern and change in the middle of the cycle the audio stops ringing. And if I make the change at the end of the cycle there is a very small hole in which nothing sounds. I'm using the rounds synthesizer to do it. I would like to know if there is any way to do it better so that when changing in the middle of the cycle the note keeps sounding (and not from the beginning painting another note in the middle of the cycle). I hope to have explained myself well because I am taking my first steps in sampling to save cpu in large projects. Reaktor sucks too much cpu ....

I want to use it live and the gaps kills me. If there is no other solution, I suppose that I need to wait until the end of the pattern to change to another pattern.

Thanks

 

Knowing the minimum "Middle" time-measure is the key. Let's say for example that you always change patterns with a 1 bar interval, that means you need to chop your sample in 1 bar chops, so that at every bar, there's a middle C note that re-triggers the sample.
A more practical way is to chop the samples in keyboard mode and have a lot of slices so the sample bits are triggered more often.

Here is a practical example:

Group A1 has the heavy reactor stuff with 8 different 4 bar patterns.
I would re-sample each pattern to Group B1, each pattern is now a big wav in 8 different B1 pads.
Use the Sampling Slicer with a setting of 16 slices on each B1 Pad and then Apply it to the same Pad, (not the group, the pad) now Pad1 has a Sample in Keyboard-Mode that is triggered every quarter note, like so:

Now create a new Pattern in B1 and do the same thing for Pad2, repeat the process for all Pads.

Now you can change patterns as fast as 1/4 and your sample will always retrigger. Technically you can chop each sample with more than 16 slices, it might be worth it if your patterns are big.

 

OMG this is amazing!!!. Thanks a lot!!. Thanks for your time!! This is a great forum!!

 

If you use lemur I have built a "breakbeat generator" which effectively does this.

I use it with 10'000's of samples using Kontakts DFD.
Or just importing sounds directly into Maschine zones.

Uses Reaktor OSC in converted to midi on sequenced clock.
Be happy to send over files

 

Welcome! No problem.
I forgot to say that you might want to either save/export the original Reaktor Group or save as a new project before you delete the original group just in case you want to go back and change things.

 

Hey D-One,
i have done this a lot in the past to save CPU power and print my analog Synths, but yesterday I found out that I get little clicks/artifacts between the slices.
Have you tried to slice a long modulated Pad sound with out any artifact?
In my current setup (MK3, 2.6.10, 10.12.5, Apollo Interface) the slices are not playing without little artifacts between the notes.
Any idea?

 

If you just disable the plugin from slot, it no longer uses CPU.
No need to delete-export-save as. Keep in project but disabled is more effective IMHO.

 

yes, disable is perfect.
Can some confirm my problem with the artifacts between the slices?
F.e., use the SOUND/instruments/pads/"feather" and play one note along 2bars.
record it and put it in a new sound/sampler.
after slicing (grid mode "4th" and with pattern "on"), the sound don't sounds the same.

 

I hate to disagree but disable/bypass does not fully save resources in all situations.
I won't argue though I have tested and afaik it's pretty much proven especially for sample based instruments but to each his own. For more info read this:
https://www.native-instruments.com/...group-to-save-cpu-bypass-all-group-fx.320159/

 

Bypass from the instrument, not from channel lowers CPU.

I mean bypass like this:


Not like this:



Test and see

 

@Frank P. slicing works best when you cut from noticable trancients.
It doesnt work good with sounds like pads. But you can try ADSR & attack and release time to fix.

 

I know what you meant. Yes, it lowers it, but its still loaded into memory and taking resources. It "lowers" it more if it's removed.

 

thank you Alpert, so it is normal behavior that sliced sounds are not sounding exactly the same as the whole sample.
So this is not an optimal method to print CPU hungry VSTs.
Its a pitty that the slicing dont work sample accurate.
so I will search another method for printing VST patterns. with the upcoming audio loops this will certainly work.

 

You can slice it manually, or adjust certain slices. But there shouldn't be any clicks/pops if the sample is sliced at exact time divisions and played in the same BPM though.
Does that issue happen often?

 

Yes, I can hear this artifacts every time when I slice "Pads" or long sounding textures or chords.
I dont change the tempo and let the slices exact as they are generated from the slice tool (with patterns on)
I also thought that there should be no clicks. But they are there.
May be you can try it on your system with a pad sound (f.e. that one I mentioned above "feather".
I think you will hear it on your system too. When you slice beats this is not noticeable. But when I use this method to "print" VST Patterns with Pads it don't work fine.
My theorie is that Maschine do not work "sample accurate" witch means that the timing of the slices are no 100% in sync. I know this behavior from earlier versions of Pro Tools.
Tkank you D-One !

 

here are 2 WAVs. one with the complete sample and one with the sliced into 4th.

https://www.dropbox.com/sh/2dblb4ajjmnswd9/AACQ_WmQEcdzFgHiWwitqjICa?dl=0

you can try to slices the first file on your system to confirm if it makes the same problem.

 

WOW... That's extremely noticeable... ill test and investigate this further, this might deserve it's own thread.

 

Hi D-One,
have you allready tested it?

 

Why have you sliced a continuous note/chord/sound like that? Doing this will almost always introduce some artefacts, though changing the attack and decay on the zone tab of sampling mode for each slice could help a bit.

 

Hey Mr36, the File I sliced is an example to show that some sounds make artifacts.
You write that "Doing this will almost always introduce some artefacts"
but D-One wrote "But there shouldn't be any clicks/pops if the sample is sliced at exact time divisions and played in the same BPM though"
so what are we talking about ?
this thread is called "Sampling to save CPU"
and this kind of slicing sampled VST tracks would be a good way to save CPU
I think the reason for my question is clear
And, no, "normal" there should be no artifact when you play slices "sample-accurate"