The age-old question about the dark art of Resampling
 

Hi Infekted, first-time poster here. I've had my Virus TI for a couple of years now and I've been slowly ramping up my 'production game' as they say. One thing that continually frustrates me, however, is this 'resampling' business. My goal is to be producing dark, aggressive D&B and breaks, and I seem to take a crack at this technique every few months, only to become more frustrated.

I understand the technical processes behind it. I've watched hours of tutorial videos, browsed countless threads and I am extremely detail-oriented. I've gotten some pretty great sounds directly out of the Virus - from straight detuned reeses, to FM throbs, etc., but they tend to lack cohesion. Reconciling the technical with the creative in terms of when and what exactly to load into a sampler, is a part that confuses me. I can load a simple detuned reese into Live's Sampler and play with its shaper and morph filter, and play a boring timestretched melody, but creating interest and maintaining a cohesive sound seem to fall apart very quickly. Add to it the fact that Live and the Virus tend not to get along and it gets very frustrating very quickly. Any tips to this end? Thanks.

Interested in this subject too. How can resampling change the sound in the intended way? From the technical it is intended NOT to cause any sound distorsion while resampling. What can be said out of the box:

Any kind of resampling is a mathematical multiplication with a square wave and as with all mixing techniques like this will cause new frequencies. A square wave has harmonics itself already so all of them will produce mixing results like Freq A * Freq B = Freq (A-B) + Freq (A+B). To make this become sound sensible: The modulation frequency will have to be selected appropriately. I would consider to perform it that way, that A-B as well as A-B makes sense to the music. One could choose it that way, that e.g. 3/2 of the freq is used and the half and the double freq occur:

440 Hz * 660Hz -> 220Hz + 880 Hz

This will sound very harmonic, I guess. Not sure if it is what you want.

Resampling is nothing more then recording an instruments and placing the results on a sampler so as to be able to play the samples.

So what you really need to focus on, I guess, is in how you should go about making a sampler instrument. Then there's some stuff to look out for, such as:

your audio editor (or daw) must have a snap to "zero crossing" points - so as to avoid audio clicks. you can sample all the notes you want (a couple of octaves or more) and then assign them to the respective keys, if you're planning to use this as you would an instrument (play it, in other words). or you can place one sample and rely on the repitching algorithm within the sampler to do it. sampler instruments can of course be mono or polyphonic or legato - and you can use portamento to.

from the moment you record something, all further processing to the source will have to be either the sampler's own modules, like filters, envelopes and lfos; or further post processing.

think that's about it. but before that, watch your gain staging: you shouldn't record to loud or to low. to loud because it can create distortion if the signal clips and it's a good idea to have some headroom left, not to low 'cause the lower you go, the closer you are to the noise (search signal to noise ratio), hence you'll be bringing it up along with the signal with further processing (mastering included).

I think a big part of your question has more to do with the arrangement and composition then with sampling routines. As far as that goes: watch your timming, try and use scales so that things sound more cohesive and harmonic, questions and answers are always a good way to go as far as placing sounds together goes. Then you should think about distributing sounds across the spectrum and panning (and depth) - this starts with the octaves you choose to play... Time stretch changes the timming, adds some kind of repitch algorithm to the sounds that changes its harmonics entirely, along with its tunning (gets out of tune), so this can also be your problem. There's repitching plug-ins (and nowadays you get that inside some daws to) like melodyne or auto tune (by antares) that you can use to correct the pitch even after some post processing that changes pitch (like time stretch).

I'm using the Virus C with Ableton (and Logic), absolutely no problems at all. The mistery island editor also works for the TI range. it doesn't work with the usb connection, just good old midi and then the analogue inputs of your sound card. don't think the other editors available support the virus ti.

hope it helps. cheers

Thanks, TweakHead, for the detailed response!

The latter half of your post was most helpful, but I get the technical side for the most part - having worked in Ableton for years, and having a background in recording. I know how to record/cut/adjust, and how to employ a sampler. I'm poor with conveying ideas and putting things into words, but I'm mostly having trouble with the creative side of the 'resampling' question. It's such an open-ended process that I've been having trouble nailing down an effective workflow. For example, I don't know whether to write a whole bassline - modulation included - and simply record it and add more modulation after-the-fact, or to take that bassline note-for-note into a sampler, to play with timestretch, granulation, etc. etc., and re-create an entirely new sound. I'm finding that once I involve a sampler into things, I'm losing some element of creativity.

I like Evoke's Resampling tutorial kind of put things into a workflow that I could utilize.

https://www.youtube.com/watch?v=QEjM3bRmmE4
https://www.youtube.com/watch?v=c_FR0JOr03U
https://www.youtube.com/watch?v=dwUK9MrsGq8

I resample my virus b to my mpc to make dark twisty neuro basslines. There are a few different things to try, of course, here are a few suggestions.

Develop a patch thats one dirty detuned reese, and pay attention to the modulation routings. Record a really long single note into the sampler where you tweak and twist those knobs like a madman. Alter lfo rates and depths (especially to cutoff but not just that) until you get a good long sample filled with all sorts of everything (still just one note though).

Now, set your sampler to play a note as long as you hold a key, and then for every key or pad you have, set different start points in the big sample. So now when you mash away on your controller or pads you are hopping around in the sample. You can quickly try out lots of ideas this way and some kind of groove or pattern should click pretty quick. Don't forget you can use the pitch wheel or pitch individual bits of the sample. You'll probably want to add touches of eq and compression and adjust levels to even it out but this is the basics of how I go about it anyway.

Alix Perez has some videos of him making bass sounds on the Virus TI and recording straight in the DAW then manipulating them - probably worth a listen if your in to that kinda dnb.

Thanks guys, helpful stuff! Evoke's tutorials and Innovine's post have really put me in the right direction.

Unfortunately, it doesn't look like Ableton's Sampler has an option to spread a sample across its keyrange without pitch-shifting it chromatically, with the keyfollow controlling sample start position - like your MPC's slice function :( It's a shame because it's got some really nice modulation envelopes and filters that I was getting some cool sounds with. In Evoke's tutorials, he chooses his start positions manually which I'll probably do anyway... But since I'm a gearslut, now I'd like to hear some opinions about sampler VSTs that have a similar sample playback option... Kontakt? Alchemy? What are your opinions?

Thanks again, everyone. I'm on the right track. :)

but of course you can have different samples for different keys on Ableton's sampler. it's just a matter of digging into it, before you start thinking you need something else. these packages that come with most daw software is actually pretty much all anyone's going to need, imo. and it's much more then people used to have access to, like a decade ago, feature wise.

so yeah, best advise here is to learn your tools, see where you can take them. then you'll presumably make much more informed choices on what comes next. but to really answer your question, then Kontakt would be the way to go. alchemy can play samples and do all sorts of things with them, but don't think of it as something to be used for that kind of sampling, more like granular odd stuff then good old school kind of sampling - know what I mean?

Unfortunately not... At least, not in the way I and Innovine described. That is, one sample spread across the keyrange WITHOUT pitchbend, such that playing different notes only alters the sample's start position (like slicing a REX file in ReCycle) It's really unfortunate because it's such a simple thing that's been a common feature of samplers for years, but Live's Sampler can't implement it. There are other ways to achieve a similar process - namely, Live's 'slice to new MIDI track' feature, but the 'morph' filters, FM and pitch envelopes I'm so fond of in Sampler are missing from these instances. Although, even just typing this out now, I'm thinking of a couple of workarounds that sacrifice minimal features.

I think I should still invest in Kontakt, though, at least eventually. I know exactly what you mean about bogging oneself down with too much gear and software. I've dramaticallg shifted away from that attitude recently, restricted myself to just the Virus, Battery 3 and Live as my primary toolkit and my work has improved, I think. It's something I have to keep reminding myself; Keep it Simple, Stupid.

I'm going to repeat myself here:

you can assign DIFFERENT samples to different keys on Ableton's sampler. That's the same as:

"one sample spread across the keyrange WITHOUT pitchbend"

You can slice the sample from transients or what have you, which translates into:

"playing different notes only alters the sample's start position"

I know Reason and how it works. That's how I started: I've seen it grow. Later moved on to Logic and then Ableton. In the former two, both with ESX24 and Ableton's Sampler you can do what you're saying - that I understood correctly the first time, by the way.

You can have it both: spread a single sample across a set key range with a pitch algorithm (that you don't want, at least for the time being); and you can indeed place different samples (or slices of the same sample, there's no technical difference, btw) to different keys - like you would with Dr Rex, Recycle... So the problem here, is that you haven't took the time to fully explore the tools you already have and started wondering if it wasn't their limitation, but it's not, it's really just your own. This is a very tired argument that comes about everywhere on the internet. The answer is: Ableton is capable of that and much more, how else do you think it would suit professionals? Don't say it doesn't do something, just say you can't do it and ask how it's done.

Better still: google it, search youtube.

Agree on the restriction of gear and software. My post is something within those lines. The more time I spend making music, the more I notice how some things have always been there and are easily overlooked and how nearly everyone would profit immensely from simply exploring all the menus and options to really get to know the software they already have - along with other tools. Hope this doesn't come across the wrong way.

I know it can be done 'cause I've done it myself, btw. And I know many people, specially hip hop beat makers would kill if such feature wasn't there, for example.

Kontakt is a whole different story: it's worth it just for the libraries you get access to alone and the technologies implemented in there. It's also great for creating one's own instruments, but you'd be impressed to find out it's much more used as a rompler then a proper sampler - most people don't go under the hood with their stuff. So make sure you do! ;)

Well I appreciate your assertion, but the reason I brought up that limitation is that I'm trying to simplify my workflow, and unlike the MPC or ReCycle style of transient sampling, Sampler has no one-button, automatic way (that's all) to spread samples that way by splitting one huge file into a zone per key - not by transient, but 61 even-length samples - kind of like there's a one-button option that automatically slices a large sample by transient into a Drum Rack - excellent for syncopating loops but using instances of Simpler instead of Sampler and the slicing options are defined by Live's transient-detection algorithm (which I'm aware you can manually change) but in my experience isn't the greatest for sounds that aren't really transient.

I'm fully aware that there's a way to do this manually, in a handful of short steps - it's exactly what Evoke's tutorials show that were posted on the last page, that I mentioned in my last post ("even just typing this I'm thinking of several workarounds"). I don't mean to be antagonistic or at all to suggest that this is anything more than a 'PEBKAC' issue. BUT, the whole reason I started this thread was that I'm trying to find a relatively simple, quick workflow with as few technical steps to stifle creativity. Having such a feature 'on automatic'; a one-button click in a drop-down menu, would help immensely. But you're right, I am 100% comfortable doing it manually by duplicating zones (or entire Samplers) and manually choosing start position, like in the Evoke tutorials. It's not that much more work, but like I said, I started on the topic in search for the path of least resistance to satisfying sounds - isn't that what we're all after?

Again, to reiterate, I'm not trying to be deliberately obtuse with you, I greatly appreciate what you're trying to impart. I'm a bit verbose and poor at explaining, but I've just been looking for every little way to simplify the process so I can focus on the creative.

Thanks for your help!

Don't get hung up on the one-button-does-all solutions. Assigning a sample to a key might take one button, or a couple, but this is only a small, small part of the workflow. If you can do it within a few seconds per key thats good enough.

You really wanna look at how the editing works: ideally you want to hit the key and have a knob which moves the start position and some other knobs for the cutoff and the amp envelope, maybe even pitch. Then, when yow press a different key, you want those same knobs to edit the newly selected sample/slice. This, for me at least, is a really importart part of the workflow, much more so than assigning the sample..

You could take this to the extreme, and just assign your big source sample to one key, then hit that key over and over while tweaking the sample start and recording the results. Then just pick the best bits, and cut THAT up and continue.. the only thing you really need is patience and a discerning ear.

Yeah, it's a very small part of all of it. The more I approach the sampling with my hands on, the sillier I feel for having brought up such a mild annoyance. ;)

How you described it is basically exactly how I've been working: creating a long tone with lots of harmonic thickness and movement, sampling it and playing with short loop lengths and different sample positions, duplicating the instrument when I'm satisfied with a particular sound. It eats CPU pretty quickly, and I'm still working on getting used to tiny MIDI regions for each note - I'm more the type to work with larger regions for my eyes' sake but I find myself zooming in and out more than ever now.

I've gotten some particularly gnarly results by doing the whole process over again once I have a one-note 'melody' made of different throbs and donks, etc., by throwing it back through the Virus' filters and effects, and then loading the recorded output back into the same bank of Samplers I've created.

Like I said, I feel pretty silly about harping over the 'less button pushes = more efficiency' side of it. The speed comes pretty naturally and I'm beginning to really enjoy the process. I'll be posting some neurofunk soon, thanks everyone :)

But Ableton does have a way to split samples from transient markers and you can, indeed, assign each of those slices to different keys - and doesn't (even) take you long to get there. This is similar to what you'd get with a REX file.

I'm not going to go into the details here, as it's all covered (ad nauseum) in the manual and on many (!!!) online videos you can find easily.

Takes a while to get used to the features in a DAW, but it's certainly worth it if you want to get some serious work done in there ;)

I mentioned 'slice audio to MIDI' a few times times above - I'm well aware of it, I use it all the time to resyncopate breakbeats. Anyway I'm not really hung up on any more 'one-stop shop' solutions; I'm pretty happy with how it's coming a bit faster each time.

Actually one thing I've been thinking about but haven't tried to implement yet (I'm not at home right now) is using the Virus' sequencer (er, 'Arp>Matrix') in creating a crazy moving reese sound. Applying the sequencer to something like an LFO rate or Phasor frequency could probably yield some pretty nasty movement - like the LFO's sample/hold mode, but with more control. Thoughts?

watch out for the phaser on the lows, as it produces a big mess and shreds your mono compatibility in one go. absolutely fine if you're doing this on a mid rangy or higher layer, obviously - pretty common practice in DNB ;) chorus also works very nice for reese kind of sounds. that and dynamic distortion, choosing your saturation mix as destinations can deliver ;)

aside from that, that's a clever way of sequencing some evolving patterns, of course. Arp>Lfo rate is good, specially combined with, say Lfo>filter cutoff. doing that to wave select, coupled with some fm in wave mode can be interesting to.

Oh yeah, that phasor definitely is not friendly to mono coherence. I use Flux Stereo Tool a lot for its phase vectorscope, and it's really funny to take a look at the low end of something put through the Virus' phasor/chorus/comb filter... what should be a straight line looks like an amoeba.

You know, I've never thought to automate wave select in FM mode - that's an awesome idea. I've played around with the shape/index parameter to change a sine into one of the digital waveforms but it doesn't really do much. I'm gonna try the wave select modulation tonight, thanks!

^flux is great. they've got great tools, that's a pretty nice analyser. it's a good idea to keep things centred on the very lows for most EDM music.

but you still need to watch that correlation meter when doing wacky stuff, if it goes negative it means chunks of your audio will be gone once its collapsed to mono - like on some night club systems. some people like a steady (means no modulation depth) phaser, just to rotate it a bit - guess an all pass filter will work to the same effect - and make it gel a little better with other elements.

if you're talking about the classic oscillators, then full to the left you'd be on "wave mode" - where you can choose 64 waves, and the "wavesel" destination also affects this. At center position you have a perfect (kind of, watch with an oscilloscope...) sawtooth, then to the right it starts to morph into a pulse wave (perfect pulse all the way right); then from centre to left it starts to morph to a perfect sine wave, so it's kind like additive synthesis where you'd be taking of the harmonics gradually - this should presumably give you some ideas btw.

random>wavesel makes it change on each note press (note in), and can produce interesting patches; like an arp fm patch, where the timbre is rapidly changing, for instance. also interesting, since arp patterns have velocity information, is vel>wavesel.

Yeah, mono coherence/correlation is one of the parts of sound design and psychoacoustics that I find extremely fascinating. Back when I was in school I liked to make my drum kits sound miles wide... then one of my profs collapsed a project of mine to mono, held up a thumbs-down and made a farting sound with his mouth. That was embarrassing. I know better than that (mostly), these days, hahaha.

On the topic of the classic oscillators' Shape parameter, I'm curious about something: Since it serves to add harmonics (changing a sine or triangle into a saw and then to a pulse wave as it's increased), would it be correct to say that it can create a similar effect to a lowpass filter? For example, applying a sharp, plucky envelope, inverted, to the shape of a classic oscillator set to sawtooth, should have a similar effect to applying the same sharp envelope to a lowpass filter for a 'pluck' sound?

So the 'Random' modulation source changes on each key press? That's really good to know. Man, there are so many modulation sources on this synth that can be used in really, really awesome ways.

^well, it morphs to the wave that's selected. so if that's a sine wave, you'd have something similar to a low pass filter going from saw to sine, yes. you can create a perfect sine with a sawtooth if you filter it enough - check with oscilloscope. you do have more filters on the fx section, on distortion - I think the TI keeps it just so that it can reproduce older presets, but it can have its uses.

that experiment would result in a similar sound, but not exactly the same as having a pure sawtooth with a plucky envelope modulating a low pass - sort like the trance standard bass patch. 'cause then you'd have to take into account the characteristics of the filter, that's some steps beyond simply adding and removing harmonics, specially if there's some resonance on it.

yeah, random changes with each note - different from Sample&Hold LFO in that department, even though if you had all your notes snapped to a grid, you could presumably achieve the same effect with it if it was synced to that grid tempo...

the fun thing about the Virus is that this frame of mind tends to come and stay: it seems like some pretty regular subtractive synth at first, but then there's much more to it and it's inspiring to explore all the holes in it.