Best Gear to Record Samples on to be Played Polyphonically on Keys

[Yoozer]

Akai S5000 or S6000 or Z4 or Z8. You really want something with enough memory and .wav compatibility. The modular will bring the character to the table.

[Gringo Starr]

Originally Posted by Yoozer

Akai S5000 or S6000 or Z4 or Z8. You really want something with enough memory and .wav compatibility. The modular will bring the character to the table.

Thanks Yoozer. I can always count on you for good advice. It's always appreciated.

[Yoozer]

Of course it also depends on the patch - if there are lots of timing and formant related changes you want to sample as many notes as possible and for that you need generous memory. It is better to sample two octaves in detail if that is the range you want to use thant to sample 5 octaves coarsely - e.g. every 6th semitone.

[Gringo Starr]

Originally Posted by Yoozer

Of course it also depends on the patch - if there are lots of timing and formant related changes you want to sample as many notes as possible and for that you need generous memory. It is better to sample two octaves in detail if that is the range you want to use thant to sample 5 octaves coarsely - e.g. every 6th semitone.

I'm a total rookie when it comes to making samples. I was under the impression that I made one sample and the note automatically gets shifted to whatever key I'm pressing on the keyboard. So if I play a G B D chord it will be the same sample on all three pitches.

So if I make a sample of each note for lets say three octaves then does the sampler find the right sample for that exact note within the three octave range? I'm a little confused how this would work.

And about storage or memory... I would assume making samples into a computer program would be better. But as far as sound quality of the sample coming back out of my computer and then running it through hardware would the sound quality be less than if I got the Akais you mentioned?

[dalshim]

I'm a total rookie when it comes to making samples. I was under the impression that I made one sample and the note automatically gets shifted to whatever key I'm pressing on the keyboard.

Sure. However, the sound will not have the same lenght, except if you use something like a roland variphrase or kontakt. That's why multisampling exist: you sample as many notes as you can, then you map them on your keyboard.

So if I make a sample of each note for lets say three octaves then does the sampler find the right sample for that exact note within the three octave range?

That's multisampling. Newbie questions, mate.

You mentioned an asr 10. It's for advanced users, but those are great for sampling synth because of transwaves. It allow to re-use in real time your synth modulations. Also, one sample is enough most of the time.

[Gringo Starr]

Originally Posted by dalshim

you sample as many notes as you can, then you map them on your keyboard.

So basically to do this I would go down the line one by one and assign each sample to the proper key. I would imagine computer programs are easiest to do this with. So what about the sound quality coming out of a computer through a good D/A converter(Lynx Aurora) as opposed to the Akai, Roland Variphrase, or ASR 10 for example? Different ears might have differnt opinions but I still want to ask.

[dalshim]

I would imagine computer programs are easiest to do this with.

Not necessarely. Many HW samplers include a routine for this, like E-mu's, where most of the process is automated. Also, you still have the possibility to use a midi editor, or to build your program with a software and then dump it to your sampler. It really depend of which machine you would settle for.

So what about the sound quality coming out of a computer through a good D/A converter(Lynx Aurora) as opposed to the Akai, Roland Variphrase, or ASR 10 for example? Different ears might have differnt opinions but I still want to ask.

Also depend of the machine and what you're looking for. If it's colorless hi-fi, you're as good with your lynx than with a later akai or yam, yes.

[Yoozer]

Originally Posted by Gringo Starr

I'm a total rookie when it comes to making samples. I was under the impression that I made one sample and the note automatically gets shifted to whatever key I'm pressing on the keyboard. So if I play a G B D chord it will be the same sample on all three pitches.

It gets shifted automatically - but guess what happens if your sample has an LFO routed to pulsewidth; higher pitch is faster playback, so the LFO will move at a higher speed.

Also, if you sample 4 seconds, shifting it an octave up cuts the duration in half - so then you only get a note that's 2 seconds long. Another octave, and it's only 1 second long. Unless you loop the sample - and whether that works depends on what you're trying to sample - you're going to notice this effect in a rather obvious way.

Samplers work with a grid - left to right, note, bottom to top, velocity. Each sample can occupy a rectangle of arbitrary size on this grid; those are generally called "zones" (or keygroups). Such a zone is specified as ranging from say, C-3 to C-4 and from velocity 0-63.

If you want every key to be different, each "zone" is only one key wide - so, bottom would be C-3 and top would be C-3 as well. If you don't want to bother with velocity, then each zone spans the maximum height.

You could start recording a wave file and then simply play each note at a time, with a sufficiently generous amount of silence in between. You could then use something like ReCycle to cut it up in separate pieces. Otherwise, this cutting up of waveforms is supremely boring work suitable for interns. Alternatively, there's also SampleRobot - it'll play MIDI notes for you and start listening and cut automatically when the volume drops below a certain level.

With Kontakt, when you use a smart naming scheme like calling your samples something like modular_01_nc3_v127.wav) for (n)ote C3 and (v)elocity 127, you could write a script (still not for the fainthearted) to automagically pick the right sample when you press a certain key.

Either way, nothing wrong with just recording all the files on your computer first. You'll have to find out for yourself whether something sounds better when it goes through your Lynx A/D or the sampler's A/D. The general evolution of A/D (and D/A) seems to be aimed at neutrality, not character.

[Gringo Starr]

Originally Posted by Yoozer

The general evolution of A/D (and D/A) seems to be aimed at neutrality, not character.

Absolutely!!!! This is exactly what I want. I want it to sound as close to the actually modular sound as possible.

This sounds like it can be difficult to perfect. But I'm thinking that perfecting this and having a good supply of various oscillators and filters from various companies could end up creating amazing unique custom polyphony. I think it would be worth the many many hours to get this down right. I guess I'll have to research and see which way I should go. Thanks guys.

[Yoozer]

The most simple test would be as follows:

- take your DAW
- use any polyphonic synth or plugin to compose a quick chord progression, by using a MIDI track
- copy that track as many times as it contains simultaneous notes. So, if at most you play a 4-note chord, 4 times.
- delete all the excess MIDI notes on the other tracks. Example: a Cmin7, track 1 has a C, track 2 has a Eb, track 3 has a G, and track 4 has a Bb (and so on for every note and chord)
- mute all tracks but the first and route it to the modular instead of the plugin
- record the result as audio
- repeat the two steps above for each track

Voila, you've now ended up with the ideal result of what your sampling efforts would culminate in (each note captured fully and independently with zero sampling artifacts).

- does it sound good?
- if not, you've learned an important lesson; not all sounds are good in polyphonic form.
- if so, is it easy to sample (e.g. do you need to sample each separate note?)
- if so, is duplicating/filtering/multitracking like that less work than creating a set of samples? Might be! But - it's easy to recall a sample set once you've recorded it.

Do you care about efficiency? No? Then you might not even need a sampler. Yes? Then by all means build your own sample sets; more effort at first, but directly usable afterwards.