Should I record my KORG Poly Ensemble for making samples?

Discussions about anything related to samplers and sampling techniques.
Post Reply
paulhimself
Newbie
Newbie
Posts: 3
Joined: Tue Feb 12, 2013 10:08 pm

Should I record my KORG Poly Ensemble for making samples?

Post by paulhimself » Mon Feb 18, 2013 4:22 pm

I have a 'feeler' thread about my synth in Buyers Guide here but this is a different topic.

At the moment I'm playing around with samplers, namely Highlife and Shortcircuit - the possibilites are quite astounding. It occurred to me I could record my PE-1000 before I sell it, if I ever do. I'm not a synth expert but I would imagine that one used by Jarre and Vangelis must have something going for it.

If anyone is interested I'd need to know what and how to record, being a sampling noob. From my studies so far on the web I've got the impression that recorded note length should be between five and eight seconds with a silence between each note. Key mapping allows one to use two or three notes per octave, but what the heck I might as well record every note.

Should I record 60 notes for each instrument? Should all fx be off? Would 44100khz be okay or 48000 khz? Either way the bit rate would have to be 24. What about headroom? Four or five dB below zero? Whatever you want, I can do it.

Once the recordings were done I'd post them up in a permanent box.com folder and anyone interested could take then down to store for posterity and maybe get down to making the samples. I don't yet have enough know-how to provide pro-quality finished samples, just the source audio files.

Let me know, then, and I can start recording. There don't seem to be any Poly Ensemble note recordings on the web, so it might be prudent to get some done whilst my synth is working well.

User avatar
nogginj
Junior Member
Junior Member
Posts: 181
Joined: Mon Mar 19, 2007 9:30 am
Location: tejas

Re: Should I record my KORG Poly Ensemble for making samples

Post by nogginj » Mon Feb 18, 2013 7:34 pm

You only need to record a note for as long as the tone 'loops'.

The best thing you could do would be to record the raw oscillators with no filters, no envelopes - just record a few octaves (or maybe even a C then a G for 2 notes / octave). You then take the tiniest portion of this that you can and loop it. Ideally a 'single cycle', but in my experience it helps to have a couple.

So with that thinking, 5-6 seconds is way over kill.

However, if you are recording fully sculpted 'tones', you need to record it as much as you can to where you can find a loop.

The point is, remove things like volume envelopes that can easily be recreated in another sampler. Record the sound of the synth as 'raw' as possible, then recreate the tone via another sampler.

It takes a lot of work but provides the most flexibility later.

User avatar
ppg_wavecomputer
Expert Member
Expert Member
Posts: 1010
Joined: Wed Feb 06, 2013 3:37 pm
Gear: more keys than hands (and feet)
Band: ['ramp]
Location: teutoburg forest, eastern westphalia, germany
Contact:

Re: Should I record my KORG Poly Ensemble for making samples

Post by ppg_wavecomputer » Fri Sep 13, 2013 4:20 pm

I did just that. If at all, the 1000 sounds even worse sampled than it does anyway.

Stephen
"Like the light from distant stars, Stephen Parsick's music has existed for some time, but is only now reaching us on Earth." Chuck van Zyl

https://doombientmusic.bandcamp.com/

https://ramp1.bandcamp.com/

https://stephenparsick.bandcamp.com/

User avatar
CS_TBL
Synth Explorer
Synth Explorer
Posts: 1677
Joined: Wed Jul 12, 2006 3:47 pm
Gear: All "In-The-Box"
Mainly FM8
Location: NL
Contact:

Re: Should I record my KORG Poly Ensemble for making samples

Post by CS_TBL » Sat Sep 14, 2013 11:23 pm

I've sampled tens o' synths/keyboards in the late 90's. As mentioned above, I did whatever I could to get the raw ROM, by bypassing any filter, envelope, scaling, LFO, DSP etc. Then I tested where the key groups were and sampled the lowest note of it. Sampling the highest note of the group would implicate less quality, so even though memory wasn't sky high in those days I picked the best and biggest option. I also managed, quite often, to exactly spot the looping area of the original samples. In that case you can just set these loops without needing to crossfade the loop area!

For synths that aren't based on samples (e.g. oscillators, physical modelling, additive etc.) I'd rather just sample every key, you won't have to be annoyed with all the artefacts you get with shifting the root note of a sample. If the tones of the Korg have their own hardwired flow (envelope, LFO etc., note that I don't know the ins and outs of this Korg model!) then at least they'd be captured accurately.

As for looping and sample size.. as you obviously want to use 'm in-the-box: what do you care anyway? You can capture 10 seconds of 'm, that should be like 6 bars at 128 BPM, and be done with it. After all: streaming means 'who cares about size'. Besides - and again I don't know the Korg P.E. - I can imagine there'd be some oscillator instability, by quickly looping the smallest cycle you'd loose all that. Even if at some point you still want small 'n looped versions (perhaps to fit 'm into a small rack sampler), then at least you still have the original recordings!

By the way: looping two cycles isn't the best idea. If ever there's a slight deviation in one of the cycles, this deviation will cause a sub octave to sound. Getting a correct loop like that is especially annoying at high frequencies, as cycles may have a length of a fraction. As having multiple cycles in a loop isn't a good idea, this is exactly why you best sample the whole 10 seconds out of it and just let it fade out. Or 20.. or 30.. what do you care? :)
"You know I love you, CS, but this is bullshit." (Automatic Gainsay)
s: VSL/FM8/EWQL/LASS h: DX7/FS1r/VL70/SY77/SN2r/JD800/JD990/XV88/Emu6400/Poly61/Amek35:12:2/genelec1030 r: Violin/AltoSax/TinWhistle c: i7-4770/RAM32GB/SSD
FM8 vids

Post Reply