A quest to construct a bagpipe ensemble from one sound only

Question

First of all, I would like to point out:

1. I do not have any means to record real instruments.

2. I only work with midi, samples and plugins (which includes nice ones like spitfire).

3. I just have a tablet with a two octave midi app.

4. I can use other apps to emulate faders, in order to work with some volume dynamics.

5. I use the Reaper DAW (but you can suggest another one)

6. The whole context of my creations is cinematic/movie soundtrack music.

7. The question is written in section 3).

Also I'm trying to emulate the following bagpipe ensemble "sound widths" (specially in (B)):

(A)

(B)

9:31 min

1) Introduction

Well, using the bagpipe as an example (which serves to my purpose), a single bagpipe player playing a single note is, for a far sound detector, a source of plane sound waves; a harmonic oscillator. Many of these oscillators (bagpipe players) forms a ensemble. The tonic of my question is: "How can I create a ensemble out of a single instrument?", i.e., I have two plugins:

I

1. Ilya Efimov Uilleann Pipes
2. Nexus 3 - Scottish Champion

None of them can reproduce the atmosphere given in (A) and (B). The answer of my tonic question is simple though: "It won't sound like (A) and (B) because its a massive crowd and real sounds are being recorded; it is a real performance!". Nevertheless, I would like to try to emulate the whole thing.

Now, a class of programs that can emulate width, or at least help on ensemble creation (introducing some width), are plugins like (I listed the ones that I use):

II

1. Ozone
2. Halo upMix
3. StereoSavage

2) Trying to construct a bagpipe ensamble

Now, using the plugins in I and II, I reached out a mild and poor ensemble-like atmosphere; it is not even close to (A) and (B). Using the principle that every bagpipe player is a oscillator, my next attempt was, as it is shown in the pictures below, to create a little "ensemble" of plugins with different volume values (I was trying to emulate some distance). First, I recored the drones using a set of 8 bagpipe plugins with different volumes (figure 1)

Figure 1

Then, due to my computer specs, I rendered the drones and recorded the melody using the same setup + the drone .wav:

Figure 2

Unfortunatelly, the attempt in section 2) was frustrated. I mean, I rendered a nice sound but not a ensemble atmosphere. The usage of layers and volume dynamics seems to play almost no role here. Together with the bagpipes, (Nexus plugin), I've use also the StereoSavage plugin, like in figure 3.

Figure 3

Also, I not aware of good plugins that have bagpipe ensemble sounds; in spitfire standards for example.

3) My question

I would like to ask: how can I create a ensemble of bagpipes (or any instrument) that approximate the width and "power" of (A) and (B), using just midi plugins?

I've already tried reverbs too, without significant modifications.

Answer 1

You're actually kind of hitting the laws of chaos theory.
One thousand pipers in a large space is chaotic [my feeling about bagpipes aside ;)

Each bagpipe has four simultaneous notes, three drones and the chanter. None of these are ever exactly in tune. So your 1000 pipers are producing 4000 simultaneous sounds, at actually 4000 separate pitches.

They never play exactly in time. They play around an average of the correct tempo, but if we're down to the sub-millisecond, no two will be exactly in sync.

They are at different distances. No matter how you mic this, the sound from some is going to take a lot longer to arrive than others.

Some are so far from the mics that you will never really hear them - they may as well be reverb.
From a synthesis angle, this may actually be a bonus.

Right. So, how to approach this in synthesis when you have only two pipers… not only two pipers but actually two versions of each & every note [plus perhaps round robin instances, I don't know the details of each of those sample sets.]

First - randomise timing. This is probably simplest to to achieve by just playing the part over & over again, once for each of your theoretical pipers. No quantize, just play it.

Randomise pitch. Slip the overall tuning of each virtual piper 'a bit'. You will never achieve the totally random tuning of real pipers, but over several iterations, this might do as a substitute.
One trick might be to 'cheat' entirely & force some pitches out of their comfort zone. Assume your sample set has one sample per chanter note. Play a semitone flat & pitch-bend it up a semitone. Same note, different sample. That will throw in a bit of variation. Repeat for a tone & then in the opposite direction. Try adding a simple delay, quieter, so you fake another player behind every player you have to make a performance for. Delay mod could be useful here - swaying the timing & pitch of the delays.

Brighten some tracks, dull others.

You're never going to get a thousand pipers this way, but you might get enough. Eventually you're going to run up against that these are all the same set of bagpipe samples, whatever you try to do to them.
You are going to hit issues with phase-cancellation or flanging, because you are re-using the exact same samples over & over.

OK - now to sound staging.
Throw away anything that claims it can make 'stereo wider'. This is only going to make any phase issues even worse - & worse still, you may not notice in stereo, only if you sum to mono. Avoid this stuff entirely.
You need to make 'distance' not 'width'.

Reverbs may be your get out of jail free card.
Use something with variable pitch too & you might just gain another dozen distant pipers. You want short, low resonance reverb, heavy on early reflections, thin on late bloom - otherwise it will quickly start to sound like a warehouse, not outdoors. Outdoors doesn't really ever have reverb. It may have a distant slap-back off a tall building, but it doesn't bloom, it dies. By trying reverb on this you're in effect trying to emulate those far 500 pipers you can't ever actually really hear.

Honestly, I don't think you'll ever get to 'truly convincing' but it might be worth a shot.
Randomness is your key here. Simple copy/paste duplication is not.

Late thought - distance reduces top end. Make your 'far examples' duller than your near.