Sound Design Feedback Thread, Dec. 2023

[nathan]

@Lukiss

I would try EQ out some frequencies that goes slightly out control on reverb, (the wet side) for the percussive sounds. Something with the reverb is going out of control. If me and possible, I would test to move the metallic feel/quality on the reverb from the percussive elements and add the metal to the vocal elements. Then i would would test to dial back the reverb on the drums & percussive elements just about until i hardly notice them.

This is useful feedback, thanks. There’s some subtle room reverb on master but also some very obvious metallic reverb that gates on very occasionally, which was intentionally incorporated into a Looperator-style thing that randomizes effects on every beat. The room reverb is fine, but the latter is just not hitting how I want it, so I’m going to try alternative ways to create space. My first instinct is to put some very, very subtle noise sweeps in there, with some inherent 'verb that is contained entirely to them.

Revised version with less reverb, tonal elements less annoying, some additional mixing:

Reminder to others here, the reason I made this is that sound design isn’t a spectator sport or a theoretical pursuit — let’s keep discussion focused on workshopping specific sounds we have posted here.

@Lukiss

Okay… my turn to be roasted (

I don’t know where to go with this…

This is based on a sample, right? It’s a nice texture, but to me feels like it’s starting over and over, an intro that doesn’t lead to a drop. Keeping this as an intro, here is my drop idea: add an original bassline, a different drum loop, and aggressively filter the sample to make room for the original material. Maybe repeat a single beat of the sample when that happens, and add some more negative space to contrast with the constancy of the sample. Call and response with another element could work too.

[dietcv]

I really like this reworked sparse version. Im personally not a fan of the robotic voice. The harmonic material is cool. I think the “pling” sound at 0:14 could be an octave higher and maybe hit a different note. The drums are solid as always.

[dietcv]

@nathan

If this is the only or primary element in the mix, maybe add in another modulating oscillator that’s more in the 20-40 Hz range, with really really dramatic exponential FM across many octaves?

cool idea, i will try that. In generral i have figured out this needs different modulation sources to transform its shape over time.

With reverb i tend to use just some early reflections to colour the sound but without any bigger tail. Otherwise all the nice transients are washed away. I personally think the reverb in the first reference could be less.

I havent finally settled on the accompanying material and have already tried out different stuff. There is some room in the higher registers and for base frequencies. But in general i dont want to use drums.

@scztt

You might try creating banks of each of these squeakers, where you slightly detune one or more parameters for each iteration - then you can either sum them, or progressively subtract them from each other (subtracting alternating channels of a multichannel expansion is simple - something like (multiSig * [1, -1]).sum.

I will try that thanks

[nathan]

You’re right, the voice synthesizer needs work. I think it should be more synthetic, or I’ll drop it entirely. I had a hard time getting it to fit in the mix in the first place, so maybe my gut was telling me something there.

[dietcv]

i think the other material is aesthetically more open and this is more “closed” with some retro reference at least how i hear it. if this makes any sense.

really instructive discussion on the Mastodon thread you have shared!

[smoge]

This second version is better, the reverb gate is dominating the beast. It’s beautiful.

This discussion made me think about situations where reverb can be REALLY interesting, for example a reverb with absurd parameters (infinite decay, or something very artificial), but which is controlled by very judicious gates (both input and output)

[fmiramar]

Some people say that reverb and delay are like salt and pepper: they tend to enhance rather than characterize, it goes well with literally any kind of food, including sweets, but the proportions can spoil your dish (or plate haha)

[jamshark70]

In defense of cavernous reverb…

hjh

[shiihs]

I guess the effect of reverb also depends on the space in which a piece is played. In headphones, a big reverb can sound great, but in a large reverberant place (say: a cathedral) it might interfere with the natural reverb to create a muddy result.

[jamshark70]

@shiihs

I guess the effect of reverb also depends on the space in which a piece is played.

Yes, and on the style of the source material. In that clip, I was looking for material that would make sense with an exaggerated reverb, and since you mention a cathedral, it would make sense in that space too. The artificial reverb + cathedral reverb probably wouldn’t destroy it too much either (though if I were producing it for a large space, I would cut back the digital reverb).

hjh

[dietcv]

@nathan

If this is the only or primary element in the mix, maybe add in another modulating oscillator that’s more in the 20-40 Hz range, with really really dramatic exponential FM across many octaves?

dont roast me for this one, i just made a quick jam with the additional exponential FM. It definetely can give you the “buffer scratching” vibes from the second reference without using any buffered audio.

[nathan]

I particularly dig the sudden downward sweeps near the beginning, an almost kick-drum like sound. I’d suggest you really lean into those as this develops.

[smoge]

@nathan

I think it should be more synthetic, or I’ll drop it entirely.

I much prefer real voices

[dietcv]

i have created these coloured noise impulses by using PV_MagDiv and the single side band PM from this thread and nathans Hasher approach for creating deterministic noise. But im currently a bit limited in getting different timbres out of this beside changing the modulational index (with MouseX here). I have tried to put it into a Non Linear Filter for more variation but thats not adding alot, does somebody has other ideas for more timbral variation?

(
var randomPhase = { |triggerRate, randomness|
	var minDuration = (2 ** randomness) / triggerRate;
	var maxDuration = (2 ** (-1 * randomness)) / triggerRate;
	var demand = Dunique(minDuration * ((maxDuration / minDuration) ** Dwhite(0, 1)));
	var duration = Duty.ar(demand, DC.ar(0), demand);
	var phase = Phasor.ar(DC.ar(0), SampleDur.ir / duration, 0, 1);
	(phase: phase, trigger: (HPZ1.ar(phase) < 0) + Impulse.ar(0));
};

var channelMask = { |trig, numChannels, channelMask, centerMask|
	var rate = if(trig.rate == \audio, \ar, \kr);
	var panChannels = Array.series(numChannels, -1 / numChannels, 2 / numChannels).wrap(-1.0, 1.0);
	var panPositions = panChannels.collect { |pos| Dser([pos], channelMask) };
	panPositions = panPositions ++ Dser([0], centerMask);
	Demand.perform(rate, trig, 0, Dseq(panPositions, inf));
};

var raisedCos = { |phase, index|
	var cosine = cos(phase * 2pi);
	exp(index.abs * (cosine - 1));
};

var singleSideBandPM = { |carrPhase, modPhase, index|
	var raisedCosWindow = raisedCos.(modPhase, index);
	var mod = sin(modPhase * 2pi);
	var carr = sin(carrPhase * 2pi + (mod * index));
	carr * raisedCosWindow;
};

SynthDef(\pv_magDiv, {

	var numChannels = 5;

	var tFreq, trig, phase, randomTrigPhase, inA, inB, chanMask, chainA, chainB, chain, sig;
	var modPhase, randomModTrigPhase;

	tFreq = \tFreq.kr(100);

	randomTrigPhase = randomPhase.(tFreq, \randomnessA.kr(1));
	randomModTrigPhase = randomPhase.(tFreq * MouseY.kr(1, 10), \randomnessB.kr(1));

	trig = randomTrigPhase.trigger;
	phase = randomTrigPhase.phase;

	modPhase = randomModTrigPhase.phase;

	inA = Hasher.ar(phase);
	inA = inA * IEnvGen.ar(Env([0, 1, 0], [0.001, 0.01], [4.0, -4.0]), phase);

	inB = singleSideBandPM.(phase, modPhase, \modIndex.kr(10));

	chanMask = channelMask.(trig, numChannels - 1, \channelMask.kr(1), \centerMask.kr(1));

	chainA = FFT(LocalBuf(2048), inA, hop: 1 / 16);
	chainB = FFT(LocalBuf(2048), inB, hop: 1 / 16);
	chain = PV_MagDiv(chainA, chainB);
	sig = IFFT(chain);

	sig = (sig * \boost.kr(3.dbamp)).tanh;

	sig = PanAz.ar(2, sig, chanMask * \panMax.kr(0.4));

	sig = sig * Env.asr(0.001, 1, 0.001).ar(Done.freeSelf, \gate.kr(1));

	sig = sig * \amp.kr(-15.dbamp);

	sig = LeakDC.ar(sig);
	OffsetOut.ar(\out.kr(0), sig);
}).add;
)

(
Routine({

	loop{

		s.bind {

			~pvMagDiv = Synth(\pv_magDiv, [

				// custom phase
				\tFreq, 24,
				\randomness, 1,

				// amp & outs
				\amp, -15.dbamp,
				\out, 0,

			]);

		};

		rrand(2, 5).wait;

		~pvMagDiv.set(\gate, 0);

		rrand(2, 4).wait;

	};

}).play;
)

EDIT: these releases are a nice inspiration for this kind of sound:

Yasunao Tone - Solo for Wounded • Part I

[smoge]

This is not a direct technical suggestion, but I think Richard Scott touches exactly the core of the issue in this clip. And he is an absolute master at doing this. Whatever electronics sound, we can never abstract the real world, don’t get into this trap. I think you could find techniques that consider designing the sound with an acoustic resonance, with “gravity” (as he said in the clip), etc… as if it hitting a material (even nonexistent), or it is influenced by the “wind”, he put well in few words

Richard Scott - Solo Modular Synthesiser Concerts promo

[dietcv]

@smoge

Whatever electronics sound, we can never abstract the real world, don’t get into this trap

what do you exactly mean by that in relation to the snippet / code / question i have shared?

[smoge]

I wrote it above. I gave two examples. For example, give these short sounds some “familiarity”, for example some kind of resonance, or gestural rhythm.

[dietcv]

okay, and your suggestion to the snippet i have shared is what?

[smoge]

Your snippet has a lot of things, but the resulting sound is lacking those aspects, that is exactly what you’re missing. It can be a a million different things.

I mean, it’s not pure mathematics, we’re actually talking aesthetics here…

[dietcv]

and how would you change it, that its not lacking those aspects?