Sparkle from Apesoft gets the AU treatment.

dobbs

I really don't get what this app actually does and the app description makes it sound like a vocoder (and when I plwy around with it it also kind of sounds like a vocoder), which I've read in many of the forum entries here, it isn't. One forum poster described it vaguely and to me what he said sounded like the app takes small portions of both audio streams, fourier transforms them and then multiplies the resulting spectra with each other. but that can't be all because then there wouldn't be a difference between source and target? can somebody enlighten me?

(and while we're at it. the basic principle of a vocoder is: both signals are split into a few bands, then you take the volume/amplitude envelope of the modulator signal and apply it to the carrier. did I get that right?)

(so can this app maybe be understood as a vocoder with an "infinite" amount of bands?)

espiegel123

@dobbs said:
I really don't get what this app actually does and the app description makes it sound like a vocoder (and when I plwy around with it it also kind of sounds like a vocoder), which I've read in many of the forum entries here, it isn't. One forum poster described it vaguely and to me what he said sounded like the app takes small portions of both audio streams, fourier transforms them and then multiplies the resulting spectra with each other. but that can't be all because then there wouldn't be a difference between source and target? can somebody enlighten me?

(and while we're at it. the basic principle of a vocoder is: both signals are split into a few bands, then you take the volume/amplitude envelope of the modulator signal and apply it to the carrier. did I get that right?)

(so can this app maybe be understood as a vocoder with an "infinite" amount of bands?)

Basically. I could be wrong but I think Sparkle is more of a convolver than a vocoder. If you do a web search with vocoder and convolver or convolve, you’ll find some articles about the differences.

Telefunky

With certain settings Sparkle sounds quite „vocodish“, like Autotune does with it‘s pitch shifting artifacts, or so called phase vocoders that probably do some convolution-like (?) processing.
The classic filter bank vocoder is an entirely different thing indeed.

dobbs

so I've read a few articles and for example this one
https://www.izotope.com/en/learn/the-basics-of-convolution-in-audio-production.html
seems to agree that it is indeed the frequency spectra that are multiplied. so for this to work in a practical sense I guess you have to take tiny segments of the audio that must be long enough that you can actually get a frequency spectrum.

but reading the app text again, I gues it's still a bit more complicated than than

mojozart

From the user guide:

Sparkle is a tool for advanced spectral hybridizations made of several algorithms that operate on frequency domain. By means of envelope preservation and phase-sync processing it can creates high-quality sound transformations.

The basic idea is to transfer the temporal structure of a sound (source) onto another sound (target) of which the spectral properties are preserved.

Simply put, you can take a voice and make it sing a Beethoven’s symphony!

To have the magic, it will be sufficient to take a source sound that is very variable in time (such as a voice, a drum pattern or a melody played by an instrument) and a target that is very rich in spectrum (such as an orchestra playing, a synth pad or an environmental sound).

RanDoM_rRay

Has Sparkle function been replicated, or surpassed by Bleass Phase Mutant? Perhaps i don't fully understand them both yet. They both look like fun tools to experiment with

HotStrange

@RanDoM_rRay said:
Has Sparkle function been replicated, or surpassed by Bleass Phase Mutant? Perhaps i don't fully understand them both yet. They both look like fun tools to experiment with

Both are great but they’re not all that comparable, imo. They’re found for totally different sonic atmospheres. Sparkle is more like a spectral vocoder where you have to route the source to a target and Phase Mutant is just an FM based audio effect you can add onto any track. I get totally different results with them both.

ejacul337

Fellas, how is this?!

bygjohn

@ejacul337 said:
Fellas, how is this?!

Well, I can’t give an authoritative answer as I bought this a while back in a sale and never quite got round to doing anything much with it. But I’ve just had an enjoyable few minutes messing with it, thanks to you bumping this thread, and I’m thinking I should use it more. A lot more!

Hopefully someone can give you a more useful answer, but in the mean time, thank you! 🙏

ejacul337

I've been playing with it, it's seems more muddy than sparkly. I was expecting something like GliderVerb I think.

bygjohn

@ejacul337 said:
I've been playing with it, it's seems more muddy than sparkly. I was expecting something like GliderVerb I think.

I think it’s very dependent on what you feed into it. To my idiot brain it seems more like a resonator where the thing you’re exciting is a sample playback. Which probably says more about my ongoing resonator obsession than the app.

I’ve mainly been feeding in percussion from Noir and using a sample of me playing an Indian banjo (bulbul tarang), which has loads of harmonics etc. The outcome is very different using some other samples as targets, mostly I’ve been using the ones supplied with it.

ejacul337

I've been feeding straight vox to it bounced from iVoxel believe it or not- as you said, it's all what you feed into it. I don't know why I expected something like GliderVerb.