Hi!

I've been looking for a way to recreate this sort of shifting, syncopated delay sound in Bjork's track Desired Constellation, and I think I'm just not experienced enough to identify any obvious clues:

I've read that the main theme is constructed from a chopped vocal sample, but what I'm really interested in figuring out is how to create the same delay feeling she achieves here — it sounds like a pattern that is morphing it goes along, randomly dropping in and out of these little syncopated bits. Sort of like it's being pushed and pulled at will as the song moves along.

Anyone have some tips on what sort of thing I could look at/play around with to get a similar effect?

 

I used Echo Boy and Tantra, that's how it sounds:

https://instaud.io/2iRJ

links:

https://www.soundtoys.com/product/echoboy/
https://dmitrysches.com/products/tantra

instead of tantra, you can use any other bitcrusher

 

I don't think it's delay.

 

Last edited: Jun 11, 2018

https://www.whosampled.com/sample/158040/Björk-Desired-Constellation-Björk-Hidden-Place/

 

oooh echoboy sounds like a good place to start! thanks!

 

These two I would have to test, but probably the effect will be the same

http://d16.pl/tekturon
http://d16.pl/repeater

 

Aside from the quality of the design of the sound, it almost sounds like simple intervals of arpeggiation using 8th notes.

For the rhythmic variance of the arpeggiation, you could vary the arp pattern by just manually playing 16th notes in time, or even with some 'swing' (8th or 16th triplets).

Set the release to reset with the retriggering of new notes, run it through a slight decimator and ring modulator. You could even automate the time divisions of the arp after you lay your intervals and/chords down. e.g. 1/4, 1/8, 1/16t, whatever. I think the retriggering here is key to creating a modular, generative feel.

It's sounds like an alien 'static-radio' type of sound, doesn't it? I love it. I love Bjork and those that she chooses to work with.

 

Last edited: Jun 11, 2018

Yeah, an arp was on my mind too! This is a great suggestion, I am totally going to try this later tonight. "Generative" is a perfect way to describe what I'm going for here.

I think Recoil had the right idea with layering multiple instances overtop of of each other as well.

Thanks everyone!

 

Yes, it does seem like a voice down below would need to be addressed and programmed separately, if you took that approach, but it behaves like a mirror image of the upper voices.