In the last couple of weeks on my MA course, we’ve been learning about editing audio using phase vocoder based technologies, and granular synthesis. This is frankly, all a bit technical and mind bending for me, but I think i’m starting to get my head around it and I’ve been playing around with editing some of my field recordings to create ambient/drone pieces.
On the 2nd of November I was out and about around the Harbourside in Bristol making some field recordings.

Field Recording on the pontoon beneath Prince’s Street Bridge
I borrowed some kit from the universities asset store (Sound Devices 552, Rode NT4) and made some basic stereo recordings from various points around the Harbourside.
Here’s one of those recordings, after normalisation and trimming in Audacity.
[soundcloud url=”https://api.soundcloud.com/tracks/529872984″ params=”color=#ff5500&auto_play=false&hide_related=false&show_comments=true&show_user=true&show_reposts=false&show_teaser=true” width=”100%” height=”166″ iframe=”true” /]
As you can hear in the above clip, theres some pretty prominent violin playing from a guy to my right at the start (he was pretty good!)
I decided that it might be fun to try and use some spectral editing to remove some of the partials that were not his violin, with the aim of making his playing more prominent and hopefully without completely trashing the original sound in the process.
I used SPEAR to do this, which is pretty fun to play around with and free! Here’s the result.
[soundcloud url=”https://api.soundcloud.com/tracks/529871568″ params=”color=#ff5500&auto_play=false&hide_related=false&show_comments=true&show_user=true&show_reposts=false&show_teaser=true” width=”100%” height=”166″ iframe=”true” /]
As you can hear I (somewhat) succeeded in making the violin more prominent, however I kind of turned the rest of the recording into a strange, banshee like sine wave fest. Not exactly what I was aiming for, but since SPEAR actually re-synthesises the sound, it’s kind of impossible to avoid this if you remove lots of the quieter partials. At least for me it is, more practice and time with it will help i’m sure.
Anyway, I then took this file and fed it into MacPOD (another cool piece of free software) and did some granular synthesis mostly just using snippets from the first 10 seconds or so where the violin is playing. Here’s the result of that.
[soundcloud url=”https://api.soundcloud.com/tracks/529872387″ params=”color=#ff5500&auto_play=false&hide_related=false&show_comments=true&show_user=true&show_reposts=false&show_teaser=true” width=”100%” height=”166″ iframe=”true” /]
I’m pretty happy with this as both something substantially more listenable (to my ears at least) than the original re-synthesis, and significantly far removed from the original field recording.
I’m currently playing around with Max, building a very basic granular synth patch, in part 2 I’ll talk about that, share the patch and hopefully some interesting sounds I’ve made with it!