On August 9, I released my third album as Spiral Island, Evacuation’s Out. In my next few blog posts, I’ll share some of what went into this album, starting with vocals.

Recording

All of the songs on this album originally came out of February Album Writing Month, and I had rough demos to iterate on. But the demo vocals generally weren’t worth keeping, because when I recorded those demos, I was finding the melody in real-time, and singing quietly because it was 11pm by the time I wrote lyrics. As a result, 99% of the vocals got re-recorded, and parts got added over time.

(There was one vocal track I didn’t redo, the “echo” vocal on In de Woestijn. I liked it better unpolished, and as a subtle Easter egg that it’s not actually a delay of the lead vocal.)

Acoustics: I refuse

My home studio is in the corner of a rectangular room that’s about 10’ x 15’. Aside from my junk, there’s not much to absorb reflections. I don’t have a great ear for acoustics, but I know that the acoustics in the room are garbage. Besides that, there’s a window facing the street and a vent for the radiator and AC.

The acoustics are something that I don’t feel qualified to fix or even judge, really, so I’ve tried to just avoid them altogether. The sound of an acoustically pleasing space is something I can always add later, in the mix.

I use a Shure SM7B, for that sweet NPR tone and its focused cardioid pattern. It’s mounted on a standing desk, so that I can sing standing up and have my laptop at that level to punch in / punch out.

For lead vocals, I get pretty close to the mic, and for backing vocals, especially when I need to hit high notes, I’ll back up a couple feet and project more. The mic has a proximity effect that picks up more low-end when you sing close, so I typically EQ out any frequencies below the fundamental note of what I’m singing (200-400hz).

In addition to EQ, I’ll typically use a de-esser, compression to tame the dynamics, and auto-tune. At the end, the vocal will be routed to reverb and/or delay buses. But I’ll get into that more later, in my post about mixing the album.

Auto-tune

Recording my Spiral Island albums has been an opportunity to try out new studio tricks and technologies. On my previous album, I did things like use pitch shift to reach melodies I couldn’t sing otherwise. I used a sample-based vocal synth to fill the role of a gospel choir, and a Yamaha vocaloid to reach notes too low for me.

The vocals on that album were all tuned as well, using Reaper’s factory default ReaTune, but without a lot of hand-editing. Listening back to it nowadays, I don’t think I quite achieved the slick pop sound I was going for, and this is one reason why.

Starting on this album, I wanted to make songs that were energetic and digital, and to have vocals matching that vibe. So I did a shoot-out of all the auto-tune options I could find at the time:

  • Melodyne
  • Nectar
  • Waves Tune Real-time
  • ReaTune
  • bx_crispytuner
  • Voloco
  • Little AlterBoy
  • Antares Auto-Tune Pro
  • RePitch

I could write another blog post on my impressions of all these, but for this album I settled on bx_crispytuner.

This was one of the more artificial-sounding plugins I tried, which I was more comfortable with on this album: the goal was to make catchy songs, not necessarily to fool anybody about my singing abilities.

It also does something to the vocal timbre, which I actually like on my vocals. And it does a pretty good job of figuring out what note you’re aiming for, which reduces the work of hand-editing it the rest of the way there.

How I used auto-tune

With a home recording project like this, I just kept doing takes until I had what I wanted, as close to the target notes as I can get. So generally all the notes you hear, I sang as best I could.

Even the low parts on Cinemascope–I sung those notes and then shifted the formant to make it sound like it was pitched down, but it’s not. (The rap verse on Wij Bevonden Ons uses a similar effect.)

One exception would be the harmony on Perfect Imagining, which is a pitched-up copy of the lead vocal. Harmonies made this way often get lost in the mix, and end up sounding like overtones/undertones of the lead vocal, but in this song’s arrangement, I needed the timing to be an exact match.

Other vocalish stuff

The gang vocals on I’m Gonna Sleep were just multiple tracks of me doing grunty voices.

Wij Bevonden Ons also includes a bit of Chipspeech. While I used a vocal synth for a gospel choir on the last album, I think these days, with the proliferation of deceptive AI-based tools, I’d maybe rather inlist actual backing singers, or use a more obvious fake, like Yamaha Vocaloid or Chipspeech.