My wife celebrated her birthday a few weeks ago. It also happened to be the 20th anniversary of our engagement. I’ve been wanting to write for her a song for some time, and this seemed the right occasion.
So, how does one sum up 20 years in a 3 1/2 minute song? Not easily, I can tell you. This is also the first time I sat down to write a song about something not starting with a melody, or phrase, or hook to begin the process.
I began by sitting down with a notebook and just jotting down ideas, words and phrases. Interestingly, the opening line of the song, “Once upon a time when a boy met a girl at a dance,” was written word for word in my notes. I didn’t expect it to make the cut because it’s an odd meter and length, and I didn’t see it fitting in. Of course, I was writing everything I could think of, knowing I would match up and pare down later. But, it ended up as the opening line of the song.
I could only go so far with just a pad and paper. I’ve only ever written one song completely without a musical instrument in front of me (The Old Bethel Church, mentioned in this post), and I had the melody for that one in my head as I was writing it – and I had to rewrite some parts when I did sit down at an instrument to make the lyrics work…
So I sat at my piano and continued working. Once I did this, it came together relatively quickly. (Teresa is working evenings this summer for the Census, so I’ve had some time to work on this without her knowing – it was all a big surprise). As I wrote it, I envisioned it as a standard band arrangement – piano, drums, bass, maybe some strings. As I started recording, though, it changed completely.
It became an A capella doo-wop song. If you remember the old Growing Pains TV show (the one with Kirk Cameron), they re-recorded their theme song late in the series (it’s the one they use for syndication now). In retrospect, this song has a very similar feel. One friend described it as the theme song for the sitcom of my life.
Anyway, this ended up taking longer than I had expected, and I didn’t have the recording done for her birthday. I gave her the lyrics on her birthday, and she thought they were very nice. My failure was thinking (somewhat naively and somewhat egotistically) that I could just stand up and improvise all the parts in real time.
After that fell on its face, I broke the song up into sections, worked out the parts on the piano, and sang them in a little at a time. I sang along with the piano part for timing and key. Due to my being out of practice singing and rushed for time, the resulting tracks needed a lot of editing. V-Vocal to the rescue! By the time I was done cutting and pasting, here is what the track sheet looked like (I hid the original piano part):
The top track is the lead vocal. Below that is the track of adlibs at the end and the dah dah dah lead at the beginning. The next two tracks are the tenor and baritone harmonies. Those were the hardest to write and record. The bottom track in gray is the bass. That one I did just stand up and improvise against the piano part.
You can see the volume automation I used to balance the parts (the colored lines across the audio data). The tenor and baritone needed the most. I tried something a little different with this song. I first set a reasonable balance with the volume faders. Then I inserted compressors into each track. The automation is actually controlling the output level of the compressor.
I used Sonar’s VC-64 Vintage Channel on the lead vocal. It has a really smooth dual compressor that works quite nicely. It also has a built-in gate, de-esser and equalizer. The lead vocal also received the BootEQII equalizer (with saturation stage), the TLS Saturated Driver and the Antress Modern Exciter to help the part come to the top of the mix.
The Tenor and Baritone each received the Push-Tec EQ, a high-quality compressor with a saturation stage, and the Free Haas stereo channel delay. More about that in a moment. The bass received an Antress compressor and the Push-Tec EQ.
You may have noticed that I have used plug-ins with saturation stages, or in the case of the Lead, a dedicated saturation plug-in. Additionally I’m using the JB Ferox tape saturation plug-in on the master output. I recently read an article called Sonic Varnish that gives an insight about what is often missing from today’s digital recordings. The author’s contention is that what’s missing is many subtle layers of saturation, which is a form of musically pleasing distortion. In the days of analog recording, this saturation came as a matter of course from the pre-amp, all the outboard equipment, the board, and the tape machines. The advice I gleaned from the article is that adding subtle saturation to audio in many layers is analogous adding many fine layers of varnish to fine woodworking. The individual layers are hard to detect, but the overall effect is quite pleasing.
Now, as promised, about the Haas plug-ins. The Haas Effect is a psycho-acoustic effect in which the brief delays between sound arriving at one ear before the other helps us determine which direction the sound is coming from, even if the volume in each ear is identical. This works for up to about 40 milliseconds of delay, with greater delays increasing the strength of the effect. After that, it begins to sound like flanging. This is especially useful when you want the mix to work well in headphones as well as regular speakers.
The Free Hass channel delay is designed specifically to create this effect. It gets inserted into a mono track and outputs a stereo signal, with a delayed copy of the signal coming out the opposite panning as the signal coming in. The right knob changes the volume balance between the original and the delayed signal.
My experiment with the Hass Effect on this track wasn’t too convincing, though, because the harmonies use a lot of “ahhs” and “oohs” which don’t apparently respond well to it. I think that what is needed is a more percussive attack so that the delay is “noticeable,” at least on that subconscious level.
Here is the mixer window showing all of the final EQ settings, inserts and sends. I rolled off the lows on all the channels and the highs on the harmonies before any signals got to any processors or effects. This helps clean up the signal and keeps things from getting too muddy.
Each channel has two sends – both to reverbs. Both are using my go-to convolution reverb (that generates its effect from a recording of actual reverb from something like a real space or a piece of equipment), Reverberate LE. I think I’m going to upgrade to the full version of Reverberate pretty soon. I really like the LE and the full version has this great feature where you can modulate between two impulses. If they are similar the result is a nice movement within the effect. Two dissimilar impluses can be used for more creative effects.
The first reverb is an ambiance, which exists to create a sense of space around the vocals. I used a short plate impulse for this. The plate reverb was an early artificial reverb (used to make the sound sound like it was produced in a large, naturally ambient space, like an auditorium). It actually transmitted the sound into a large suspended steel plate with microphones around the perimeter. While it didn’t sound that much like a “real” space, it did have a pleasing sound that ended up becoming the sound people loved and expected on vocals and drums.
The second reverb is for the live “tail”, that shwoosh of sound hanging in the air that most people associate with reverb. I couldn’t find any IR’s of real spaces that I really liked for this song, so I turned to algorithmic reverbs (that generate reverb from a complex formula rather than a recording of “real” reverb). I almost settled on DaSample’s Glaceverb, which I’ve not yet actually used on a published song, but changed my mind when I found something special. (It also looks like DaSample is no longer around – the website link is dead).
The biggest name right now in digital reverb units is the Bricasti M7. No, I hadn’t heard of it, either. But then I caught a thread where a company that specializes in creating Impulses from classic reverb units, Acousticas was releasing a free set of Bricasti IR’s. People all over were gushing about the unit, so I downloaded the impulses (nearly 500 meg worth) to give them a shot. I have to admit, I am quite impressed. The 4-second reverb tail is beautiful without overwhelming the rest of the sound.
I adjusted the send levels of each of the channels so that the harmonies were a little deeper in the space than the lead. The master channel received my favorite 2-bus compressor, the Density Mk II. Then the aforementioned JB Ferox tape saturation plug, and then Sonar’s Boost 11 limiter to make sure the signal behaves.
Ladies and gentlemen, for my sweetheart: Every Day