(As always, skip right to the bottom for the finished product.) While the origin of the lyrics for Away in a Manger are shrouded in mystery, the two most popular music settings are known well. The more popular of the two is by James Ramsey Murray in 1887. Slightly predating it is another version by William J. Kirkpatrick, written in 1895. The two have different forms – Murray’s version has a simple repeating verse while Kirkpatrick’s is in an ABA form with a very different B section at the beginning of verse two. I have long played a medley of these two versions on piano at Christmas events.
As I was putting together a list of Christmas songs and hymns I wanted to include in Moments of Comfort and Joy I knew that this medley I’ve been playing for years would be part of it, and knew it should have a very folksy instrumentation. Also, all the times I have recorded my Appalachian Dulcimer to date I have just strummed it, and I wanted to do a little fingerstyle for a change. This was a great match and an arrangement was born.
By the time I was done the instrumentation was dulcimer, alto recorder, piano, harp and fiddle. The form is:
- Intro featuring solo dulcimer
- Verse 1 (Murray) with recorder lead and dulcimer accompaniment
- Verse 2 adding piano and fiddle
- Interlude with piano and dulcimer
- A section of Kirkpatrick with piano lead and harp accompaniment
- B section of Kirkpatrick with harp lead and pizzicato fiddle countermelody
- Reprisal of A section with piano, harp and arco strings
- Coda with restatement of Murray melody with recorder and dulcimer
I’m happy to say that everything in this recording except the harp was recorded live, and the harp was only computerized because I have an impossible-to-play-live key change in the middle of the harp part. It is literally a part that could only be played on a keyboard with a fully chromatic sampled instrument.
The dulcimer was recorded with a single small diaphragm condenser mic placed about 18 inches above the instrument. I have to confess that I cheated a bit on the dulcimer recording. I was unable to play the fingerstyle intro live, no matter how much I practiced. To make it work I just played the individual notes slowly and edited them to tempo in Cakewalk. This means that I can’t easily use that footage for the video, but if it’s the only way to get the performance, then it’s the only way.
The recorder was done in a few takes and only needed minimal editing after the fact – mostly to fix tuning problems with the instrument itself. It’s interesting that recorders have what are called “forked” fingerings, where you put down lower fingers on some of the open holes to tweak the tuning of the note. This makes it extra difficult because you are often not just changing one finger but sometimes putting more than one up and down at the same time. And this doesn’t always bring the note fully into tune for the unskilled player (like me).
I first recorded it with my large diaphragm tube condenser but the sound was a bit harsh. I went back and did it again with my ribbon mic and the tone was absolutely perfect. The ribbon smoothed out the highs beautifully and captured the tone of the instrument wonderfully. One more note on recording recorders – for many woodwinds the sound comes out of the entire instrument, including all the open holes. As a consequence, the best way to mic them is to place the microphone above the middle of the instrument rather than right at the bell. Flutes get the best sound right at the mouthpiece. When I tried to mic the recorder above the body of the instrument the sound was distant and anemic. Through experimentation I discovered that a recorder produces most of its sound right out the end of the bell, so that’s where I placed the mic.
One of my best friends is a concert violinist. She has been playing professionally for more than 30 years; her tone is beautiful and playing agile. What it isn’t, however, is folksy. A new friend plays violin at my church and the first time I heard her play I knew she was perfect for the part. I have a write-up on our recording session. Her woody, rich tone was just what this song needed.
Once I had everything recorded and edited, I bounced the tuned and edited tracks down to clean tracks, archived the edits, and copied my project and renamed it “Mixdown”. I’ve started doing this where I create a new copy for each major phase of the project, archiving old, unneeded tracks to save resources. By the time I’m done I frequently have a writing, scoring, recording and mixing version of each song.
The composite mix window shown above give you an idea of how each channel is set up. Every channel has a mixer emulator on it (except somehow I omitted channel 1!) and the standard Quadcurve EQ and most of them have channel compressors as well. The buses on the right follow suit, with the addition of the tape emulator on the master bus. There are buses for the reverb and ambiance effects and you can see the sends for those on the individual instrument channels. With only five total instruments in the arrangement I didn’t bother bussing any of them together before sending to the mix bus.
The composite screenshot shows each of the final tracks and the automation lanes I used. Typically the volume automation is pretty subtle, giving small boosts to featured instruments for each part of the song. I’m also automating reverb sends and some features of my compressors or eqs. Also, the one V-Vocal edit on the dulcimer track is a bizarre bug my editing copy where the timing played back correctly, but when I bounced the track down it was screwed up. I had to re-fix the timing in the bounced version.
My biggest struggle with mixing the dulcimer was in taming the attacks. Previously I used a de-esser to manage the pick noise, but since I finger-picked this part there wasn’t a consistent frequency to latch onto. It was just too pokey-outy in the mix. In the end I tried a number of different transient designers until I settled on Transient Master which came in the Komplete package from Native Instruments. I doubled this up with a compressor – the trusty (and old) Sonitius Compressor – to bring them under control. There is also a CA-2A compressor on the channel, and both the send and the peak reduction are automated.
Piano and Harp
The harp and the piano both have very similar processing, which makes sense since they have similar acoustic characteristics. Similar EQ dips are found on both channels, as the two together caused an unpleasant build-up in the 230-240 hz range. I also added the Kush Clariphonic to each. The primary difference here is that I only use the Clariphonic during the portion of the piece that features the harp – otherwise it was too strong. The automation lane marked Clariphonic on the Harp track is the bypass – up is bypassed, down is active. I added the smooth Clariphonic top end for the harp feature section and turn it off otherwise so it doesn’t compete with the piano.
The fiddle just has Prochannel processing on it, including a high-pass filter and slight top-end shelf on the EQ and the CA-2A compressor slightly evening out the dynamics. The send to the reverb is also automated, increasing the reverb amount some during the pizzicato section and all the way at the end during the ethereal double-stop played over the last notes.
The 2-bus has the most processing, which is typical for the way I mix. The channel emulator, tape emulator and Density Mk III compressor (set to mid-side) are pretty standard. Some new things I have added include the Unlimited brick wall limiter and the DPMeter LUFS meter. I have been finding that as my ears improve the Boost 11 limiter is sounding worse and worse to me. Essentially, if I engage any limiting at all the music begins to sound constrained and choked. There are a huge number of freeware vst limiters available so I downloaded and tried several out. I really like the results I’m getting from Unlimited, but a major downside is that the GUI is HUGE.
DPMeter and the Bluecat gain plugin work in concert to bring my overall average loudness to -14 LUFS so that YouTube will neither turn up or down my delivered video. I know that experts are saying don’t limit for a platform, limit for a sound, but once I have the sound I want to go ahead and fix my final volume so that I know exactly what I’ll be getting.
The video is the most ambitious that I’ve done for one of my songs. It consists of a scrolling score video captured from Finale, and I also overlaid tracking videos from the recording process. I’m pretty happy with the way it turned out. Please see my previous post for a detail of the video software I’m now using.
So, without further ado, here is Away in a Manger: