My Jesus I Love Thee started out as a poem written in 1864 by William Ralph Featherston. In 1876, Adoniram Gordon added the melody. I first became aware of the hymn (not having grown up with it) when it was included in The Acapella Project II by Glad. Ironically, I didn’t care much for it – but it was mostly the rather atonal-odd harmony arrangement they used in the recording. As the years progressed, I kept running into it and eventually grew to love it.
The second thread that ties this together is a magazine I had in my early 20s – I think my mother picked it up at a used book sale somewhere. It was Sheet Music Magazine, which only ceased publication this year (2013). In this particular issue was a mash-up of Beethoven’s Moonlight Sonata and Cole Porter’s Night and Day. Now, I can’t say where I got the specific idea to give My Jesus I Love Thee the same treatment, but somewhere those two came together. The sheet music I wrote has a copyright date of 2007, so it’s been at least 6 years now since I finished it. I do remember that once I completed the piano part, I thought it needed a little something more.
I did a whole series of posts on the original recording. In October I mentioned that I had completed writing the piano part but wanted to add an “ethereal” pad. By November that had matured into a cello part. I struggled with getting a decent cello recording until the following May when I hired a string player in Germany to program the part using the Garritan Gofiller cello instrument. I thought I was done and published, but as I listened to the track over time, I was never happy with it. For starters, it was too fast. Second, I’m not that happy with the Garritan Steinway (at least, the least expensive version) for gentle, soft playing. It just gets too bright, too easily. Lastly, the Garritan Gofriller just didn’t have the sound I wanted for a contemporary recording – too classical. I solved the piano issues last Fall when I took to Norfolk First Baptist Church to record their Steinway live and in person. This piece actually represents the last of the tracks for this project that I recorded that day. I pondered a bit about what to do for the cello part, and eventually posted an ad on Craig’s List looking for a performer. BJ Griffin responded and laid down the track for me back in April. Finally all the parts were in place, and I have a version I am HAPPY WITH!
As mentioned in the earlier post, we recorded in the main sanctuary of Southside Baptist Church here in Suffolk, VA. Their brand new sanctuary is decked out with significant acoustic treatment and sounds really very good for a converted gymnasium. Actually, while I wouldn’t want to record a banging drum set in there, it sounds just wonderful for my purposes. The pink tracks in the screen shot represent the mid channel, and the grey tracks are the side channel.
We ended up with five complete takes of the song and four or five more stabs at the ending. I spent several evenings in a row comping together a best take from all the recordings, and the screen shot to the left shows the results.
I was a little underwhelmed by Sonar’s track grouping capabilities (at least version 7, which I am still using). I was hoping that I could group the tracks and any edits I made to one would be reflected in the other. Unfortunately, what I had to do was manually select all the clips I wanted to perform a specific edit on (generally two, one from each track) and to the split or drag or whatever. Too many times I forgot to grab the second clip and had to undo my edit and make sure I had everything I needed. This is why I bounced the 5 tracks of piano recording down to a single stereo track before doing any editing/comping.
This shot shows the final audio tracks with all the bits and pieces put together. While I never specifically wrote about it, I comped the piano together the same as for the four previous tracks I recorded at First Baptist. Also visible (if you blow the screen shot up) are the automation lanes. The envelope in the piano window is the volume – and it’s static right up to the final couple of notes. Essentially, the first of the final three notes was too strong, so I used the automation to clean it up. A compressor would have affected the entire track, and I only wanted to fix that one note. There is no automation in the separate cello tracks, but they are combined to the first of the two aux, or bus tracks at the bottom. The automation on the two bottom tracks is also very simple – the cello is faded down at the end to more closely match the natural decay of the piano, and the reverb is ramped up at the end.
The piano channel is relatively simple in this mix. I’m trying a new compressor – the TDR Feedback compressor II. It comes Bootsie recommended, and seems to provide and really nice, clean compression. Granted, I’m not driving it very hard – just taking a little off the top. The equalizer is just the standard Sonitus EQ that comes with Sonar. Since I’m only making cuts with this, I wanted a good, clean EQ. I cut some of the lows because the piano was a bit boomy in the mix, and then cut a little more at ~1100 hZ to clean up some mud.
The two individual cello channels are bused to the same aux channel, and then all the processing is taken care of there. First is the Zoom M/S decoder, which takes the mid and side channels and turns them into stereo. I used two EQ’s on this track. This has become my standard practice when I want to cut some frequencies and boost others. Typically I do this for two reasons – first, I generally want the cuts to be clean and surgical, and the Sonitus EQ excels at that. Then I want the boosts to be musical and character-full, and the BootEQ does great at that. Second, I usually want to cut before the compressor and boost after. Here’s a good article on that. In this case I’m not using a compressor on the cello because I didn’t need one, but the first principle still applies. I’m using the Sonitus as a high pass filter and then the BootEQ adding a strong presence peak in the low mids for richness.
I’m using two reverbs for this track. I’ve struggled this entire project putting together a reverb sound that I like for the more ambient pieces. The in-our-face tracks are a piece of cake, and ValhallaRoom works that space masterfully. I’ve had less luck using it along for longer reverb tails. I’ve also not been happy with the longer convolution reverbs, as they often sound more distant than I would like, even when I adjust the predelay and wet/dry mix.
On this track, I decided to try and get the best of both worlds. I inserted ValhallaRoom first, set to the Large Ambiance patch. It is a 0.8 second long reverb, so when I loaded up SIR convolution reverb, I set the predelay to that same 0.8 seconds. This means that the second reverb does not start working on the sound until the first is completely done. The impulse response itself is the Timeworks Church from the Echochamber collection. (Echochamber no longer appears to have these IR’s posted, but if you would like them, drop me a line.) Anyway, I think I’m happy with the results. I may experiment some more with the last two tracks on the project.
The reverb channel itself is very simple. Note that I haven’t mentioned the Slate Audio RC Tube plugged into every single channel on the mix. That’s because that is now a standard and I’ve pointed it out in every one of the last four mixes I’ve published. Including, now, this one. The equalizer is placed after the reverbs to thin out the bass and keep everything from getting too muddy. Maybe you noticed that even though I only have two instruments in this song, fighting mud is still a priority. The Classic T-Racks EQ is free from IK Multimedia, and I decided to use it here somewhat on a whim. It works just fine.
The final mix before the output is again a fairly simple affair. After the Slate RC Tube, I inserted Bootsie’s Density Mark III compressor – still my favorite for the mix bus. It’s set to mid-side processing and driven just enough to take a little off the top. The sharp-eyed will notice that I’m using a different metering plug-in that my usual. I’ve recently discovered the Sleepy Time Records Stereo Channel, and I like it because it gives me options. The meters are also nice and easily readable. Now all I need is a peak hold indicator (like LSR LVLMeter) and I’ll be set.
So that’s it. It’s been a long journey for this one, but I’m really happy with the results. If you haven’t already, scroll back up to the top and have a listen to the video. I can’t believe that I only have two tracks left and this project is done and ready for prime time. To that effect, keep an eye out for another Recording Session post – this one should be pretty cool.
And remember, I love comments, but if you got here through my Facebook link, please remember to leave the comments here and not on Facebook, so that everyone can appreciate them!