My first mix to come out of my Steinway recording session is an Interlude, subtitled Peaceful.
I have a number of back issues of Keyboard magazine from the 90’s that my mother picked up for a song at her library’s book sale. I have read and re-read those magazines over the years. To give an idea, current issues run about 64 pages – these are all nearly 200. Lots and lots of content – only Sound on Sound comes close in the current market.
Dave Stewart was a regular columnist at the time and would frequently post rather unusual chord progressions as part of his music-writing lessons. One particular progression caught my interest and I wanted for years to do something with it. As I started composing interludes for this project I thought of that progression and pulled it out of the file, and I’m glad I did. Thanks, Dave, it’s beautiful – I hope you like what I did with it.
The time signature for this piece is a rather unusual (5+6)/8 – i.e. a total of 11 eighth-notes broken up into a group of five and a group of six per measure. I didn’t really have a solid musical reason for doing this other than experimentation, but the end result is really quite engaging, I think.
Once I had the basic structure worked out (the arpeggiated eighth notes in the left hand) it didn’t flow quite as nicely as I had hoped. I had originally thought of putting a solo instrument with the piano and began working out a melody line in the right hand. The more I developed the melody, the more I liked it and less I wanted to crowd it with another instrument. The resulting melody turned out beautifully and ties the whole together very nicely.
When I brought the recording for this piece home, I had five separate tracks of audio. There was a microphone over the treble strings, another over the tenor/bass strings, and third back from the piano and raised up high to collect a more ambient tone, and a stereo pair located under the piano, just behind the front legs.
I recorded all the multiple takes straight through without turning off the recorder. If I messed up I stopped, waited a moment, and picked back up from before the clam. I ended up running most of the song twice (the long unbroken splotches on the screenshot), and had to work fairly hard to get the ending right (all the short, broken splotches).
When I imported the tracks into Sonar and listened, though, there was a lot of noise in the recording – hiss and rumble. I don’t know where it came from – the mics, the setup or the recorder. Without having the recorder to test (I returned it to its owner after the session) my best guess is that I didn’t have the input level turned up enough and I’m getting a lot of room noise in the recording. I know it isn’t the mics because I have successfully recorded with them in the past. The other possibility is that the converters on the R16 are really noisy, but without the unit I cannot confirm that. Given the high quality of the converters in the less expensive H4 this seems unlikely.
First, I mixed the five tracks together (see below) and imported them into Audacity and used the Remove Noise function within that software. When I listened closely, however, there was a very low-volume weirdness going on. It was sort of a modulated digital distortion that wobbled from left to right. I tried to capture it in a clip to post here, but I don’t know that it turned out. Listen on headphones and turn it up…
[soundcloud url=”http://api.soundcloud.com/tracks/65050251″ iframe=”true” /]
Not very loud, but totally unacceptable to me. Despairing that my entire Steinway recording session was wasted, I called it a night. Fortunately, I had an idea…
I went back to the original tracks I recorded and loaded each one into Audacity and de-noised them one at a time. The theory was that the algorithm would have a much easier time with mono tracks than stereo, and mixing the five separate tracks together would then cause the resulting audio hash, at some level, to cancel out. Whether either of my theories were true, the final results were not just acceptable, but wonderful. Project saved!
Mix, Stage 1
The mix of the five original tracks was fairly straightforward. My goal was simply a good balance of the mics. I rolled off the LF rumble on each track (checking how high to take the frequency by listening for when I began to lose actual music from the signal) and adjusted the balance.
The two main mics were panned a little bit left and right (the offset in the panning is to help balance the stereo image). The room mic is dead center, and the stereo pair underneath are hard left and right (like a stereo pair should be…). The room mic has a little more EQ applied to help it support the other mics better, and has some additional gain because it was really quiet, even when compared to the low levels of the other four mics (I also added about 12 dB of gain to that track in Audacity).
I added more gain with the Boost11 limiter – about 20 dB according to the screen shot. My levels were really low, but now I’ve figured out how to make them work. Thank goodness for 24-bit recording (which pushes the digital noise floor down low enough to make it possible). If were recording in 16-bit I might not have been able to save the recordings.
Mix, Stage 2
Once I had a stereo track to work with, I created a new Sonar Project and imported the file. I did this mixing in two stages because I believed the that the editing would be much easier if I was dealing with one stereo file rather than 5 mono files. I know that when comping a take you can link the tracks and edit them all at once. However, I was planning on such complex editing that I didn’t want to mess with that.
In the end, I used all of full take number 2 and just tacked about the final 10 seconds on from another take. I tried to do a little more fancy editing together, but the volume difference between the two takes I wanted to merge was too significant to get them to flow smoothly. The edit I did end up with turned out really nice, though. It goes by seamlessly.
I applied processing to this final, edited take. The screenshot shows both the piano channel and the master channel. I could have just put all the effects into one or the other, but I did it this way just so that all of them would show on the screenshot to show the order in which I placed them.
The effects are a veritable Bootsie-fest. I started with his BootEQ Mk II, which I’ve found to be a musical and useful equalizer. I had originally wanted to use his BaxterEQ, because I love the concept, but it turned out to be too subtle for the tonal shaping I wanted here.
Next in line is the Density Mk III compressor. I ended up compressing more than I had expected to. The initial attacks of the notes stood out too strongly for my taste (likely a recording issue, since I thought my playing sounded like what I wanted in the room). So, I ended up driving the signal into the compressor a bit to take those tops off. I ended up with 4-6 dB of gain reduction on the peaks when all was said and done.
p.s. If you are getting here from my Facebook link, please remember to comment here on the blog post and not on the Facebook link. Thanks!