When Danielle informed me that she had booked me as a singer at The Gin Rummy, I approached the idea with no little trepidation. After all, I’ve never really done this sort of thing before. But that means I have to come up with an hour of material somehow…so I applied myself to the task, and hopefully the end results will not feature me falling flat on my face.
Danielle and I differ somewhat in our methods for producing our “live” vocals. She scouts out karaoke tracks on the Net (as MP3 files or similar) and then sings the lead vocals to them. I work with MIDI files, which is a technology I’m quite familiar with, as I used to do arrangements of pop songs as MIDI as a hobby. Trouble is, I didn’t have a decent sequencer program on tap. I scouted around, and located the freeware version of Quartz AudioMaster, which handles both MIDI and audio recording, and does so with a user interface that looks and feels quite similar to Cubase, the last sequencer program I used. True, it handles only 16 tracks of MIDI and 4 tracks of audio, but my needs aren’t great…and, when I need audio tools the freeware Quartz lacks, I can always export audio and load it into Audacity for further processing.
So…now I have to pick out some songs. Here’s the criteria I go by:
- The song has to have a good-sounding MIDI file that I can find somewhere. For instance, I’d love to do Aerosmith’s “I Don’t Want To Miss A Thing,” but the only MIDI files I could find for it, quite frankly, sounded like ass. Yes, I know how to sequence MIDI, but I just don’t have the time it would take to sequence my own files right now.
- I don’t want it to be a song that’ll sound stupid if I sing it. That eliminates most, if not all, songs done by female vocalists…though I could see myself making an exception for Nena’s “99 Luftballons” in the original German.
- I have to be pretty sure the song’s in my vocal range. That’s why I’m not doing Marillion’s “Cover My Eyes”; I can’t be sure I’ll hit the high notes in the chorus (“Pain, and heaven!”) properly. And if either Quartz or Audacity has an equivalent to the “vari-speed” control on a 2″ tape deck, which is sometimes used to lower the key of a song to where a singer can sing it properly, I haven’t found it yet.
- I have to have the lyrics down pretty cold, or be able to chart them out correctly. That’s why I had to pass up Oingo Boingo’s “Dead Man’s Party” for now; it would have been too much work to properly chart out the lyrics for that version of the song (which is not the same as the one on the album). Maybe later…
- And, last but certainly not least, I don’t want to step on Danielle’s toes. I have an excellent MIDI file for the Smiths’ “How Soon Is Now,” and I’m pretty sure I can sing Morrissey’s range, but Danielle wants to do that song, so reluctantly, I set it aside. That’s also why I’m not doing my own version of the Eagles’ “Hotel California,” either.
Okay…decent MIDI files in hand, I set forth to produce some songs, hampered only slightly by the nature of the tools I’m using and undeterred by the fact that about all I know about music recording, mixing, and production is what I learned reading The Daily Adventures of Mixerman.
First, I load up a MIDI file in Quartz, set up an audio track to record, check my levels, and lay down a vocal. (Easier said than done sometimes…at one point, I had Pamela peeking in, flashing the lights at me, and trying to tell me to “keep it down.” WTF? I’m recording here!) I go over the recording, listening to see if there are any egregious screwups. Sometimes, I’ll have to redo the whole take; other times, I can just “clip out” a bad portion of the audio track and record a quick “punch-in” fix for that line or verse. If there are long instrumental sections in the song, I’ll clip out and erase the audio track in those sections, to keep from having excess noise in the recording. I adjust the equalization on that track, too, to kick up the “low end” of my voice a bit.
Then I have to prepare an audio mixdown of the MIDI track, to one of the other audio tracks. I mute my own vocals for this, of course, and, if the MIDI file has a track or tracks containing the song’s vocal line (most do), I mute those as well.
Now I have two tracks of audio, the raw vocals (with some EQ) and the mixdown of the MIDI tracks. I export those from Quartz and import them as two audio tracks into Audacity. To my vocals, I apply some dynamic range compression, to balance out the dynamics and tighten them up. I adjust the gain on the two tracks to make them as loud as I can without distortion (since they seem to lose some gain in MP3 conversion), and give my vocals about a 3 dB edge over the music to make sure they’re not “buried.” (Not a hard-and-fast rule…I just go by what sounds right.) At this point, the track is done, and I export it as both an MP3 file (for previewing and a time check) and a WAV file (for later combination into the final “set” recording).
Lather, rinse, repeat for each song I want to do. (I don’t have to do all the songs in this order…in this case, I laid down vocals for a bunch of songs, then waited until later to mix them down.) And still I’m not done…because, as Danielle keeps reminding me, I have to talk in between songs. So I need to write some appropriate “patter” to go between the songs, record those little bits, and then splice them together with the completed tracks as WAV files to create one monster WAV file for the set, which I will then crunch down to MP3 for transmission.
The performance will be on Saturday at 7 PM, at the Gin Rummy’s new location in South Sunset. Just in case you want to come and throw tomatoes. 🙂
UPDATE: Holy crap, these tracks are loud! I don’t know why Winamp attenuated the volume as much as it did, but the gain on these tracks needs dropping by about 3 dB or so; I only found this out when playing the tracks under mpg123 on Linux. Fortunately, that’s as easy as cranking down the “gain” sliders in Audacity and re-exporting the tracks.
UPDATE: I glossed over Danielle’s techique briefly at the start of this entry, failing to do it proper justice…I understand she employs a similar set of techniques to get her voice to sound right. Certainly it takes her as long as it takes me to get workable tracks…and I think her results sound much better. But that’s what happens when you start with way more talent…she’s an entertainer, I’m just trying not to completely stink up the joint.