links |
articles |
blog |
| Anatomy Of A Demo Recently a friend of mine asked me to help produce her demo. Her name is Tanya Roberts. She is an excellent singer, but a little inexperienced, and wanted me to produce the session for her. We began with a brief meeting where we discussed the song selections, and the overall concept. Since this singer covers a lot of variety we decided to record a collection of excerpts, each one under one minute in length, and each one reflecting a different style. We would end up with short pieces fading into each other, totaling less than ten minutes overall. We picked ten songs, and resolved to shoot for a natural sounding recording that would honestly reflect what she might sound like live, with her own band. We would try to resist overproducing things as much as possible. The ten songs were: 1. Love Sneakin Up on You 2. Genie In A Bottle 3. Killing Me Softly 4. Chain Of Fools 5. Almost Like Being In Love 6. Fever 7. Turn The Beat Around 8. Guantanamera 9. Perfidia 10. We Are Family Since my friend doesn’t have a regular band, we needed some backing tracks for her to sing along with. Ideally, these tracks should present the illusion of a great band backing up a great singer, live. We had some great players lined up, but one little problem. Nobody was able to come at the same time. I decided to start with a MIDI sketch, and build our “live” tracks up from there. This would allow the various musicians who were going to participate, the option of coming in whenever their schedules permitted. We resolved to create a MIDI sketch, outlining the form for each tune. The MIDI sketches would include a basic drum loop, a MIDI bass part, and a rhodes or piano comp part. Once this was accomplished we could replace the MIDI parts with “real” instruments, using the click track generated by the MIDI file at the musician’s convenience.. DAY ONE MIDI sketches. We were very lucky to have Mark Liuzza, a great keyboard man on hand so I didn’t have to play anything. This left me free to work the controls, and made things move along very smoothly. Using Performer, on my G4 Mac, I quickly established tempos, and recorded Mark’s parts. First, four bar drum loops, followed by bass and piano parts. We really caught a break because this particular cat is used to doing both at once. I actually recorded the bass and piano parts at the same time. He wasn’t real happy with everything he played, but it was good enough for my needs. I kept having to remind him that these were just sketches, outlining the form. Many of the MIDI instruments would be replaced with the real things later. We were only doing one verse and one chorus from each tune, and Mark knew the material very well so we only spent about twenty minutes on each song, give or take a few. After about two hours we had eight of the projected ten MIDI sketches. Tanya and Mark left and I spent about an hour cleaning things up with quantizing, note corrections, and cut and paste stuff. 3 hrs total DAY TWO Recording MIDI parts in PARIS. Lead vocals. Working on my own, I created the remaining two MIDI sketches and began the process of recording each MIDI part into PARIS, which is the hard disc recording software I like. I sync PARIS to Performer, and record each track one at a time. I create a separate PARIS project for each sequence (song). So why take this step? The midi machines have mixing and effects capabilities so it would seem unneccesary. For some reason, everything sounds better, mixed this way. I think it has to do with the synth devoting all of it’s processing power to one sound, but I’m sure there are a multitude of factors at play here. One advantage of this is the way it forces me to focus on each individual MIDI part. If I need to change the sound, or I detect errors in the part, I can fix things as I go. Effects are off at this stage so I really get to hear the rawness of the sounds before effects. Once the tracks have been recorded, including a click track, I can more easily mix the MIDI instruments together. Now that the MIDI parts are in an audio project I can close the Performer program, and I’m looking at a mixing console that represents an accurate picture of each sketch. If something is too loud I simply grab that fader and turn it down. The PARIS system comes with a control surface that behaves just like a mixing board with faders, so it’s much easier to mix this way. I don’t have to grab something with a mouse if I don’t want too. Most of this stuff will get replaced as I go along. The midi bass part will be replaced with a real bass, midi drums get replaced with real drums, etc.. I’ll probably end up keeping the midi keyboard parts for the most part. The neat thing about doing things this way is the flexibility. I can do drums last if I want to. At this stage, I am left with a pretty good audio representation of the MIDI arrangements in the right key, and my singer wanted to sing, so we recorded most of the lead vocals, and a few backgrounds (AKG C-1000). We cut all the leads first, and came back for a few overdubs. Couldn’t go too far with the female overdubs because we wanted it to be performable. This was another spot where it was important not to overproduce. 5 hrs total DAY THREE Bass parts. At this point I am trying to play every part perfectly so I won’t have to come back. I used my Fender Jazz bass through my trusty Manley instrument pre amp. From there it goes to an Alesis 3630 compressor/limiter, which is set to catch peaks. I have to make sure I’m not facing my monitor when the tracks go down or I get a nasty buzz. Usually I’ll do a few practice passes, then hit record and step away from the monitor a good six feet, and turn my back as well. 3 hrs total DAY FOUR Guitar parts. Used the Manley again, with various guitars. Strat for this, 335 for that, etc.. The Manley works great for any clean tones. For the nasty stuff I use my Mesa Boogie. I use the effects in PARIS on the guitar. Great chorus and delay options in there. I set a delay time to an eighth note, with zero feedback, and I mix the dry/wet ratio to only about 30% wet. On the guitar channel I set the input level on the effect send to -10. This seems to be just the right amount of delay for my guitar sound. You don’t really hear the delay, it’s more of a feeling. The delay feeds the dry signal back on top of the channel’s dry send so there is a doubling effect too. Set the effect send to “post” fader and pan the effect away from the original signal. This spreads the guitar out for a stereo effect and makes it appear bigger. This is the key to my recorded guitar sound. 3 hrs total DAY FIVE Drums Jeff Mills (excellent drummer) came over. I miked the drums with a kick (D-112), snare (SM 57), and two overheads (SM-81), and we whipped all ten tunes out in two hours. We were picky about the playing, but not too much about the sounds because we kind of wanted that “sloppy sound man” live feel. Balanced, but nothing fancy. A little room reverb on everything but the kick. 2 hrs total DAY SIX Remaining vocals Tanya came over and sang the remaining female parts. 1 hr total. DAY SEVEN Keyboards Additional keyboard parts Alesis 081, Young and Chang Baby Grand Piano (2-C-1000), male background vocals (C-1000). Used the synth for organ parts, and background pads where needed. The real piano was just the thing for one of the tunes because it was a jazz standard.. I sang low male parts. Tried not to go too far with overdubs and made an effort to vary my tone and inflection, to create the illision of two men singing. 2 hrs total DAY EIGHT Mixing Mixed (PARIS), edited (PEAK LE), mastered (T-Racks), and burned CDs (JAM) of the edited mixes. 1. The mixes are accomplished in PARIS with the “bounce to disk” function. Up till now everything has been recorded in 24 bit resolution, but for the final mix, we need 16 bit format. Once I have the mixes, I select the stereo L and R files and choose the “Export stereo file” function. This creates a new file from the two mono files. The new file is what the CD burner needs. It is in sd2 format, and is an “interleaved” stereo file containing both the L and R channels. 2. The next step is to open the sd2 file in PEAK LE. In this program I can trim the start and end points. Set fade ins and outs, etc.. Once that is done I save that file with the changes I’ve made. 3. Next I open the new file In T-Racks, which is a lovely program that emulates an analog mastering suite. It has a very friendly interface designed to look and operate just like analog gear. It includes parametric eq, compression, and limiting, and the effects are first class!!!! Here is where I set the overall level. Up to this point, I’ve tried to leave a little headroom in the mix so I can drive the compressors at this stage. The result is an “in your face” mix that really is radio ready. 4. Finally, I open Jam, assemble the tracks in the order I want, and audition all the transitions between tunes. When I’m happy I burn a CD, and check it in the player in the living room to make sure it plays on everything. If I’m in a real picky mood, I’ll ride in the car with it for a day or two. 3 hrs total DAY NINE Final Tasks Added one last background vocal part to one song, remixed, edited, and mastered that tune, and settled on a final order for everything. Burned the finished disc. Done! Total hours used: 22 Musicians: Drums, Bass, Guitar, 2 Keyboards, one male and one female vocal. Studio costs: $35.00 per hr x 22 hrs = $770.00. Fee waived, in lieu of being first call for any gigs my friend gets from the demo. (We’re real good friends.) All the players donated their time so there were no musician’s fees. (Probably would have been about $1000 to $1500.) Software: Performer (Mark of the Unicorn), PARIS (Ensoniq), PEAK LE (Bias), T-Racks (Multimedia Prod.), Jam (Adaptec). Value of project: $2270.00 Actual cost of project: $0.00 This would have all been faster and easier if the musicians could have all come in at once, but doing it this way let me spread things out over a nine day period, which turned out to be good for the details. When you are planning your budget, allow a little extra time for the engineer to live with the tracks. He can scrub things so they are easy to mix, digitally edit out mistakes, experiment with eq and effects, and generally get to know the mix before he runs it. Many people shortchange their project at this stage, but it is ill advised to do so. By choosing short, abrevated versions of the songs we were able to keep every session short, so nobody got burned out. What we ended up with was about 6 minutes of music that fades smoothly from one song to the other. Each tune is less than a minute long, and there are seven tunes in all. (We cut three.) Each tune reflects a different style of singing so it is a real fast way to get to know Tanya’s capabilities as a singer. It is also a fairly honest representation of what one of her gigs might actually sound like. That was our goal all along. Mission accomplished. |
|