Now that the basic studio has been constructed, it's time to collect some thoughts about recording. I'll admit, up front, that I'm an engineer by trade, a musician at heart with no claims to anything but personal enjoyment but I'm not a recording engineer so the following is a mix of personal experience while researching the experiences of others.

The web has provided a few general concepts which appear to be sage advise:
  • When using processing effects, adjust just enough to hear the effect; then back off a little
  • The better the original source track the easier it will be to mix and master
  • With modern equipment / software, never record "hot"; there is more than enough dynamic range
  • Unless you're using old digital, record in 24 bit and 96 kHz sample rate - disk space is cheap
  • If you're frustrated and tempted to "just fix it in the mix", you need to fix it in the source tracks
The following are commonly discussed topics so worth of attention:
  • Sample Rate and Bit Depth
    • 44.1, 48, 88.2 or 96 kHz -- in short, I chose 96 kHz
  • Recording Level
    • -20? -18? -15? 0? +3? dBFS -- in short, I shoot for -20 dBFS
  • Tracking
    • Getting the best original source tracks makes the rest of the steps easier.
  • Pre-Mixing EQ
    • Just like the importance of starting with good tracks, those tracks should be EQ'd before mixing
  • Mixing Spectrum
    • Terminology has developed over time to describe 6 major sub-bands of the audio frequency spectrum
  • Equalization
    • Adjustment of individual sub-bands within the audio spectrum to compensate for non-ideal sound
  • Mixing
    • This is a 'mixture' of the results of the previous EQ, individual faded adjusts and balance (pan)
  • Mastering
    • This step is focused on the 'song sequence' and how the mix will translate outside the studio
Sample Rate and Bit Depth
  • Sample Rate defines how frequently the music waveform's analog signal is measured and converted to a digital numeric value.
  • Bit Depth defines the resolution of that measurement where the maximum range of values begins at zero '0' and ends at 2^BitDepth. A 10-bit depth means the measurement range is 0 to 1023.
  • CDs are 16 bit, 44.1 kHz LPCM encoded with about 96 dB dynamic range
  • The Studio One DAW processes in 64 bit and encodes lossless at 24 bit 96 kHz for 145 dB dynamic range (although using real hardware such as the Presonus FireStudio Project the usable dynamic range is closer to 120 dB).
  • Dithering (adding a small amount of noise) is frequently used prior to output LP filtering of the recovered analog signal to reduce the noise artifacts (distortion) caused buy the sampling.
  • Hint: Use 24 bit, 96 kHz if your system can handle it.
Recording Level
  • There are three (3) metering units to be aware of when recording: dBFS (dB relative to Full Scale), dBu (+24 dBu = 0 dBFS and +4 dBu = -20 dBFS in the US RP155 standard) and dBV (+12 dBv ~= 4Vrms). For mixing with a DAW, the display shows dBFS. ( eg. 0 dBu = 0.775 Vrms, 0 dBV = 1.0 Vrms so +4 dBu = 1.23 Vrms )
  • There are four (4) important metering levels to remember when recording: ~-120dBFS (roughly equal to the noise floor), -20 dBFS (the "Reference Level"), -12 dBFS (Nominal Peak Level) and 0 dBFS (digital clipping limit).
  • Setting the "Nominal Recording Level" to 20dB below hard clipping, distortion on peak transcients will be very rare as most peaks are no more than 10 dB above the nominal music level. This still provides 100 dB of dynamic range above the assumed noise floor.
  • Common claims for "reference levels" include -18 dBFS and -15 dBFS but they never talk about either their noise level or the dynamic range of the music being recorded. If the dynamic range required by the source is know to be small and there is a need to maximize the recording's dynamic range, then reducing the headroom by using -15 dBFS might buy an additional 5 dB but at the risk of unexpected clipping.
  • Hint: Record with a -20 dBFS Reference Level.
  • Garbage in, garbage out. There is no substitute for starting with clean, crisp, un-colored original tracks. Trying to fix bad source tracks, remove bad room effects, un-do hollow or boomy mic placement using EQ just adds more work and prevents getting a natural sound from instruments -- especially acoustic instruments.
  • Acoustic guitar is very difficult to record so one option is to use 1) a pickup, 2) a specific guitar EQ like the Fishman Aura, 3) a pair of X-Y positioned small condenser mics and 4) a large condenser mic for ambiance. This will provide 4 close-in-phase tracks, each with a unique palette which can be later mixed, panned and EQd into a final mono or stereo field with greater flexibility in the mix.
  • Hint: Be liberal with assigning separate tracks for each part or instrument.
Pre-Mixing EQ
  • Sometimes call a "Starter Mix" or other term combinations, this is the initial pre-mixing preparation phase.
  • Each track is separately prepared using Pan (or Balance for stereo tracks) and EQ (only if required).
  • The objective is 1) final "touch-up" (remember, start with only quality tracks) and 2) EQ to "position" the track within the final mix (estimating where the track will "sit" within the final mix goals).
  • EQ-Cut: First, remove frequencies that either 1) do not add value to the track (i.e. rumble or hiss) or 2) will likely compete with another track where those frequencies in the other track are more desirable.
  • EQ-Boost: After EQ-Cut, boost track frequencies for clarity, crispness or added exposure for the future mix
  • Hints: Forget level and first cut frequencies before boosting others.
Mixing Spectrum
  • Sub-Bass (16-60 Hz)
    • Adds a sense of power -- more felt than heard
  • Bass (60-250 Hz)
    • Adds "thickness/fatness" -- where the rhythm resides -- cut to reduce "boomy" room responses
  • Low-Mids (250-2kHz)
    • Lower tones of musical instruments -- 300 Hz is the "center" of "muddy" sounds
  • High-Mids (2k-4kHz)
    • Vocals reside here -- carefully dip the instruments and boost the vocals
  • Presence (4k-6kHz)
    • Affects the "clarity" of vocals and instruments -- distance feels closer when levels boosted
  • Brilliance (6k-16kHz)
    • Clarity, brilliance, sparkle, crispness -- but also harshness when boosted too much
  • Fixed band EQ - The center frequencies and widths of the adjustment bands are fixed -- all that can be adjusted is the degree of 'cut' or 'boost' for each control
  • Parametric EQ - Both the center frequency and the "sharpness" (Q, the 'quality factor') can be adjusted -- allowing wide dynamic range of adjustments and precision in cutting offending frequencies.
  • Q: <1.0 is wide, 1.0 is medium, 1.4 is typical and >2.0 is narrow.
  • LPF - Low-Pass Filters are used to "roll-off" upper frequencies where the whole high-end is too harsh
  • HPF - High-Pass Filters are used to 'Roll-off" lower frequencies such as cutting out unwanted rumble
  • Low/High Shelf Filters - Have a "flat" response over a lower or upper sub-band and that whole range can be cut or boosted uniformly. High end EQs often have this feature.
  • Quick Rules
    • Be very careful in the amount of cut or boost used
    • Use high-Q for cuts and low-Q for boosting
    • Roll-off the bottom to "stick out" and Roll-off the top to "blend in"
    • Cut if you want things to sound "better"
    • Boost if you want it to sound "different"
    • You cannot add something that isn't there by boosting
    • Cut at ~250 Hz if it sounds "muddy"
    • Cut at ~500 Hz if it sounds "honky"
    • Boost the "mid-range" up to 6 dB to make a guitar "fat"
    • Boost 125-250 Hz range to make vocals more "chesty"
    • Boost 2-4 hKz to make vocals sound "closer"
  • Still to write ...