In its simplest definition, mastering is a post-production process whereby a song (or multiple songs) is (are) prepared and transferred to a data storage device - the master - which is in turn used for duplication and distribution.
On a high level, that's all a mastering engineer does. Conceptually, it's very easy but in practice, it is not.
To illustrate why lets all climb into a virtual time machine and venture back to 1975 where "Steel Dodos" is recording their first album, "We can't fly."
The mixing engineer, under the direction of the producer, just finished eight songs resulting in a 1" stereo tape of this effort. Bob the intern is handed this reel of tape and runs down the hall to the mastering engineer, who through his seemingly magical ways, converts that stereo recording into something that can be mass produced - a record - a plastic disk with some grooves on it, sleeved in cardboard, to be sold in mass quantity for $12.99 at K-mart. "Steel Dodos" is already counting their projected royalties so it's up to the mastering engineer to do this thing, and do it well.
Tape, especially 1" studio quality tape, has a fairly flat frequency response as well as an "okay" dynamic range.
Records on the other hand, does not have a flat frequency response and it has a worse dynamic range than tape - but here is the kicker - the frequency response as well as dynamic range
changes depending where on the record surface the needle is. and is also very different depending on the needle type, how the needle is suspended, and whether it's a ceramic needle cartridge or a magnetic cartridge. Wow, that's a lot of variables, don't you think?
It gets far worse too
Consumer phonographs, even good ones, have the needle mounted on an arm and that arm is mounted on a pivot, at the back of the unit often on the right side. So the path of the needle is not a straight line, but rather an arc. Because the path of the needle changes with respect to how the grooves are cut, the needle will be more sensistive to right-side grooves when the needle is closer to the outside edge of the record, and more sensitive to left-side grooves as the needle progresses closer to the center of the record. With these needle position changes (i.e., the arc) frequency response as well as basic balance of left to right also changes.
Why does this matter? Fancy record master-cutting metal lathes cut the grooves linearly, meaning that the arm is not fixed on a pivot and instead slides across the machine like a print head - so the needle is always perpendicular to the cutting surface
and to the groove being cut. The record is cut differently than it's played, and that by itself introduces a lot of audiophonic abnormalities.
The mastering engineer knows this, and compensates for this.
Also, because the dynamic range of the record is far less than the dynamic range of the original stereo tape , that has to be compensated for as well - through compression and limiters of varying types. Some songs may have more volume then others, and levels have to be adjusted so that all of the songs on the true master (the metal disk that the cheap $12.99 copies are stamped from) is approximately equal so consumers who buy records don't have to constantly adjust their volume knobs song to song on their phonographs. Sometimes that volume differential is overall volume, other times it's because of bass frequencies. Sometimes harmonics are an issue with records whereas on tape it wasn't.
The mastering engineeer knows this too, and compensates.
Finally, and less understood, are the irregularities of the mediums, tape to record.
Tape has a fairly accurate speed regulation system - often a roller attached to a flywheel driven by the motor via a non-stretchy belt. This roller pinches the tape against another free-spinning roller, and speed is maintained. The motor has a smaller diameter pulley than the flywheel, so irregularities in motor speed is "geared down" and reduced. We measure the inaccuracy of tape speed by "wow and flutter".
Lower quality decks use smaller flywheels and crappier belts, high end decks use bigger flywheels, more precise motors, better belts, and often a stroboscopic system to make a feedback loop with auto-correction - kind of like a tachometer in your car with a little "brain" to maintain a specific RPM.
Anyway, the tape is played back start to finish at the same speed - "X" inches per second, with minimal variance. Start to finish, the dynamic range and frequency response is fairly consistent.
Records suck in this regard - because records are round and the recording follows a concentric path that gradually gets smaller in diameter as the needle gets closer to the center, the speed of the grooves passing by the needle changes constantly - continually getting slower. At the outer edge of the record, the plastic spins by far faster than at the center - and that can be calculated out like so:
Record = 12" diameter on the outside, 5" on the inside.
- 12" diameter x 3.1416 (pi) = 37.6992 inch circumference
- Rotate the record at 33-1/3 RPM and we have:
- 37.6992 inches x 33.33 RPM = 1256.514335 inches per minute.
Lets do this again with the inside groove diameter, 5"
- 5" diameter x 3.1416 (pi) = 15.708 inch circumference
- 15.708 x 33.33 RPM = 523.54764 inches per minute.
Obviously, 1256 inches per minute is a faster rate of plastic zipping by than 523 inches per minute.
With that continual speed change (slowing down), the dynamic range of the audio changes as well - there's more audio fidelity available available to the first song on the record than there is on the last song on the same side. Add to that the arc-issue I mentioned above, this becomes fairly complex.
The mastering engineer knows this, and compensates accordingly.
There are other anomilies with records - certain frequencies can resonate the needle and cause unwanted audio artifacts, and if a particular recording has a heavy paylod of these frequencies they need to be dampened precisely with a parametric EQ without damaging the overall sound of the recording.
The mastering engineer knows this, and compensates accordingly.
Another factor that's very important is what the purchaser of the hot new "Steel Dodos" album is going to play the record on - a $19.95 fisher-price "batman" phonograph, or a top of the line Denon with laser tracking? The record needs to sound "good" on both devices, and everything in between.
You can try this yourself: Take your favorite album (that your purchased) and play it on a boom box, a bookshelf-sized stereo, your car stereo, and your home theater system. It sounds different on each system, yet sounds pretty darn good on all of them.
Somewhere, at some point, a mastering engineer put a lot of time and energy to achieve that "sound good" aspect on all those devices, using his or her knowledge and experience to make that happen, without having those devices at his fingertips.
In my mind, that's the true "art" aspect of the mastering engineer - the ability to accurately imagine what the recording will sound like on many playback devices, without those devices being immediately at his disposal, and get it right every time.
As you can see, mastering is a bit more than twiddling EQ and setting a limiter - it requires actual knowledge - it requires skills honed through experience - and in fact, "art".
---
In today's world a mastering engineer is still necessary to "adjust" the better quality, higher bandwidth studio recording to fit onto a generic, minimalistic cheaper-than-dirt consumer platform - a CD - with potential to be played on a $19.95 fisher-price "Barbie" sing-a-long karaoke machine all the way up to a high-end Sony home theater system in the presence of snotty, self-indulgant audiophiles who argue the importance of 1" diameter speaker wire.
Mastering hasn't changed, only the devices used.
Enjoy
