Apr 5, 2003 - by dbfA column made out of messages on the Mailing-List of MacMusic!
|The history of Mastering|
Before the blessed era of the CD, we transferred our delicious 1/4 or 1/2 inch master tapes on vinyls. I remind you in passing, and I’m extremely moved, because that brings us back to the Middle Ages, that before carrying his tape to the "mastering project", the sound engineer at the time spent a few sleepless nights to clean his mix, to decide with the composer on the order of the pieces, all that was to be combined meticulously on a large reel with white markers from 3 to 4 seconds between each title. When by misfortune and because of the absence of Dolby, the master tape was a little too noisy, or polluted, the producer would yell while listening to the final editing with the headphones and force us to replace all the markers by printed tape with an equivalent noise to that of the mix in order to avoid the listener to hear an absolute silence between the pieces... What a time…
After it had been listened to a thousand times and finally accepted by the producer, the master tape finally took its way to the mastering studio to be transferred on the original mould: a metal sheet covered with a tiny layer of vinyl that a mischievous chisel was going to engrave as good as it could be done. The unfortunate ones as me who have gone through this difficult time are marked for the rest of their life. The engraver would make us a sample, that was called "flexible" but was not at all flexible in view of the fact that it was made of the same metal as its twin mould (I still have a hundred of these in my archives, this nightmare only stopped in 1991) and that was used as reference compared with the first press-sample.
For the record only I remind you that the vinyl was engraved in stereo by a side movement of the chisel for the informations on the sum (left+right) and vertically for the information on the difference (left-right). But, if the side movement was only limited by the duration of the musical program (broad furrows: reduced timing) the vertical movement was limited by the tiny thickness separating the vinyl layer from the scrap iron. THUS: it was impossible to put the kick drum on the left and the bass guitar on the right , impossible to put « off-phase » information at an average level, impossible to have a level of difference higher than the sum level, impossible, impossible .. a real nightmare. Fortunately, relentless and competent men knowing their machines better than the manufacturer did, managed to make miracles, and to make the reference vinyls that are listened still today with an extreme pleasure. But on the sole condition that the sound recording process and the mixing had been done with an exceptional quality care so that the result on the disc was to be correct. If not, the expert would throw us outside of his laboratory saying to us nicely, but firmly, to go to learn our engineer job... "Can’t engrave this my man, gotta remix it...": What a damn shame. What a total decline in front of the realizer who with an angry glare, would make us understand that we were fired, banned for life, and that it was for our own good to take the first plane for the Bengla Desh.
I do not exaggerate. Quite a lot of my colleagues lost their reputation on the path of the engraving studio . Others, which I’m part of, gained theirs through the magic sentence of the engraver "Well there, Champ, this mix, nothing to say, easy as pie to engrave it" These engravers loved us, because they could improve our work and do theirs correctly. They knew that the person in charge of the record company would not call them three days later to insult them because the press-sample was unbareable compared to the one they had in their office. And then one day arrived the CD. We were told « This is it, the nightmare is over. No more defects, no more stress ».
The master tape goes to the factory, no more humans needed, the CD goes out of the press and it is a clone of your master tape. »In fact the nightmare had just begun » . The engraver has put away his Neumann machines in the closet , put out his SONY cutting benches out of them boxes and the greatest fraud of the century of audio has finally started... I leave you to meditate on this beautiful history, I light up a Marlboro, I drink a coffee and I release a little band-width before the continuation of this new serial « Oh what happened to my beautiful sound ? »
|No more CD Burning , but PRE-MASTERING|
Approximately: you do your mixing in the studio, on an analog tape, a DAT or whatever suits you. You go to the mastering studio where someone puts the pieces in order with a complex numerical system. At the beginning it was the Sony system based on the PCM 1630 on Umatic cassettes, which is still very current.
If t’s an analog tape, it should be copied on a Umatic cassette, before assembling it on another Umatic cassette, which will be the final master sent to the factory. If it is Dat tape, it should also be copied on a Umatic to be able to be assembled. (We speak about the beginnings of the mastering).The levels between each pieces have to be compared and corrected or limited if necessary.
Once the editing is done, the Umatic has to be encoded to give it the PQ codes. The PQ codes are used to index the CD. It is them that make the CD player locate the beginning of each title. There is an index at the beginning of each title, and an index at the end of it ,even if the pieces are linked together. The ending index is used to know the exact timing of the piece. The CD player is able to position itself on a title because of these indexes. Without getting into boring technical details (especially for us who have to deal with them) the beginning and end indexes have "offsets" of a few milliseconds, allowing the player to position itself slightly before the piece in order not to not encroach the beginning of it when it starts.
The Red book specifications are very strict. Any master tape being sent to the factory which does not respect these standards scrupulously is pitilessly rejected. In order to be sure about the respect of these standards, the tape and the Umatic recorder undergo an ANALYSIS, which is printed on paper and which details all the necessary points: The number of the error corrections, the comparison between the engraved PQ codes and those described, the number of exact packages of recorded data etc... There is no laughing at all in the factories. The master tape arrives at the factory. It is put in another Umatic, another analysis is performed in order to be sure that the machine that reads the master tape does not generate more errors than the one that has recorded it.
If everything is OK, the Umatic tape is put in the rack of the LBR (Laser Beam Recorder) and a Glass Master is engraved. Then all is said and done, the robots take control and the CDs are piled up in plastic boxes, a press does not make less than 500 CD each time. No samples made by unit, your CD goes directly in stocks of the record company and the record store. The least error is reflected on the totality of the engraved CDs, without the least chance of a control.
All this history was before the virtual world did interfere. Nowadays the mastering on Umatic is almost as obsolete as the steam engine train , for the simple reason that since the advent of virtuality it is possible to engrave a Glass Master at double speed if using a computer. Double speed = 2 times more Glass Masters, 2 times more presses and thousand times more CD per day...
Today, a Glass Master can be done, depending on the factories, starting from a Umatic, a DDP tape (8mm tape Exabyte of datas containing an image disc of the CD, the current standard), a PMCD (master CD containing the same information as a Umatic) or of simple home made CD. In accordance with the informations above, it thus appears that to finish on a DDP which is now the standard reader installed in the LBR rack any other format arriving at the factory is transferred in DDP. WHICH MEANS: That there is indeed a human intervention after the phase of pre-mastering and that if this intervention is made without competence, it can purely and simply ruin the work of a few months. But we leave the subject, let’s keep that for another debate.
|Doing your own Pre-Mastering|
The third section of this fascinating article concerns more particularly what occurs today to the pre-mastering, and if it is really possible to do it by yourself, in order to send your own work directly to the factory.
As everyone knows it today, the CD is in 16 bits. In accordance with the processing done on the signal, if you have to compress, equalize or boost the master, the audio files used for the pre-mastering has to be done generally in 24 bits. I repeat, in general, because in the absolute, to copy back in 24 bits an analogic rotten cassette and reproducing the diaphragm of a 100 watts marshall speaker having made the world tour of Metallica will surely not change the face of the world, but... The 24 bits is standard in the mastering studios, where the audio files are handled with SADIE stations on PC or Sonic Solutions on Mac. The Umatic is in 16 bits that’s why it’s obsolete. Once the editing is done with the eq and various processing that has been carried out, another final master is made by reducing 24 bits to 16 bits thanks to complex Dithering programs like UV22 Apogee, or Superbit Mapping Sony.
An audio CD can be only be burnt in 16 bits, 44,1kHz.
When the CD appeared, in 1982, the standard indicated that the maximum level of the CD, that’s to say 0 dBfs, was 18 dB above the zero of the mixing desk, in order to have some margin (headroom) for the peaks. Over the years, and the increasingly systematic use of new limitings device and ultra-powerful compressors, the situation today is , in 1998, that CD have their margins reduced to 2 if not 1 dB, that means that these level at 16 dB higher than those in 1982. The war of the sound level is limitless (: -) ) and each mastering studio competes, as at the time of the 45 rpm, in order to produce a master even louder than the one of the competition. Even if that makes a good impression, there is nevertheless a problem, because the average CD reader having been built for a margin of 18 dB, has a lot of trouble to find ressources to produce 0 db of average level out of its cheap pre-amp. It even sometime happens that the level of the CD requires the reader to get more juice of its electricity supply than that it can provide, the result is : distortion and an horrible sound.
The fact of having a lot of level on a CD does not relate to the fact of being done in 24 bits, but of being limited and compressed in such a way that not having peaks anymore, you can raise up the volume by 3, 6 or even 9 dB without any problem. Useless to say that if the music was not thought accordingly, the final result can be artistically disappointing.
Very loud does not always mean good. Everyone today can make a master at his place, with a little care. A device like the TC Finalizer, well used, can give the same theoretical results as those used for a pro mastering. It is especially necessary to have some experience of these techniques, because any error is of course fatal. Because it is difficult, without a specialized installation, to be sure to be really to the standards of the factories.
|Speaking about the mix|
Be sure about the importance of the nuances you desire. In the case of a CD intended to be listened by a normal individual, not having a hi-fi system 2X1254 w installed in a concrete bunker perfectly isolated from sound pollution of the vicinity, it is very difficult to benefit from nuances higher than 6 dB without having endlessly to raise the volume to hear the ppp and to lower the sound when there are fff, etc.
If you are positively sure of yourself after checking that these nuances are REALISTIC, try to limit wisely the sharper instruments, those that make the peak-meter going towards the zero, even when you have the impression that there is no volume. In the list, ALL the percussions, certain piano notes, the consonants of voices (T... and K... too compressed)
If each instrument is carefully planed without harming its quality, the mix will already have a peak/volume ratio nearer to reality. If for a reason, an instrument releases a peak of 6 dB higher than the others, the mean level of the CD will be 6 dB below, by taking this single peak of the piece like maximum level. Too bad...
Once the mix is finished it must be carefully checked; on this matter, the mastering will consist in putting it at the standards of the external world. Which means : equalize the master tape, if needed, to cancel the defects of the listening test on which the mix was made. Needless to say that to carry out this operation, it is necessary to have a reference listening test on which everything is heard, the qualities as the defects, a reference proved reliable, which means that has already sent some samples in the nature successfully.
Then, you need to boost the electric level of the master tape, if possible, so that when the mastering has been carried out, the CD is comparable to the other productions on the planet, and that the only difference between your magic disc and that of the best CD of the year is only musical, and not technical.
To boost means raising the level, if the peaks allow it. If the master was made on a DAT not calibrated, it is possible to gain a lot of level. If boosting is needed, the background noise behind the music will also be raised. If the mix is flimsy about the "volume", the mix has to be compressed. Planing down the inaudible peaks doesn’t mean compressing or squashing the mix, it is just a way of raising the overall level upwards. The analog tape had the talent (it has always it, besides) to plane these peaks with softness.
The human ear hates the peaks, which are planed naturally by the tympanum. To leave them does not bring anything musical, only an obvious feeling of hearing strain after a few minutes of listening. A sharp sound is without thickness. It goes without saying that these processings reduce the signal-to-noise ratio of the mix. This is why the mastering established its standard in 24 bits, so as to have more margin downwards for this kind of processings.
But as long as the musical contents IS IMPROVED, the result is optimized, too bad for the noise and the distortion if the music is better. Only the result counts. When the work is done alone at home, the only advice that I can give is to compare what is done with a lot of discs that you are familiar with and that are close to the music on which you work.
In order not to be cheated by cheap converters , try to find a CD reader with a digital output. For ProTools, the least expensive solution that is close to the pro-mastering tools is the Waves Maximizer plug-in. Doing some experiment, you will rather easily find optimization settings of the level very convincing.
Translated from french by Dominique Zbiegiel, aka DOMZ.