What is Mastering and what it is not.

How has it changed over the years.

My own mastering history.

When I started my career as a cutting engineer, way back in 1972, it was all simple. There was only one format: the vinyl disc.
http://paulvanderjonckheyd-mijnaudiocv.blogspot.be

Different procedures were used to distribute the single- and album masters to the different territories around the world.

- Sometimes master lacquers were cut in one location and send out to the different pressing plants.
- Other companies send positive metalwork (mother matrixes) for local production.
- Master-Copies, EQ'd as in the initial cutting room, were send for local cutting and pressing.
   - Most of the time however 1/1 copies of the original would be send round, sometimes with cutting instructions.

There were 2 schools of thought:
(1) Is the disc-mastering of the master lacquer part of the mechanical process of duplicating,
or (2) is it the final stage in the artistic process ?

So on the one hand there were cutting rooms cutting a 1/1 copy of the mastertape and there were others trying to polish the studio master with minute EQ and limiting to produce an even better result.

EQ was most of the time very subtle with 0.5 to 1.5 dB changes.
As a general rule, an album would be recorded and mixed in the same professional studio, by the same engineer and under the supervision of the same producer. Mastertapes were always send in the correct running order, with the needed silences between the tracks.

Crest factor, the difference between peak- and RMS-level was around 14 dB for pop- and rock productions and even bigger for ballads, jazz and classical.

Limiting was only used to control the occasional peak, with the gain reduction needle barely moving (0.5 to 1.5 dB max.)

All processing was performed “on the fly”, in realtime to avoid extra tapecopies. To adjust between tracks, a double set of stereo-equalizers was needed to switch-over between tracks.For example some 1.5 dB low roll-off could yield a gain of 2 dB in overall level.

Recording level was, and still is, an important factor to make the radioprogrammers choose a single for broadcast. We all know (?) that this is utter nonsense, radio is the great leveler as I will explain later, but there is no denying: First impression makes all the difference.

Until the early 80's, all changes to the sound of a track to polish it or to adapt for the next medium, vinyl or later compact cassette, (pre-mastering) were made at the mastering stage: cutting a masterlacquer or making the bin-master for cassette duplication (mastering). There was no stage between the mixing and cutting the master-lacquer (mastering).

Up until that time, the purpose of mastering was to be “as invisible as possible”, making minute changes just to adapt to a different medium and to have the “loudest cut possible” with a minimum of interference to the original studio-master. It used to be a very delicate balancing operation.A lot has changed with the advent of the digital recording and of the CD, later even more with home- and project studios.

Radio ready (?)

Radio is the great leveler. The argument that your track will not be as loud on the radio as others just does not hold. All radio-stations use 5-band compressors (Orban Optimod) that will make sure all songs to be broadcasted equally loud; radio is the great leveler. All will depend on the way your song has been recorded and its frequency content. By the very nature of the radio compressor a hyper-compressed song might sound less loud on the radio than a well balanced, open and moderately compressed one. A hyper-compressed track might even start to distort. Tests have shown that different degrees of compression on the same title makes hardly any difference to the loudness on the radio.

At top, five mastered cuts of the same music, with increasing loudness and visual density. At bottom, the same cuts passed through the Orban radio processor.(Bob Katz: Mastering Audio - the art and the science - Radio Ready: The Truth) 

Over and over I have been hearing this sentence: My song is not as loud on the radio as the others! This is how I explain this phenomenon:
Someone is listening to the radio. Suddenly his song comes along. A natural reaction is to turn the level up. With the next song the level is put back. The resulting conclusion being: My song was not as loud as the others. I call this: “Negative Psychological Feedback”.
It is all in-between the ears!!

Analogue vs Digital distortion.
Analogue distortion is usually described as “warm” versus the “cold” or “harsch” digital distortion.This is because analogue distortion is, for the most part harmonic, with 2nd order distortion producing a pleasing effect to the ear, while higher order harmonics rapidly being rolled-off by nature of the medium.
When overdriving an analogue tape the harmonic distortion comes in gently, producing some gentle compression in the process.
Digital sound on the other hand sounds clean until it clips the convertor or processor, producing very nasty unharmonic distortion instantly.Early digital designs were not all that wonderful either, giving digital a bad start.
http://www.soundonsound.com/techniques/analogue-warmth

The new digital age.

With the arrival of the CD in 1982, the glassmastering became an integral part of the production process.
The glassmaster is a 1/1 copy of the PCM1610/1630 U-matic master.


The Sony PCM1610/1630 with DMR U-matic machines and DAE-1100 digital audio editor were very expensive and very slow to use. Very few studios could afford to make their own U-matic masters.A new stage or layer in the music-production chain was born: the (pre-)mastering studio.
At first these (pre-)mastering engineers would just take separate tracks from the Studio U-matics or copy other formats to U-matic, stringing them together to make an album-master U-matic, put the PQ-codes onto analog track 2 on the U-matic cassette, make an error-control list and send to the glassmastering.



The first affordable digital recorder on the market in 1982 was the Sony PCM F1, recording onto a Betamax VCR. Many studios were using it but could not even put an album in the correct running order.
So when HHB released their CLUE digital editor in 1983, I bought it to string the tracks together. This was before the days of computer editing as we know it. CLUE used 2 5" floppy disks and a dedicated computer to control the two C9 Betamax units for editing.
All mastering processing was still done completely analogue and an AMS digital delay at 50 kHz sampling rate was used for groove control of the cutting lathe.

In 1987 Sony launched the first DAT-machine, the DTS100ES, with several other manafactureres following.
This (consumer)format took the industry by storm, every studio adopted it overnight.

Luckily the same years saw the birth of the first computer hard disk recorders, later know as DAW'sAudio Design's SoundMaestro finally allowed me to do the most difficult musical editing.
With a big investment in SoundMaestro, a PCM1630 processor, DMR2000 U-matic, DTA2000 digital tape analyzer and a Philips-Studer PQ-editor, I was now able to supply industry standard U-matic CD-masters. Most processing still being done in the analogue domain.

(Pre-)Mastering for CD wasn't all that different from vinyl (pre-)mastering, still making subtle EQ-changes and leveling between tracks.

Digital Overs and Inter-sample Overs.

PCM1600 and -1610 level-meter with his calibration mark for a VU studio level of +4 dBV gave the processor 20 dB headroom, enough headroom for peaks to pass without clipping. When using the PCM1610 to digitise analogue tapes (AAD), it became clear that these recordings were about 6 dB louder than all-digital productions (DDD). Design engineers are always a bit naive in thinking everybody will stick to their rules. It is the very nature of creative people to explore boundaries and work outside the rules or break them. Now everybody started to push for higher levels. But, where in analogue distortion comes in slowly, crossing the digital 0dBFS-line means immediate non-harmonic. 

When Sony released their subsequent model, the PCM1630,they changed their metering to 0 dbFS, with a studio-level marker that hardly anyone used, trying to hit 0 dBFS as much as possible, with disastrous effects up to the present day. Hitting the 0 dBFS on the 1630's meter, or indeed any non-oversampled peak meter gives rise to inter-sample distortion, as explained here:
https://benchmarkmedia.com/blogs/application_notes/intersample-overs-in-cd-recordings?utm_source=Benchmark%27s+Application+Notes&utm_campaign=490edc434c-EMAIL_DAC3+DX_2017_02_10&utm_medium=email&utm_term=0_7c8c792ee5-490edc434c-161966933

Now that it had become clear that, because of the natural tape-compression of analogue tape, songs recorded analogue and later converted to digital (AAD) sounded +/- 6 dB louder than all-digital productions (DDD), digital productions were now being more heavily compressed. The Level-war had started.

The next important step happened in 1989, when German Harmonia Mundi in collaboration with Daniel Weiss of Switzerland developed the first Digital Audio Processor, the bw102. Sony also released a DSP, but it was a closed box, where the bw102 was completely modular.

The bw102 allowed very high quality digital processing: EQ, compression, limiting, expansion, gating, de-essing, reverb and level-changes, dynamically programmable to SMPTE timecode. It had upsampling/downsampling modules to make processing at double sampling rate possible.

It was a dream come true. Although very expensive, I put an order in the moment I saw it at the 1990 AES convention.

It was also a dream for vinyl disc-cutting. It was now possible to dynamically program the processing of a whole album, going back and forth between the different tracks of an album, constantly comparing, tweeking and changing between tracks.

Soon other manufacturers followed, marketing hardware digital processors. Digital lim/comp had the possibility of previewing the signal, reducing the amount of artefacts and distortion inherent in severe reduction in analogue lim/comp.

As a result the level in productions increased and the crestfactor or PLR (peak-to-loudness ratio) went down again.

In the following years more and more software processing (plug-ins) were hitting the market, much cheaper and affordable compared to their hardware counterparts.

Complete sets of “Mastering Suites” were now offered, with as the only sales pitch: “To make Your Music Louder than the Competition”. (Some of the “presets” offered in these made my stomach turn.)

With the increase in Loudness, the distortion caused by deliberate clipping, overloads an inter-sample-overs kept on rising.

Now, a lot of masters I received for vinyl were becoming more and more difficult to cut.

Their Peak-to-Loudness Ratio (PLR) became so low, the level of a side being high all the time, using a lot of space on the disc. I was forced to drop the overall peak-level of a side.

Asking for non-mastered studio versions allowed me to cut better sounding vinyl.

In 2001, when Apple introduced their iPod and in 2003 iTunes, we pleaded with Apple to make “Sound Check” ON by Default. To no avail.

It soon became clear why the Fraunhofer-Gesellschaft advocated dropping levels before converting to any data-reduced format.

Already present overs and inter-sample-overs, theoretically +3dB but in practise around +1.5 dB, are augmented by several dB caused by the overshoots in the many filters present in the algorithms of the data-reduction formats, increasing with higher reduction rates (lower kB/sec).

Again, designers were being a bit naïve, nobody was dropping their levels. The sound of data-reduced formats became much worse than it really had to be. Some Mp3 and AAC formatted tracks became nothing more than headache inducing racket.

My tests on some iTunes top-10 titles showed overs to +6dB.

Meanwhile the opposition against this crazy loudness war was getting itself organized with initiatives like TurnMeUp.org, Dynamic Range Day and Pleasurizemusic.com.

With the introduction of the TT-meter in 2009, we finally had a meter to properly check the Dynamic Range.

Using this meter, it became clear that, as a general rule, sound is really suffering at a DR below 8, DR6 being the borderline between an acceptable and an overcompressed track.

More and more I realized I could and would no longer compete in this crazy war. During the 40 years of my career in professional audio I had invested several 100.000 euros and countless hours, days and nights of work, always being at the forefront trying to give my clients the best possible quality. Now I was asked to ruin perfectly good recordings.

Mastering is the art of making the final Master sound “Better”, not necessarily “Louder”! But “Louder” is what most clients expect.

Why mix your album for days, adjusting every single track of a song to within 0,5 dB to make your mix just right and then ask the mastering engineer to completely alter your entire album in a couple of hours by slamming a compressor-limiter over it, compressing it to death. Compromising your sound, all this just for the sake of being louder than the competition!??
I never understood why mix-engineers were not up in arms against the mastering rooms for ruining their work. It is my sincere conviction that a lot of albums would have sounded a lot better without passing through a “Mastering Room”.

That's when I decided to make DR8 my lower limit, only allowing DR6 for Dance or Heavy Rock tracks. I fully realized that it would cost me clients. For several years now I hated what I was asked to do. I lost all joy in the work I loved so much before. I just could no longer be part of this crazy level-war. SoundCheck and ReplayGain were already around for years, switch them on or use your level controls please.

Luckily I could still have my professional pride and joy in providing quality master-lacquers for vinyl.

A most encouraging thing happened in 2011 when Apple announced a higher iTunes quality: Mastered For iTunes (MFiT). We finally had the tools and a set of rules to make higher quality iTunes tracks. (And why not better CD masters using the same tools and rules.)

An extra advantage is that Apple now has, in the case of MFiT, high resolution masters on file they can use later whenever the iTunes format will be improved again.

I immediately applied to become an “Apple certified Mastered for iTunes supplier”.

The beginning of the end of the loudness war was finally announced in 2010 with EBU R128 and 2011-2012 with the adoption of ITU BS.1770-2/3 laying down the rules for Loudness-Normalization.

Most streaming services have now adopted it: Spotify (-11 LUFS), Apple (-16 LUFS), YouTube (-13LUFS), Tidal (-14 LUFS), let's hope for a common reference level soon. Hopefully Soundcloud will follow.

Being happily retired now, I hear that the loudness envy and loudness fear are still very present. It was, and still is, a very contagious disease, an epidemic, that will be very difficult to eradicate. It will require the re-education of a complete generation born into this level-war.

I' am hopeful that we will get dynamics in our music again.

Why do you need (Pre-)Mastering

MASTERING is the moment for putting the final polish onto an album or single. The last opportunity to make sure that everything fits together seamlessly and that the sound does not vary from track to track.

“The secret of the successful mastering is all in the blend. You need a large measure of skilful engineer, a subtle blend of traditional and state-of-the-art equipment and a sizeable dash of well balanced acoustics.” (Bob Katz)

Usually each track is mixed in isolation. Rarely do you have the luxury to compare tracks as you mix. Some are mixed at 2 o'clock in the morning, when your ears are fatigued, and some at 12 noon, when your ears are fresh, resulting in differences between your tracks.


Many projects are mixed in less than ideal circumstances, inadequate monitors, uncontrolled acoustics. 
When the tracks are assembled into the correct running order, you may probably find that they do not sound like an album: Some tracks will jump out, others will be to weak, some may be to bright, others to weak in the bass.

The 2 most important tools of the mastering engineer are flat monitors with lots of headroom in an acoustically well balanced room and a trained set of ears.

I never liked “the average living room concept”, there is no average living room. Monitors should be flat, neutral, undistorted, in a room that works with and not against them. You have to hear what is really going on in the recording. Many clients have noticed things in their recordings on my monitors they did not hear before.

In my room, tracks that were good, sounded really good. Tracks that were bad, sounded really bad. Sean Davies nicknamed my monitors “Pravda”.

It is my experience that tracks that sound really good on a big, flat system, tend to sound good on a variety of systems. After installing these monitors/room, I never felt the need to check on mid- or close monitors, I very rarely did. My listening room is what I mis the most now that I have retired.

This is what mastering is about, polishing the sound, making al the facets shine. Not grinding it to one-dimensional flat surface.

Mastering is all about detail, subtle changes. To put it bluntly, if EQ changes of more than 3 dB seem to be necessary to solve a problem, it would be better to go back into the studio for a remix.

Pre-mastering should not be performed in the same studio as the recording or with the same engineer who recorded the work. It robs you of the advantage of a second opinion, a fresh approach and of (usually) much better listening conditions.
Take advantage of the trained ear of the mastering engineer, a lot of albums of different styles have passed through his mastering room over many years. He has the expertise and the specialized gear that is not available in most studios, he is the perfect person to make the final artistic and technical decision with you on your precious project .

What happens in the Pre-mastering room ?

First we will load your tracks into our mastering DAW in the correct running order.
Then we will take plenty of time to listen to your tracks very carefully and decide wether some or all of the tracks need any treatment, analog or digital, always with respect for your music.
Sometimes we will do nothing, good things should not be touched - good is good, sometimes we will advice you that it would be better to go back to your studio and remix.
We will equalize the songs to make them brighter or darker, tweak the stereo width or depth, adjust levels between songs, put in the correct pauses between the songs or cross fade them where you prefer, fine-tune the fade-ins and fade-outs.
We will only compress or limit where necessary, trying to keep as much of the dynamics and life in your music as possible.
We will polish your tracks and make your songs into an album.
After PQ coding we will burn you a CDR, so you can listen to it on your own system, and make the final changes before making a master in DDP format on CD-ROM for the pressing plant.

How long does it take ?

Pre-mastering a full CD (+/- 65 min) will take between 3 and 10 hours.

Different formats can be used: analog: 1/4 or 1/2 inch IEC, NAB or AES with or without Dolby A, Dolby SR or DBX II.

Digital: DAT, PCM1630 U-matic, SONY F1 Beta or VHS, CD-R., CD-Rom with your soundfiles in 24 or 32 bit - 44.1, 48, 88.2 or 96kHz, wav, aiff, SDII, stereo interleaved.

Always check with the mastering house which of the formats are accepted.

Never send quicktime- or mp3- files.

Masters for CD should be in DDP format on DVD-ROM or via network to the factory server.

Masters for iTunes should be in 16 bit, 44.1Khz, wav-format. (Apple also accepts aiff- and FLAC-files, but most aggregators' websites only recognize the wav-format)

Masters for MFiT should be at least 24 bit, 44.1, 48, 88.2 or 96 kHz wav-files.

Masters for vinyl should be in 2 big, 32 fp or 24 bit wav- or aif-files, 1 for A- and 1 for B-side, with all tracks in the correct running order with the necessary pauses. A track listing with timing of the scrolls should be included.