Getting Songs Mastered

It has been a while since I had a song mastered. I've either been happy with the results of my master channel signal chain or I've sent off remixes where the mastering was going to be taken care of.

I'm not a fan of unattended mastering. It makes me uneasy that someone may impart heavy processing on something I've worked hard at. Many hours of work can easily be destroyed by a half hour mastering job performed by someone who doesn't share (or know, or care about) my vision. I like to take full responsibility for my sound - after all, it's my name attached to the product. If the song is clipping and distorted because the mastering engineer got a little eager with a limiter, it reflects poorly on me, not them.

Anyway, here are a few things to be considered when we get our songs mastered:

1) Attended session or not?

In these days of cheap internet mastering this is becoming more difficult. I much prefer to sit in the room when my song is being mastered, to talk to the engineer about the result I'm looking for. I take in a reference CD with several great sounding songs, in the ballpark of what I'm aiming for. The mastering engineers who have worked on my songs have all been happy to do attended sessions. I can understand that some may not be so keen - I'm not sure how I'd feel if I had someone looking over my shoulder while I was putting together a mix. It would depend on the person and the types of contributions they make during the process.

If attending the session isn't a practical option, provide as much detail as possible about the result you're looking for. Some engineers will ask for reference songs. Hopefully you'll have some ready to send. I assume most people compare their songs to others in the same style. It'd be very difficult to to gauge the mix otherwise.

2) How loud should it be?

If you have a particular position on dynamics, let the engineer know this. Let them know you want it as loud as possible, or that you don't want your song squashed to death. If you've heard other examples of their work in the same style, maybe you can trust that they'll push it hard, but not too hard. In any case, it can't hurt to let them know how hot you want your song to be.

3) Where does the mixing process end and the mastering process begin?

I often put a touch of EQ on the master channel of my song. I'm happy to take off the limiter when I send it off to get mastered somewhere unknown, but I'll often leave my EQ touches on there. They're part of my creative vision and there's no guarantee that the mastering engineer will share that vision. So we need to work out where we draw the line - which decisions are essential parts of the creative process for a given song, which ones we're happy to leave in someone else's hands. Perhaps our song fades out at the end and we're happy to leave the fade to the mastering engineer. It's up to each of us.

4) What bit rate and sample rate should the song be?

Mastering engineers generally are quite upfront about the format they prefer to work in. If not, a good rule of thunb is to use the same sample rate that the final song will be in. It doesn't hurt to double it (so 88.2 kHz for songs going to 44.1 kHz, or 96 kHz for songs going to 48 kHz), but going from everything that I've read, working at higher sample rates makes no audible difference when the song is converted down to the final sample rate. Bit rates are a different matter - here it does make an audible difference when converting down to the final bit rate. I provide 24 bit files when the mastering engineer will end up delivering a 16 bit master. If the final output were 24 bits, I'd send a 32 bit file (though I've never been in this situation).

5) It doesn't sound amazing after mastering. Why not?

The first thing we should consider is: did it sound amazing before mastering? If not, then the solution is to keep practicing our mixing skills until our songs sound the way we want them to before they're sent off to be mastered. In my early days of mixing I expected the mastering process to help make my songs sound better, that somehow I and this professional engineer would be a team in delivering a great sounding song. That isn't the case. Mastering will make our songs sound more correct, will push the gain up cleanly, but it won't fix the fact that we've chosen the wrong kick sample or that we've turned our pad up too loud in the mix.

If the song sounded great before mastering, then discuss the issues with the engineer. The engineers I've worked with were happy to rectify issues. I wouldn't push too hard though - if the second version is still nowhere near optimal, leave some time before going to that engineer again. For me, this gave me enough time to learn that all the issues with the finished song came down to my inferior mixing skills!


I hope this has been helpful for some of you. Bear in mind that not every artist has their songs mastered. Some of them like to be responsible for the process from start to finish. Once again, we all make our own decisions. I can see the benefit in running a song through another set of ears in an optimal listening environment, but I also totally embrace the mindset of complete responsibility!

Keep making great music!
Fabian

The Impossibility of Self-Mastering

For me, it's not a question of whether or not to master my own music. The way I see it, even if I wanted to, I literally can't.

In terms of mastering a single song, the mastering process is a chance for a second objective pair of ears, listening in a high end listening environment (some to a higher degree than others), to notice and correct mix imbalances and deficiencies. Hopefully you can see why, even though I'd like to, I can't master my own music. I would need four ears and two brains, and my second brain wouldn't be allowed to listen in during the mixing process.

I do, however, have audio processors on my master channel. Typically I'll have some compression, followed by EQ and finally a limiter. These are tools often used by mastering engineers, who may well use them in a similar way to myself. Just because it's a similar (or the same) process doesn't mean it's the same process from a second-set-of-ears perspective. When I use these tools, I'm still at the mixing stage. The mix just happens to sound the way it will sound when the song is released.

There is a lot of debate about mixing into a chain on the master bus. Many engineers leave their master channel empty and actively try to dissuade others from processing the master channel. I can totally understand this perspective, which I believe boils down to two main factors: firstly, taking responsibility for each individual sound and group in the mix. For example, the end result is generally much better if I fix problem frequency areas on individual sounds rather than subject the entire mix to drastic EQ processing. If the levels are slightly imbalanced, it's much better to adjust the individual faders than to try to correct the imbalance on the master channel. Secondly, depending on the compression applied to the master channel, mixing into it could well feel like mixing when drunk (I haven't actually tried that). Every new sound, every level adjustment, will change the overall balance. I used to mix into heavy compression when I started out mixing and it added immensely to my mix frustrations! I don't use heavy compression any more, but I occasionally hear other people's mixes which are treated to significant compression, and it's often quite eye-opening to hear the relative levels when the compressor is bypassed – they make no sense whatsoever!

Having established that there are very good reasons not to process the master channel, I'll happily admit that I process the master channel. The reason is pretty simple – I want to hear how the song will ultimately end up sounding, as early as possible. I just find it more inspiring to work on something which sounds more “finished”. As stated, my signal chain is typically compression – EQ – limiting.

The amount of compression I apply to the overall song is fairly minimal. I typically set the threshold very low (-40 dBs) and apply a very light ratio (1.10:1). The attack and release times are as fast as possible. This treatment means the song is always being compressed, with the overall dynamic range being reduced evenly rather than just the loudest parts being pushed down. There is no cutoff point where the compression kicks in and out – the loudest parts will be pushed down the most, the next loudest parts will be pushed down almost as much, all the way down to -40 dBs. The relative levels of the instruments will need to be spot on, with or without this treatment. In practice, this treatment will drive the average level up by a decibel or two.

The EQ I apply is (hopefully obviously) dependent on the song. Very occasionally I won't need to apply any EQ at all and that does feel satisfying. Having said that, I don't lose any sleep when I need to employ a few bands of processing. Whatever leads to the best end result is fine with me. More common EQ requirements are: a highpass filter to cut out excessive sub frequencies – under, say, 35 Hz. I can't hear these in my listening environment so I use a frequency analyzer which measures RMS levels to see whether the sub frequencies are over-represented. I prefer to apply this to my Kick/Bass group, but sometimes other groups also have low end coming through and it's easier to apply it to the overall mix. I'll often apply a broad boost centered around 3 to 6 kHz (it depends on the song), to bring out overall definition. The boost will only be a dB or two, and is applied when the relative high ends of the instruments/sounds are already balanced, that is, all the sounds benefit equally from the boost. Beyond these typical EQ treatments, it's wide open – whether small boosts or cuts anywhere in the frequency spectrum, it's entirely dependent on the mix. I don't take the decision to apply master channel EQ lightly. Where possible, I treat the individual instruments/groups, since that will almost always lead to a better result.

The limiting I apply is purely to bring the RMS level to a “competitive level”, so I can compare my song to the songs in my reference ballpark. I aim my level for the midpoint of this ballpark – usually I could drive the gain up further cleanly, but maximum level isn't my goal. I've heard a number of casualties of the loudness war and I'm happy to be well back from the edge!

At some point soon I'll address “real mastering”, in terms of getting music mastered by a mastering engineer. Not every song requires it – when I put songs up on the internet (or on www.InsideMixes.com) I don't get them mastered. If I had a bigger project, such as an album, or a song I knew would have a large audience, I'd be very likely to go to my local mastering engineer and sit with him while he masters my songs. For the most part though, I'm happy to take responsibility for the entire process.

Keep making great music!

Fabian