Beginning of the week I had a very interesting chat with one of my clients about a problem that creeps up often these days. He had just released a track on Soundcloud as a free download that I had mixed for him quite a few months ago.
As pure analog formats have pretty much ceased to exist, and even albums have been superseded by the ubiquitous single, so has the job of the ME shifted, from a technical, preparatory process to a mostly artistic one.
Today the ME makes sure that the quality and translatability of the audio within the context of musical expression is maximized.
They offer that final pair of unspoiled ears to judge, and possibly correct, anything that may have gone unchecked, or can still be improved. Mastering is absolutely critical to release a mature product, which is the main reason why a good mastering engineer is essential to the process in my opinion.
This concept has wide ranging consequences on what we, as producers and mixers, can, and must expect from a good mastering job. A famous sound engineer once said something along the following lines (it may have been George Massenburg):
“I pay the mastering engineer a lot of money to tell me that the mix is perfect, and nothing needs to be changed”.
Or, said differently, in an ideal scenario, the mastering engineer does absolutely nothing to the mix, because it is perfect. Ultimately, this is our goal, and what we should strive for every time we mix a song.
However, in the real world, our mixes are never absolutely perfect, and usually the ME can improve on it, even if ever so slightly. More importantly, a good ME will be able to recognize the things that are good about the mix, and leave them untouched.
So then the mix becomes the benchmark by which to judge the quality of the mastering job. The better the mix, the fewer changes we should see happen in mastering.
And that is how I define a successful mastering job:
Of course in practice, what does that actually mean? How do we tell, if the master sounds and feels better than the mix?
At this point I need to talk about loudness. For one it will be the most obvious difference between the master and the mix, but it will also be the main hindrance keeping us from comparing the two. And, as I’ve said previously, the loudness of the master is actually fairly unimportant in the grand scheme of things. Of course, it is still something we need to check, as plenty can go wrong in the process of making a track loud.
Listen through the entire master. If it still feels as good as the mix, and all the right emotions are triggered at the same moments, we are off to a good start.
Listen through the master while switching back to the mix when it spikes our interest. Here is what we are paying attention to:
Well, if everything checks out, and we are generally happy with the master, write a nice email to the ME, thanking him for the great job they did, and pay them accordingly. If you are a DJ, and have the time and opportunity, play the master in the club to see how it compares with the other tracks you play. It can also be very revealing to hear how the master changed compared to the mix, if you got a chance to play that one at an earlier date.
If we are not happy with the mastering job, but it only involves a few tweaks, write a nice email to the ME, telling them about what you are hearing, and why you liked how it was in your mix. Then ask them if they could do another pass, keeping those things intact, while applying the other positive changes they made. By the second pass, at the latest by the third pass, they should have gotten it right. Thank them for their help, and pay them accordingly.
Every once in a while, you will come across a mastering engineer, who seems to mess everything up. The mix has been changed in ways, that were not even remotely necessary, and the master completely fails the “sounds and feels at least as good as the mix” test. In this case, write an email to the ME, kindly thank them for their effort, then pack your things and run. Nothing is worse than a mastering engineer, who’s job description, let’s remind ourselves, can be summarized as quality control, who is obviously not capable of discerning between good and bad aspects of the mix. You want to stay as far away from them as possible.
Every time I prepare a piece of work for export I go through the same process. I call it „fake mastering“.
The reason I call it fake, is that it can hardly be described as mastering at all. In fact I could summarize it in two easy steps:
However there is a process to it, and reasons why I do what I do. I’d like to share them with you. In fact, it’s what I recommend you do in all cases but when handing it over to a professional mastering engineer. And in that case you are well prepared, since all you need to do is disable (or not enable rather) one single plugin.
I’ll go through these steps whenever I know that the export won’t be processed any further. The export is intended purely for listening purposes. These may be that I want to double check the mix at home or on my MP3 player, that I want to upload the track to the net for demoing purposes, or for the mix to be checked by my clients. In any case, it’s the last step in the processing chain.
The idea behind the processing then is pretty simple. I want the music to be roughly on par with the volume of professionally mastered tracks. However I don’t actually want the process to change the sound of my mix in any way. That way I can directly compare how my mix stacks up against other tracks without me having to to constantly adjust the volume. I can judge the quality of my mixing work, because I know that non of what I hear may be due to sloppy master bus processing.
In effect, the result is simply a headroom starved version of the mix.
The baseline for this to work is of course a solid mix with plenty of headroom on the master bus. You are working with plenty of headroom aren’t you? I may have used gentle compression or EQ on the master bus during the mixing process. But non of it was for the purpose of increasing the technical level of the material.
I then insert a simple peak limiter on the master bus. My limiter of choice is the entirely free Limiter No6 by vladg/sound. It offers a whole range of functionality, but I really only ever use two modules. The „Peak Limiter“ and „Protection“. I keep the other modules disabled.
On the peak limiter, I’ve found the following settings to suite the kind of music I work on:
On the protection circuit, I set:
The „ISP Precise“ setting will keep any Inter Sample Peaks from surpassing the set „Ceiling”. This is potentially important if the exported file is converted to MP3 later on, as some MP3 encoders apparently introduce distortion if Inter Sample Peaks occur. I have to admit though, that this is purely a precaution, as I have never actually witnessed this happening. I have all these setting stored in a preset for quick access.
So now comes the important part. I set a loop around the loudest section of my mix and press play. On the peak limiter, I’ll then turn down the threshold until I see the needle in the VU meter twitching. I tend not to go beyond 1 or 2 dB of gain reduction. I also pay close attention, listening out for any obvious changes through the squashing of the transients. Ideally I don’t want the limiting to be noticeable at all.
The beauty of this approach, in comparison to most other limiters, is that the volume of the music remains the same. This makes it extremely easy to hear the effects of the limiting process.
I then turn to the protection circuit. On the „Output“ knob, I add back the same value that I took away with the „Threshold” knob on the peak limiter. This effectively removes the remaining headroom. So if I set the threshold to -12dB, I’ll set the output to +12dB. And that’s it! It’s that simple.
We now have the mix playing at its highest possible technical level. We have carefully reduced the peaks so that the volume should only slightly be below that of professionally mastered material. We also know that what we are listening to is still the mix, just like when we were working on it, since we made sure that the peak limiting process doesn’t change the sound.
The best thing about all of this? When we go back to work on the mix, all we have to do is disable the plugin (and you better make sure that you do!). If we decide to export the track again, for another round of feedback listening, we simply enable the plugin and we are good to go. And finally, if we want to hand the track off to a professional mastering engineer, all we have to do is leave the plugin disabled when we export. The mastering engineer will thank you for the extra headroom.
A final tip: As you can imagine I can’t recommend the use of much other processing on the master bus for fake mastering purposes. Tools like multiband compressors or stereo wideners are incredibly powerful tools when used the right way. But in most cases they’re more trouble than they’re worth. More importantly they actually defeat the purpose of what we are trying to achieve in the first place. Which is to not change the sound of the mix at all.
Developing a functional workflow is THE most underrated part of the music production process in my opinion.
A good workflow gives us the frame work to let our creativity run wild, while allowing us to move confidently forward by making decision that build on one another, and not questioning their validity unnecessarily later down the line.
As such, a good work flow also means knowing when to make which decisions. For example, because mixing builds on the process of arrangement, a piece of music should ideally be arranged, before it is mixed.
In business, “time is money” as they say, which I would rephrase as “time is creative capital”, in the creative process. The more time we spend questioning decisions, the more subjective we become, and the more subjective we become, the less we will be sure to have made the right decision. A good work flow allows us to work fast, maintaining a quasi objective state of mind as long as possible.
Furthermore, a well placed decision early on gives direction to the process, thus making all follow up decisions much easier to make. As a consequence we work faster, are more confident about our decisions, and we will make better music.
On the other hand it also means that, if we are indeed unhappy with a decision later down the line, figuring out when this decision was made, and why we are not happy with the result, will be very clear, making it much more straight forward to avoid the same unproductive decision in the future.
I would even go as far as saying that a good workflow is more important than using certain tools. If a tool makes the process more difficult, no matter how expensive it is and how good it sounds, we should get rid of it.
Of course, this doesn’t mean that every decision we make needs to be set in stone. If we get a burst of inspiration about the arrangement after the track is already mixed, the idea needs to be tried out. It is simply important to keep in mind that this might require the track to be mixed again, as the presentation of the material may need to be portrayed differently.
So let’s look at a few examples, why they are good to adopt, and when they could feature in the creative process.
Prep work: getting rid of distractions for a relaxed state of mind, allowing uninterrupted creative flow:
There are hundreds of these workflow tricks out there. Depending on which part of the process you focus on and what your strengths and weaknesses are, some may be more useful than others. Sonicscoop has a great article on some more work flow tricks, that my peers use.
Ultimately though, the more of these we discover for ourselves and the better they are integrated into a full process, the faster and more effectively we will work and the better our music will sound and feel. I guarantee it.
A spectrum analyzer is a bit of a double-edged sword.
While learning to use them you might discover certain “truths”, only to realize later that the same “truth” just ruined your mix. I’ve certainly had my ups and downs with them, but I’ve come to appreciate what they can do, and when it is better to simply ignore them.
First, let’s set it up right. I find a window size of 8k samples suitable. This gives adequate resolution in the lower frequencies. I then set the AVG time knob to around 200ms which allows me to comfortably follow the movement of the display. I also set the metering to “k-20” as this represents the levels I mix at. Finally I add a second spectrum that holds the peaks behind the moving spectrum. It can be reset by clicking the display.
Frequency spectrum analyzers give a very technical representation of a sound, and, much simplified, all they show is the volume level of a particular frequency at any given moment in time. As such I’ve found their functionality to also be technical, rather than musical.
For one, they are great at showing you how low in frequency a sound actually goes. In many cases sounds will have energy in the low octaves that simply isn’t required. In fact, retaining this low frequency rumble will eat up space in our low end that our kick and bass could occupy, thereby reducing their punch and definition.
Snares are often candidates with unnecessary rumble. Solo your snare and look to see if the spectrum analyzer shows any movement below about 150Hz. Chances are you won’t need that energy. Similarly, low bass notes often extend way below 50Hz, especially from synthesizers. Since most sound systems can’t reproduce these frequencies anyway, it’s pretty safe to remove them.
Spectrum analyzers are also great for spotting ugly resonances in a sound. Where these come from will most often remain a mystery, yet they are easily visible as peaks on the display that are substantially higher than the peaks around it. Reducing or completely removing them often triggers a strong “ahhaaa” effect, making the sound more open, comfortable and simply less annoying. Untamed resonances can wreak havoc on all sorts of sound processors and the oh so important balance question.
The final use I have for the spectrum analyzer is much more a musical one than the first two and thus needs to be exercised with caution. That is re-balancing individual harmonics of a sound. While the goal here is completely dependent on the particular scenario as well as your taste and I therefore I can’t recommend what you should do, it is quite possible to use the spectrum analyzer to see how the individual harmonics in a sound are balanced against each other. Sometimes an instrument somehow feels uncomfortable or is difficult to mix because a particular harmonic is to loud, or the balance of harmonics is otherwise unfavorable. Dipping or boosting certain harmonics with an EQ can then be used to great effect.
The one point I’d be very careful with is judging the quality of your mix by its overall frequency balance in the spectrum analyzer. A great mix is one that supports the emotional statement of a song in the most effective manner. For one, this says nothing about the actual sound quality of the mix, but more importantly it certainly doesn’t say anything about its frequency balance. Everything is allowed and nothing is forbidden, as long as it triggers the right emotions. It is sometimes mentioned that you should aim for a “flat” frequency balance but, while some great mixes do in fact exhibit this behavior, it is by no means a hard-and-fast rule. In any case, mixing such that the spectrum analyzer shows a “flat” response will most likely not be the right frequency balance for your song.
Mixing music is like mixing cocktails.
When you’re mixing a drink, there is a limit of how much you can put in the glass. Once you’ve filled it up to the brim, and you find that you haven’t added enough vodka, there is no way to add more without spilling the drink. (OK, of course you could just take a sip, and then add more. But for the sake of argument, let’s say you’re the bar keeper, and you can’t really experiment with quantities in front your client.)
The same principal holds when mixing music. Actually it holds true for the entire creative process.
Let me explain why.[someimages img=”10-cocktail.jpg”]
In music, everything is relative. In the same way that you can’t dictate an absolute volume for your music, since the listener will simply adjust the volume to their liking on the playback device, there is no absolute front to your mix, no absolute left or right, no absolute punchy kick drum, and no absolute huge drop.
In all of these examples, if taken to their extremes, their ‘absoluteness’ will not hold up. This is because of the simple and important fact that there is no point of reference.
If all our elements are all the way at the front of the mix, we will not perceive them as being in the front. If everything comes from the left speaker, the listener can simply turn their head to face that speaker and now, for them, that speaker will be front and center. If we hear a punchy kick drum, we can’t tell how punchy it is, until we compare it with a different kick drum.[head]The point is, in music, just like when mixing cocktails, everything is relative.[/head] If we want a vocal that’s all the way at the front of the mix, we have to give it a point of reference that is located far, far away.
If we want a huge drop, we better make sure that the part leading up to it is quite in comparison.
If we want a track to sound extremely wide, we need to make sure that the elements in the middle are narrow in comparison.
And because there is natural limit of how far we can go (the brim of the glass), the balance between the two contrasting elements is of utmost importance.
Too much vodka is too much vodka. Getting the mixture just right, while staying within the confines of the glass, is what makes a fantastic drink. It tastes good, and gives you a good buzz.
So how do we know that we’ve added too much vodka? Or rather, when do we know that our kick drum is punchy enough, or has too much punch for that matter?
That’s where referencing other music comes in. A good reference track is a track we know well, which we’ve heard on many different speaker systems, and which we know translates in a certain way. It can tell us, for example, if our punchy kick is in the ball park, or if we’d like to make it slightly punchier. It can tell us where our upper, or lower limit lies. It can also tell us if our mix has enough definition, has too soft, or perhaps even too hard transients, or if our master bus processing is choking the entire mix. (Although I recommend staying away from master bus processing all together anyway)
So the next time we find our mix is lacking something, let’s think of mixing cocktails, load up a good reference track, figure out if our glass is already filled to the brim, and then decide if we can add, or if we should be taking away.
In music, everything is relative. It is through contrast that we can express strong emotion.
If you are DJ these days, chances are you are a producer as well.
As we all know it’s almost impossible to be one without the other in today’s electronic music business (5 Reasons Why Producers Make More Successful DJs). However, if you are just starting out, you might feel very self conscious about your productions, because you feel that they don’t hold up in comparison to the tracks you like to play.
The thing is, for the novice producer and the industry veteran alike, playing your tracks in the club is just about the most important tool to get feedback on your production.
It’s just so easy to get lost in a production and consequently loose any objective way of evaluating the actual music, its emotional impact and functionality. We easily end up focusing on some intrinsic detail, while completely blending out the big picture. But it is the big picture we should really be focusing on! The best way to stay objective, and to hear the music, and not the sound, is to listen to the track outside of the production environment. A DJ set at a club offers the perfect opportunity for this. When playing a track embedded in the flow of a DJ mix, our brains are automatically conditioned to hear the track in the context of other music and a state of emotional connection, rather than technical quality. And it is in this state of mind that functional and dysfunctional aspects of a production become blatantly obvious. Questions like: “Is a particular part too long or too short?” “Does the break happen when it needs to?” “Is the emotional energy and impact delivered on the drop and the following groove?” seem to almost answer themselves.
Together with the feedback from the crowd, this very focused type of emotionally embedding your own track within other tracks has a very distinct way of uncovering strengths and weaknesses in the music.
Of course, how, under the unhinged conditions in the DJ booth, you then make sure to capture that feedback, is a whole other matter. I imagine a small text block and a pen will come in very handy.
The second main reason to play your own tracks is that fans want to hear the productions of the artist they admire. Because of today’s strong connection between DJing and production, gigs have long evolved into more of a concert environment. DJs are known for their particular sound, and fans want to experience that sound in a club atmosphere.
So the next time you are struggling nailing down a track, and have a DJ gig lined up, take the opportunity, play it, and gain some serious insights into your production.
It’s a common misunderstanding that the levels in the DAW determine how loud your record will ultimately be.
In fact, how loud we perceive the music, called loudness, and the technical levels of the material, have less in common than you might think. As it turns out, and I am sure you have experienced this yourself, two records played back at the same technical level can differ in loudness quite a bit.[someimages img=”13-Mobile.jpeg”]
The simple explanation is that loudness is created in our brains rather than through any actual technical characteristics of the material. Loudness is, in effect, how loud we perceive the music rather than how technically loud the material is. It is determined through the frequency content of the material, where the energy sits in relation to our brain’s “frequency curve”, the balance of the mix, the dynamics of the material and the development of these throughout the song. Loudness is, in many ways, also a measure of how good a mix is. A mix with high loudness tends to present the material in a very effective way, which is why it is so enjoyable and desirable.
The big pitfall in this regard is that it is generally assumed that high loudness can be achieved through high technical levels. We demand absurdly high levels in mastering, hoping to compete with the songs we are referencing, when in reality what we want is higher loudness. In the process, we introduce undesirable distortion, compromise both micro and macro dynamics, negatively impact depth and width, and generally speaking, reduce the enjoyability of the music.
Now when we get to the real world, the listener actually seeks a consistent experience of loudness, such that a sequence of songs will appear to them to be played back at the same volume. They either compensate for any perceived difference in volume between songs through the volume button on their playback device, or playback volume is adjusted automatically by the device itself. In effect, any differences in technical level, or loudness, between songs become almost irrelevant. What does remain however, is the compromise in audio quality that occurred when we tried to to push the levels during the mastering process. In a way, what we tried to achieve has turned up on its head. Instead of gaining high loudness, we got stuck with bad audio quality. The track that was carefully mixed for high loudness, however, and conservatively mastered to maintain all the finesse in the music, now sounds clean, full, deep, impact-full, in one word: enjoyable.
So what can we learn from this?
First and foremost, there is no point in pushing the technical levels for loudness sake during any part of the production process. In almost any scenario this will introduce compromises that do not benefit the listening experience in any way. Watch Ian Sheperd’s video on the automatic mastering service “Landr” for a great example of what compromised audio sounds like.
Second, we should strive for high loudness, not high levels, when our aim is to make high impact music. Again, Ian Shepard gives a simple example on how to manipulate loudness, rather than levels.
Third, as the final loudness of the music is not linked to the technical levels, we can start thinking about how to use the levels in our DAW constructively. This is what gain staging is all about.
There is one final very interesting aspect to all of this. And that is that as listeners, we are actually intuitively very good at judging if we achieved high loudness or not. If we are listening to “healthy” loudness, we tend to reach for the volume knob with the desire to turn it up. In fact, you’ll want to turn it louder and louder. And if the material was treated right, your heart will thank you for it with a jump of joy. If however, you listen to a song and wish to turn it up, but always end up thinking: “That was painful, let’s turn it down again”, you are most likely listening to a very compromised kind of loudness. So next time you are enjoying a good listening session, pay attention to what you do with the volume knob. It might just tell you more about the music than you think.