WHAT ARE MULTIBAND DYNAMICS?
This category of processing is dangerous to your audio when misused. It is a very, very powerful type of processing.
You can target just the frequency areas you want by using this tool, and it has all kinds of uses in recording projects.
It’s fairly new to the scene, and is designed to increase the transparency of the dynamics processes applied to audio.
It works by splitting the input signal into several discrete frequency bands for independent dynamics processing.
It turns out that this deceptively simple idea is incredibly powerful, which is why it is so easy to ruin your audio with it.
On the other hand, it can do amazing things about some of record production’s thorniest technical problems, as well as creating some awesome sounding effects.
Mastering is another place that multiband compression techniques come in handy. It should be obvious why. When you master music, you are mastering full-range audio, so using dynamics processes with any kind of precision almost requires it to be capable of multiband processing.
Until multiband compression came along, the engineer would have to create several duplicated tracks, each perfectly in phase with the other, and each filtered differently to remove all but a selection of frequencies, giving them a group of tracks split into different bands.
Then you would need a chain of devices (probably filters, EQs, dynamics, then more EQs and filters in series or parallel) to go on each of these tracks. It was rather too easy to have phase problems arising.
You could easily end up with mis-settings of control parameters across the different devices that covered the various bands, causing a loss of focus and clarity at those bands. Back in the days of analog, it was a daunting proposition in terms of both cabling and sufficiently resolute parameter settings.
Today, a multiband processor is a simple thing to insert on a track. It’s even more powerful than ever before these days, and you should experiment with them but be very careful if you actually use one on your latest song.
You need to develop good critical listening skills to deal with them safely, since they target only parts of your audio’s overall tonality. You must be able to target bands too, using your listening skills as well as you can, so you can tell what it is you are hearing accurately enough to make good decisions for your music.
PUMPING – OR NOT
You have a stereo mix (or a stereo submix in your song session) that you want to compress.
Unfortunately, the compressor is being triggered by the 4 on the floor kick drum pattern that the loudest thing in this whole mix, or mix stem (a.k.a. subgroup, submix, bounce). The compression clamps down on the whole mix every time the kick happens, causing a dramatic pumping effect. This can be musical and exciting, of course, but it is hardly transparent and so it is not always appropriate for your music’s emotional range.
In most cases, you will want to prevent or at least aesthetically control “pumping” of level when you compress a whole mix, especially if mastering your music, and the same is true for any sub-mixed group (e.g. your drum tracks subgroup).
Multiband lets you have your cake and eat it too. The easiest way to fix the above issue would be to split the audio into two bands, one set to keep only frequencies below, say, 150Hz (the very lows and some upper lows, the bulk of the energy in the kick and bass instruments) and one only containing frequencies above 150Hz (all the other frequencies).
This two-band method is very simple to do, and cool for mute effects in a track where you suddenly throw all the highs and mids out of a song for a bar or two. However, most of today’s plug-ins will have four or more bands you can set, each with upper and lower limits that can be set and that can and do have overlap regions. This gives you way more flexibility.
You can do some very accurate and minimalist de-essing using multiband compression.
Obviously, you don’t really want to compress the whole audio file when the “esses” happen.
It’s using a sledgehammer to crack a nut, as they say. Far better to compress only the frequencies that are really energetic in the “esses” and leave the rest alone.
In mastering, I deal with sibilant mixes quite a bit. In a perfect world, the mix you do will deal with this by inserting a three-band (or more) multiband dynamics processor on the vocal in question (or hi-hat or cymbal track) and setting the middle band to between 2kHz and 8kHz (normally around 5k or 6k with singers) with compression that does the job. Leave the low and high bands uncompressed. Sorted!
Here’s a screenshot of a multiband maximizer, just so you know about them! They allow frequency-band-specific targeted level maximization, and uses for them are found mostly in mastering. This one is a Waves low-latency processor. The biggest problem with multiband processing plug-ins is that they will eat up lots of CPU cycles. It’s a lot of number-crunching and it takes a bit longer than real-time. The latency can become a major issue unless delay compensation has been switched on in your DAW.
Across the main stereo outputs of the DAW, any latency won’t matter during mastering since nobody is trying to play along and there are no other tracks to remain in sync with.
DO THE LEAST HARM
Do the least harm. That is your motto when using these tools. I’ll leave you to explore multiband dynamics yourself. It is very cool and will help you get some crazy bass pressure in your dance music.
It bears repeating to do the least harm. You can’t remove the compression later and you can really do some serious damage, so do the least harm. Please.
Your mastering engineer will give you a dynamic, exciting, professional and awesome result if you let him do this part of the job for you, but do check it out so you can talk to them about how you feel about it. It really does require at least a few years of experience to master these tools in any serious way, but don’t be shy to try them.
For the next two posts, I’m dealing with “PROCESSING – TIME-BASED EFFECTS Parts 1 and 2″. I’ll be talking about processing that relates to time more than frequency or amplitude, although all are involved. That means delays and modulation effects like chorus and flange, for the most part. Then we’ll move on to reverbs for two more days of processing fun before finishing up May’s blogging with lots of details about microphones, and a quick tour of preamps and DI boxes.