Behringer x32

“One thing I’ve found to be true over the last week is that if you want to get to know a device or piece of software well… teach people how to use it. You’ll probably learn something too.”

This observation, which I posted as a tweet, seems to hold true for several others. As I continued to help nick with his x32 broadcast configuration, which I began discussing in my last post, we made some interesting discoveries. These discoveries lead me to a fresh sound for my streaming setup as well.

The desire to revamp my broadcast configuration started with Nicks discovery of a rather disturbing problem with the combinator, which I will discuss in a bit. However, to fully appreciate this oddity, it’s a good idea to understand how multi-band compression works.

A multiband compressor allows you to split a track into different frequency ranges (called “bands”) and compress them independently.
You can choose to compress only a certain part of a track’s frequency spectrum, or apply different flavors of compression to several areas of the spectrum.

This effect is useful in a broadcast environment because, if used correctly, material from a wide variety of sources can be molded in to a uniform sound. If you listen to most terrestrial radio, you’ll hear how everything is at the same volume, has the same amount of base, etc. We weren’t trying to take the combinator to the extreme of some stations though. We wanted a more subtle version of that effect.

Nick and I discovered that there were two major contributing factors as to why the combinators defaults make this impossible. These problems effected him more than I, as his configuration is more reliant on the combinator than mine, so it will be discussed in a future post, as I said above.

Band separation

I had initially discovered The first, and biggest problem, which you hear in Drues video. The combinator uses a 48 db crossover between each of its 5 bands. This means that very little sound that one band processes is evaluated by the others. The results are a hollow and brittle sound, as the frequency components of most voices and music are brutally sliced apart by the crossover. The brittle quality comes from the fact that when sound falls through the proverbial cracks, you can tell. It’s basically the audio equivalent of having a group of people sing in harmony, with the problem of each person being in there own practice room! Naturally, the people will sing much better together if they are close to each other, and can hear those around them. So, the first thing to do is narrow the isolation of the audio on each band by changing the crossover from 48 DB to 12 DB.

Now that the harsh hollow brittle sound is eliminated, its time to configure the combinator as if it were a standard single band compressor.

All of your standard parameters are available for your tweaking pleasure… Threshold, ratio, attack, release etc. However, other parameters exist for further sculpting of sound.

Crossover Frequency:

This parameter adjusts where the audio is divided. If the material the combinator is to process has a lot of base, letting it focus more on shaping the lows may be ideal, same with the highs. Fortunately, its possible to make the distribution such that a very nice balance is achievable.

Band solo:

It is possible to isolate the audio processed by each of the 5 bands. If the audio is pumping or being crunched, the problem frequencies are obvious. Granted, that may be the desired effect, and if it is… use something else, there are software plugins that let you do hard clipping.

Band specific threshold and gain:

This lets the combinator be used as a dynamic eq.
Do fun mastering tricks like compress the highs, then turn them up, or play with the low bands so that you compress tracks until they have less base than you want, then boost the low gain until the base has came back. Then notice that material with less base gets more, and base-heavy tracks lose a bit of it.

spectral balance control:

This parameter is enabled by default. It does more harm than good, as Nick discovered.

the idea behind SBC seems to be that the overall output should increase if the differential gain of any band drops by a specified amount of DB. The degree of auto-makeup is governed by the Spectral balance controls threshold and speed settings.

Nick discovered a big problem created by SBC which, in his case, broke more than it fixed. The mock scenario was a voice break with a quiet music bed in the background. His voice was driving some of the bands in to limiting, and the SBC was trying to keep the combinators overall output gain the same as its input. The result was a strangely equalized bed being boosted in volume when he talked! No good!

So like a good tutorial-following person, he went to layer 1 of the combinator, and pressed the third knob below the display to disable SBC. However, nothing happened! So, I started playing around with a freshly instantiated combinator, and due to being in a hurry, accidentally pressed the fourth knob.

, that did it! The change was almost as amazing as that made when adjusting the band separation.

The rest of the story, as such.

After a bit of coaching from me, Nick was routing channels to sends, pairing them to create single stereo outputs, and processing each of them the way he wanted.

If you have any questions, leave them in the comments so that the answers benefit others. Plus, embedding a twitter conversation that starts with the same text as a post title looks ugly to me. So, I’d rather not.