Stereo image masking
Since the period of ‘in the box' processing, I've been critiquing mixes for many years and one repeating trend I've noticed is a lack of concentration around stereo image mixing.
Our shift to digital audio workstations [DAW] as a mixing platform on a large scale must give an ever-expanding toolset, outstanding adaptability during production, and oftentimes, the ability to streamline tasks that would have taken days with a fully analogue system. It has spawned plenty of new genres and ways to collaborate musically, and the sound quality has increased in several ways. For a variety of reasons, we will never revert to the "good old days of analogue." However, in some regions of mixed construction, this new manner of working has resulted in a lack of focus. To help improve our digital outcomes, we can draw some positives from analogue mixing mentality and signal path. The use of stereo field or the stereo impression of the mix is one example.
When working with a DAW, the routing paths are simply set up by default when we create a new channel by importing audio or loading new virtual instruments. If it's a mono sound source, it'll go to the centre of the left or right channel (mono). Alternatively, if a ‘stereo' instrument or effect is formed, it is discretely routed to the left and right (mono left and mono right). Stereo-only exists because we give two mono channels a relationship, and I believe it is useful to recognize stereo for what it is in the reality of the signal path.
One issue in comprehending path routing is pseudo stereo. Where the source is in fact mono but creates a ‘stereo’ image by the use of phase in terms of delay between left / right and/or tone difference (one side is brighter darker than the other). When this folds to mono it often cancels to some extent or causes a phased aspect to the sound. Some virtual instruments and reverbs or delay even use polarity to give an impression of width, which cancels in mono. Not very useful for your mix on a mono playback system. You can test whether your effect ‘mono sums’ on input by signal into either the left or right of the process and listen only to the ‘wet’ signal output. If the ‘stereo’ of the signal doesn’t change the input to the process is summed; if it is a ‘true stereo’ effect the wet signal will alter in the image between left and right input. In some ways, a mono in, the pseudo stereo output is fine for giving a mono source give a ‘stereo’ feel, but not the best outcome for a stereo or mono panned source where you’re trying to create an ‘environment’ for those sources to sit in too.
Another good example of a Pseudo stereo virtual instrument is the piano, which tends to be placed as wide across the stereo as possible, bass through to treble in terms of register. Sounds great when played in isolation but occupies all the stereo field when in a mix. If we wanted a realistic piano sound of a concert hall, the listener would hear all the direct signals as mono as the audience could be thirty meters or more from the piano on stage; we are not going to hear the bass on the left and treble on the right. The auditorium will create the stereo and spatial impression through acoustic reverberation. But often in non-classic production, we want a more intermate outcome.
Panning the frequency of the notes across stereo makes sense as it unmasks each note from another in register (frequency) across the stereo field. But manipulation of the left-right pan controls to balance it into the stereo field of the mix will often achieve a much more cohesive effect with everything else, plus better mono compatibility. Creating a mini stereo inside the big stereo, what I like to refer to as ‘pocketing’. Obviously, in terms of a production aesthetic, there isn’t a correct answer, if you want a super-wide piano in your mix, that is a choice, but if you have not thought about the stereo impression of the piano and just gone with its default output, you have not made a choice, you are using someone else’s pre-set. Never a good way to achieve a unique sounding outcome, which is after all that makes famous tunes famous; well-crafted song writing and outstanding production.
A mix in this form does not necessarily sound awful; it simply does not access the level of detail that is attainable with the simple action of using proper panning technique to unmask the fundamental elements of the mix and release the effects, resulting in a deeper spatial impression.
If in reading this you feel you may have fallen foul of your DAW’s ease of routing in producing music, even a basic approach of placing your stereo effect elements at differing stages of the pan as across the image using a dual panner mode in your DAW with have a big impact. There’s always a way to achieve control of the pan for each mono element of the stereo part, even if this means printing the audio in split stereo and loading back to mono tracks to the pan. Once this ‘control’ is achieved you should think of each stereo element as a ‘mini’ stereo that sits in the ‘big’ stereo mix. This is in-part way engineers have been bus mixing since the advent of stereo. It is easier to control a mix in its ‘sections’ ‘mini worlds of stereo’ than as a host of individual channels. This manipulation of the panorama can be is as simple as L95 / R95 L90 / R90 and so on with the effects returns. It will create a wider stereo image and clearer detail overall. If you try this and bounce a version of each mix and AB, you’d clearly hear the detail difference as your mix opens out.
Even though there is less in the hard stereo, the mix will appear wider. This is just simple unmasking, allowing the details to come through. Also, don’t forget the mono aspects, moving parts placed dead centre even a few points can achieve more clarity. A good rule of thumb is no element should be at the same pan position as another unless it is not in the same register (frequency range). There a good rationale for why the kick/bass/vocal and snare are often all work well down the centre. The kick sits in frequency below or above the bass, the vocal above these the snare above and / below often wrapping the vocal, and let us not forget the hat on top. Frequency pocketing is just as important as a panorama.
All of the parts are working together to expose our blend. All of these elements will sound unmasked in the same pan position since they are separated by frequency. Hard left/right distorted guitars with broad stereo cymbals on top follow the same reasoning elsewhere in the mix. In the same pan position as before, but without the frequency mask.
In terms of mastering, not only does this make the mix more transparent, it needs less processing overall as depth is fuller, hence the dynamic range is reduced as everything isn’t sitting on top of everything else. This really helps with apparent loudness in the final outcome. As I said many times before; loudness is achieved through effective mixing and not by the use of excessive dynamic control during mastering. If your mix doesn’t have it, it is not going to be ‘fixed’ in mastering; mastering will just make your mix sound better.
Taking this idea of separation to the next level, you can use busses to control sets of instrumentation and their tied effects in the overall stereo field. Very effective for strings, brass, guitars, drums, piano, etc. You can not only control the apparent width but also centre focus of that image in the overall ‘big’ stereo. This concept I like to call ‘pocketing the bus.’ One note; don’t be tempted to use ‘stereo imaging’ tools to create ‘stereo width’. This often will introduce unwanted phases into the mix and on the whole, they are not allowing you to actually change the left-right panning position, just the width of the mid/side relationship. All these types of imagers are just a simple mid-side matrix in principle. You can make one in the DAW, cheaper and you’re not using someone else’s pre-set either; you know where and how your sound is being manipulated.
To summarize; there’s never a correct answer to any mix or approach towards it, but experimentation and application of pan will deliver many rewards in terms of the spectral detail in your mix. This can only lead to a more positive outcome when mastered.
What you should ask yourself is; ‘have I considered the panorama or each element’, if not, you’ve not made a production choice, just used a default outcome.
JP Braddock
Mastering Engineer
