Tutorial – Audio Depth Perception in Your Mixes
There are two techniques, which can be used separately or together, useful in adding a sense of stage depth to your mixes using reverb. We take a close look at them here.
by David Baer, July 2014
There was a thread on a forum recently where the poster asked about how to get a perception of depth into his mixes. As so often happens on forums, the responses were varied and inconsistent. I was intrigued enough to research this topic a little further and found that there was a fair amount of “expert” advice available on the web, some of it very authoritative-sounding but also completely contradicting other equally authoritative-sounding advice. Misinformation on the web? I know you’re as shocked as I am to learn of this!
But with persistence, a common train of advice started to become evident regarding the use of reverb and pre-delay to accomplish establishing an illusion of depth. Then, a few weeks later I was watching the excellent Groove 3 video tutorials on MIDI orchestration, as presented by master instructor Eli Krantzberg. He demonstrated a very simple technique in one of the first few videos of the series that was immediately evident as being straightforward and effective.
So, this is a summary of what I’ve learned. There are two techniques which can be used separately or together. Both are quite straightforward, and one of them can be done with even the most rudimentary of reverb plug-ins.
But before we dive in, let’s consider two basic facts of sound transmission. The first is that if air were a perfectly lossless transmission medium (it’s not), if you double the distance between a sound source and the listener, the loudness will be down 6 dB (assuming no reflections are added to augment the level). In real life air is not a sound “superconductor”, so the attenuation will be a bit more. And that leads us to the second factor: higher frequencies are more subject to distance-related attenuation than lower frequencies.
So even without any reverb in the picture, we can “set the stage” to promote a sense of depth with faders and EQ. The first audio clip below does just that. We have a trumpet (a stock sound from Halion 5 with all internal effects disabled) coming from front-stage, mid-stage and rear-stage. Front-stage is panned center, mid-stage panned halfway to the left, and rear-stage panned halfway to the right.
The listening position is assumed to be about the same distance from front-stage as the distance from front-stage to rear-stage (but don’t bother to get out your calculators – this isn’t precisely mathematically correct). The three positions occupy their own track. The mid-stage sound is reduced 3 dB and the rear-stage 6 dB. Some high shelf reduction was added to mid-stage – just enough to barely notice. Slightly more was added to rear-stage. No other effects or treatments are present other than the EQ and the reverb.
All three tracks have a pre-fader send (with levels and pan-position duplicating post fader settings) to separate, per-track reverb buses with B2CAudio’s Aether reverb plug-in on them. All buses are mixed in at -15 dB in this first clip. For all these demos, we use a B2CAudio-supplied hall program, modified only to modestly reduce the Time parameter.
Although there’s no need for this baseline demo to use three separate reverbs running the same program, it will be evident why this becomes necessary in the remaining demos. Here’s the baseline demo:
You can hear the sound getting farther away but the depth illusion is not all that effective. So that brings us to the first technique.
More Reverb = Further Away
The method Eli Krantzberg demonstrated that I mentioned earlier relies on the common-sense notion that the balance between direct sound and hall sound will alter as the sound source moves further away from the listener. As the source moves further back, you’ll get less direct sound and more reflected sound. So we just imitate that behavior. This clip is identical to the baseline demo except as follows. The front, mid and rear reverbs are mixed in at -15, -12 and -9 dB respectively. These amounts may be a tad more than you’d wish to use in a real mix, but they do show that this trick can be quite effective, or at least it seems that way to my ears.
Of course there’s something sacrificed in using this approach. The sounds farther back lose clarity. Now, to a certain extent, this mirrors reality. Sounds from the rear of the stage (in any large setting) will have more reflected content compared to direct content. But you do have to ladle it on a bit thick in the mix to make this illusion really obvious. And that might be an unacceptable compromise. So that leads us to another technique, one which avoids the problem of muddiness.
More Pre-Delay = In-Your-Face
Another commonly-cited technique is to use a reverb plug-in’s Pre-Delay parameter to enhance a sense of depth. A number of sources recommend adding 50 ms of pre-delay to place a sound at the front of the mix. 50 ms is about the most you can get away with before the delay starts to be perceived as an echo rather than reverberation.
Then advice starts to differ. Roey Izhaky, the author of what to me is the Bible of mixing, Mixing Audio, doesn’t advise what amount of pre-delay to use for sound sources further back. As an aside, let me say that the 60-or-so page chapter on reverb in this book alone is worth the price of the entire book. I have found this book to be indispensable and could not recommend it more highly.
One web source I found recommended 50, 40 and 30 ms values for front, mid and rear. Another recommended 50, 42 and 35. But let’s start by trying something more drastic: 50, 25 and 0 ms respectively. In the next demo clip, all reverbs are mixed in at -12 dB. The reverbs differ in only the amount of pre-delay.
To me, this effect is rather subtle. I can hear it if I’m listening for it, but I don’t find it nearly as obvious as the technique shown in the Demo 1 clip. However, it does remove the muddiness problem of rear sounds being drowned in reverb.
Let’s try one of the web recommendations: 50, 40 and 30 ms:
To me this is neither better nor worse with respect to the depth illusion. However, pre-delay has its own virtue in reducing the chance of unwanted comb filtering. A little pre-delay will nearly always do more good than harm. So, my vote goes for the 50/40/30 solution.
Double Your Pleasure
And the next obvious thing to try is combining the two techniques, so we’ll do just that. In the following clip, we will use the 50/40/30 pre-delay settings of Demo 3 and the reverb levels, -15/-12/-9 dB, of Demo 1.
Are we there yet? Well, almost. But there’s a refinement on the above configuration that seems to me to be the sweet spot. Rather than increase the overall reverb levels as we move further back in the soundstage, what if we instead limit the increases to the early reflections? The reverb tail contributes more of a muddy quality than the ERs. At the same time, the ERs do much to place a sound in terms of distance perception. So, in the final clip, we do the same thing with pre-delay (50/40/30) as the last time. The reverbs are all set to -15 dB, but the mid reverb has had the ER level raised 3 dB and the rear raised by 6 dB.
And I think we have a winner! Now, you will need a reverb with both pre-delay control and ER level control to do this. Most moderate quality (and up) reverbs can be expected to offer this.
So, let me recap the recipe:
- Adjust your front, mid and rear sounds to taste so that the level is less the further back you go and maybe some high-end EQ reduction can be added as well to the further-back sounds.
- Run your front, mid and rear sounds each to one of three buses with a reverb on it (name these effects buses appropriately to avoid sloppy errors). Use the same program on each reverb instance and use the same mix level on each of the reverb buses.
- Set the pre-delays of the front, mid and rear reverbs to 50, 40 and 30 ms respectively. Set the ER level of the mid reverb up by 3 dB (or to taste) and the rear reverb up by 6 dB (or to taste).
One optional refinement could save CPU consumption. Turn the late reflections (tail) portion of the bus reverbs off leaving only the early reflections enabled. Then route the output of the three buses to another bus with a reverb program running only the late reflections. Reverbs have to do considerably more work when generating the tail portion, so consolidating that into one instance might keep you system running a bit more efficiently.