Mono to Stereo – and (Safely) Back Again
Sooner or later, it’s likely that you’ll be confronted with the problem of making a mono track more interesting by adding some width. Here’s how to do it the prudent way.
by David Baer, Nov. 2013
Sooner or later, it’s likely that you’ll be confronted with the problem of making a mono track more interesting by adding some width ambiance to make the listener’s experience more interesting. Reverb sometimes does the trick, and there are a number of special effects plug-ins that do other manipulation that accomplish the same – the Plug and Mix line’s Dimension 3D and Cool Vibe effects comes to mind (http://www.dontcrack.com/).
But reverb is not always the most appropriate solution nor are not-so-subtle “special effects”. Sometimes you’d like a bit of restraint, and there are several tried and true tricks to make this happen. However, there’s a potential downside to some of these tricks (or reverb or other effects for that matter), and that is the degradation in your sound when a stereo mix is played back in mono. The villain is comb filtering, which can make your audio sound anemic or hollow.
The subject of the importance of mono-compatibility comes up from time to time on the music production forums. Is this problem one you need to worry about? There are two situations commonly cited about why one should, the more important being having your production aired on a public PA system, be it in a club, an airport, a supermarket or at some community function. The second is FM broadcast. When the signal is weak, FM receivers try to compensate by reducing the reception to mono.
Personally, I think that if weak FM reception is an issue, comb filtering is the least of your worries. PA broadcast might be another matter. Most PA playback is done in venues that you’d hardly deem a critical listening environment, so who’s going to notice in the first place? But it might be important to you and if so, read on.
We’re going to look at a few tried and true techniques here, with audio examples, so that you can judge for yourself. All but the first are mono-safe as well.
The Haas Effect
Helmut Haas was an audio researcher who established how wave fronts arriving at each ear at different times are perceived. When the time interval is between two and 50ms for speech (two to 100ms for music) the result will be the perception of a single sound which has a stereophonic quality. The effect is also known as the precedence effect. Haas’s research went much deeper of course, but that’s the crux of the matter that concerns us here.
So, a trivial solution to stereo-izing a mono track is to delay one side by typically between ten and sixty or so milliseconds. There’s nothing more to it than that … no filters, no amplitude differences, nothing. The most rudimentary of delay effects will get the job done.
I’ve constructed an example of using this technique. The delay settings (I used Fab Filter’s overqualified Timeless 2 in this case) are pictured below and Example 1 is a brief audio clip of the results. In all the audio clips except the second one, we begin with a few measures of mono and then cross fade to having the effect present. The Haas trick is clearly very effective in accomplishing our goals.
But if the two signals are summed back to mono, and something not so nice ensues. With time differences as short as those we use with the Haas trick, signals from the two tracks can cancel each other out at some frequencies and reinforce amplitude at others – in other words, unwanted comb filtering. What frequencies are affected will depend on the audio content and the delay amount, but it can be unpleasant. Example 2 demonstrates this. In this audio clip two short passages are included, each played first in straight mono and then with the comb filtering.
Employing Mono-safe Delay
There’s another widely known way to incorporate delay that is completely mono-safe, however. It uses a mid-side (MS) to stereo (LR) conversion. We can do some cool tricks by taking an LR signal, converting it to MS, applying some manipulation like EQ or compression on just M or S, and converting MS back to LR. Higher-end effect plug-ins often sport this capability.
In this case, though, we start with mono, which becomes the M channel. The S channel will be the M channel delayed by between ten and maybe sixty milliseconds. Convert MS to LR and we have differences between the L and R channels … in other words, stereo. The width ambience can be subtle but also quite lovely. It’s normal practice to put a high-pass filter on the S channel with a cutoff at or above 100Hz, because the lower frequencies can “gang up” on one side and shift the audio image in unwanted ways. But the higher frequency shifts are interleaved in a balanced fashion that doesn’t favor either side.
And here’s the beauty of it. If summed to mono, the side information sums to zero. Result: no comb filtering. Again I’ve used Timeless 2 to create the demo clip. Timeless 2 has an on-board MS conversion capability, so I only need one effect. Lower-end delays won’t have this (although most should have an on-board HP filter), so you’ll need two plug-ins: the delay and an MS-to-LR converter. If your DAW doesn’t supply an MS/LR converter, there are several excellent free downloads available, like this one: http://www.voxengo.com/product/msed/.
If there’s any criticism to be made of the MS-delay trick, it’s that the stereo bands favoring the respective sides will drift around based on the audio content. Depending on your esthetics, you might actually find this a side benefit and not a downside at all.
Band-stable Mono-safe Manipulation
If the previous technique doesn’t cut it, here’s another approach. You can do complementary EQing of multiple bands. In doing so the frequency bands are stable. The easiest way to do this is to use a plug-in dedicated to the purpose. In this case, I’ll demonstrate using Melda Production’s MStereoSpread.
MStereoSpread offers two stages, one using complementary L/R multi-band EQing and the other using delays. But the delays are optional, so mono-safety does not need to be sacrificed. The advantage in using an effect like this is that it will probably supply far more bands than you can introduce doing your own EQing. The demo clip is done with a factory preset using 44 bands, but MStereoSpread is capable of more than twice that. The results are quite nice, as can be heard in Example 4.
MStereoSpread lists for a little over $100USD, but Melda Productions (http://www.meldaproduction.com/) has fairly frequent sales in which it can be acquired for half that price.
But what if you have a fairly competent EQ plug-in that supports a generous number of bands separately adjustable for each channel? Do you need to acquire a specialized plug-in to do this? Based on my experimentation, the answer is no.
In this example, I’m going back to Fab Filter for its versatile Pro-Q effect. I’ve used ten bands, as pictured, even though two more could have been crammed in there. The screen shot tells the whole story. Just to be on the safe side, I did use the Medium Linear Phase mode out of concern that all that EQing in close quarters might have some unwanted phase side effects. Example 5 demonstrates the results.
OK, suppose you don’t have a high-end EQ effect in your arsenal and your budget doesn’t accommodate a specialized stereo-ization effect. Can you do it with the EQs you have on hand, maybe one that came with your DAW software? The answer is a qualified “yes”. Yes, you can get away with it. However, I was not nearly as pleased with the result as with either the MStereoSpread or Pro-Q solution.
In this case, I’ve used SONAR’s Pro Channel EQ. You can judge for yourself if you find it adequate by listening to the clip in Example 6.
So there you have it. You can have your stereo-ized mono and mono-ized stereo too, keeping the comb-filter boogey man at bay in the bargain.