Sonible smart:reverb by Sound-Guy
smart:reverb from sonible
Austrian company sonible have created some fascinating audio plug-ins that are “different” from most. When I first tried out their smart:EQ 2 I expected another spectral matching plug-in similar to several I already had. I found a significantly different approach to other such EQs. Then they surprised with a smart compressor (smart:comp) that took the idea of trying to analyze dynamic sounds to new places. Now they have applied their methods to a reverb unit. How can that be?
What is It?
smart:reverb (SR) is an algorithmic reverberation plug-in that has some unique parameters, even without considering its Learning mode. I was a beta tester of smart:reverb before the Learning mode was fully functional, and I was impressed with what it could accomplish without even using that function. But when Learning was finally fully implemented, it added some really unique abilities.
Let’s start with the basic reverberation engine. This is a mathematically “computed” algorithmic reverb rather than a convolution reverb. While both algorithmic and convolution reverbs use a lot of maths, the former compute the reverb effect based on summations of many, many delayed repeats with the characteristics changing over the repeats according to some mathematical relationship. Convolution reverbs use actual sound samples of a real space or mathematically modeled space, and impose those sound characteristics on the signal you provide. Both types can provide excellent “spaces”. Smart:reverb is an algorithmic design with a unique approach.
There are several controls similar to other such reverbs: reverb time (often termed decay time), pre-delay, spread and density. But you’ll notice in the screen-shot above there is both a Reverb Time dial and to its right, a graph (called the Temporal Shaper graph). This graph shows characteristics of the reverb time (and density and spread) over the width of the graph, marked off with 15 grid lines which marks off the total reverb time. This graphic display is not just a display – you can grab one of the handles and move it to change the height and slope, changing the instantaneous decay rate anywhere within the reverb time. Likewise the spread and density graphs can also be adjusted over time (first clicking the Spread or Density titles). This can create some very different reverb effects.
So with these controls alone you can obtain a great range of reverb effects that relate, theoretically, to the size and reflectivity of a room. And there is more. Above the reverb time dial are three controls, Width, Color, and Clarity. Width is just that, with a range from mono to wide stereo – it appears to be a Mid-Sides effect, with zero width leaving the Mid signal only and 100% being just the Sides. You’ll probably want to keep the control in the zero to maybe 60% range unless you’re going for some special effect. Color is essentially a tilt-EQ for the wet sound (doesn’t affect the dry signal). Positive values cut lows and accentuate the highs for a brighter sound while negative values do the opposite and create a darker sound. This is a very quick and effective way to adjust the reverb tone, but not the only way. Below the Temporal Shaper graph is a filter control that applies LP, HP, peaking or cutting filters to the wet signal. This can cut muddy frequency areas or create resonant effects if you’re making something really different.
Back above the Reverb Time dial is a third control, Clarity, that I found very effective. It provides a built-in ducking control of the wet level controlled by the dry signal. This is a trick I’ve used with reverbs many times, but had to do so with a separate gate or compressor using an external side-chain and some fancy signal routing. Having this right in the reverb is very handy. I found it works well even at low settings (20-30%) and setting it at 100% often caused audible pumping of the reverb time, so there is more than enough range for good results for any kind of music.
Speaking of settings in percent, all the levels in smart:reverb are set in percent except the filter below the Temporal Shaper graph which uses dB. So the Dry/Wet control below the reverb time dial is calibrated in percent (and accurately so). I found this a bit odd in that dropping a setting from 100% to 50% is only a 6 dB drop, but dropping from 50% to zero is an infinite dB drop! However, in practice the level controls work well if you click the little link “chain” between the dry and wet levels. In this mode the total will always add to 100% which enables quickly balancing the effect – except the wet signal is always lower in volume than the dry and depending on the other settings described earlier, may be much lower. This usually results in the output level dropping quickly as the dry level is reduced from 100%. This may make judging adjustments difficult if you use reverbs as insert FX. Which you shouldn’t do anyway!
I always use reverbs in a sends configuration, well, almost 100% – maybe 99% of the time. It’s rare I want a reverb effect in series with a single audio track. I put reverbs in FX buses with the wet level at 100%, and use sends levels from audio tracks to adjust how much of each track hits the reverb. In such a case smart:reverb works fine since it’s very quick to pull the wet level to 100% and with the link option on, dry goes to zero.
But There is More!
You may be wondering what is that matrix box? And what are those strange bubbles above the Temporal Shaper graph? And how does the Learn function work? In fact these are all related. Even without using Learn, you can use the matrix to modify the reverberation tones over time. It’s a control that effects the results in complex ways. The matrix has labels at top, bottom and sides: natural and artificial, intimate and rich. Moving the dot around (which can even be done using automation while tracks are being played) will change the resulting reverb significantly, changing the characteristic curves in the Temporal Shaper graph. The results sound like the descriptors, natural sounding in the upper half and more artificial reverb in the lower area. Intimate tends to sound closer with reverb pushed back while rich sounds like listening in a larger area farther from the source with more complex reverberation.
Things get even more fun after you have run Learning mode on a sound. As with smart EQ 2 and smart:comp, the Learn button will analyse the audio playing and use the results to set up several parameters. It will not change the reverb time – that relates to to size of the “room” you are simulating and is your choice. But you will see the lines in the Temporal Shaper plots change when you run Learn, and in the area to the right of the matrix (called the Particle Display) you will see a new set of lines appear. Learn will attempt to create time and frequency settings that compliment the type of sound you are processing. As with smart:EQ 2 there are also profiles available to optimize the effects. This defaults to “Universal” which works fine for an overall mix or sub-mix of many instruments, but you can also select specific types of sound such as keys, vocals, guitar, etc. if you are applying reverb to a specific type of sound. For example, I often send all drum and percussion to an FX bus and use a slightly different reverb on that bus than I use on vocals, keyboards or guitars. You can run Learn with the profile set to Universal and select a specific profile later or you can set that profile before running learn. It doesn’t matter since you will get the same result whenever you select a specific profile.
When you first run the learn function the Particle Display will add three vertical “broken lines”. What do they indicate? Notice on the far right there is a frequency scale from 20 Hz to 20 kHz and the scale on the top of this plot is time which starts at zero and goes to whatever the Reverb Time setting is. This graph shows the reverb time across the frequency spectrum and the “bubbles” or particle display shows the energy of different frequencies over time, and changing almost any setting will change the size and density of the particles which shows how the sound energy is changing over time and frequency. And just as with the curves in the temporal display below, you can change these new curves in the particle display by grabbing and moving a handle. This will change the intensity over time in different frequency bands. So not only can you adjust how the overall decay level varies over time, but also adjust how different frequency ranges decay over time.
Running Learn will result in the matrix set to neutral (dot centered) and you can then move the dot if you want more natural or more intimate sounding reverb. As you move the dot around in the matrix you will see the three “broken lines” in the Particle Display plot change shape and the curves below, Decay, Spread and Density, also shift their form. Really fun and really provides a fast way to check out a range of reverb variations.
Of course there is a preset system to save all your settings to recall instantly, and an A/B comparison mode. And as with any well mannered plug-in, all the settings will be saved in any DAW project for future recall. There is a Bypass button which bypasses the reverb, but does not have any sort of loudness compensation, although I don’t find this as serious with a reverb as I do with a compressor. And there is a Reset button which brings smart:reverb back to initial settings like opening a new instance.
And there are more special features too – the freeze button and the infinite button at the bottom provide some fun changes with freeze basically grabbing the most recent reverb sound and smoothly looping it until you turn off freeze, and infinite providing an infinite decay time which bypasses the Reverb Time setting. How long is infinite? Well, an hour after I left the studio with an infinite reverb on but the input signal stopped, the output signal level was still exactly the same, within 0.1 dB of what it had been when I went to lunch. By infinite, they mean infinite! And the freeze mode also plays on forever until switched off. I see some new vocal FX on the way!
How Does it Sound?
In short, excellent. I often think of algorithmic reverbs as “clean” or even “sterile” sounding, but smart:reverb sounds excellent and really surprised me when I compared it to some none-too-cheap dynamic convolution reverbs. Considering smart:reverb is entirely "computed" (ie, algorithmic) I was extremely impressed that with some adjustments to the curves, decay time and color, I quickly obtained reverberation effects that were extremely close to certain rooms/chambers in famous LA studios. So close that even with an exposed track (just a vocal or single instrument) I doubt many people could tell the difference, and I really didn’t find I would prefer one over the other. That says a lot for the excellent reverb these folks have created.
I've added some examples, each starting with a dry track and then using smart:reverb after learning with the nominal centered matrix result, then with the matrix control dot moved to various corners. First example is a drum mix, then a simple guitar riff repeated several times, and finally a beautiful vocal by Egda Carolyn (Google her) in Portuguese, one of my favorite languages for songs.
smart:reverb is available as Audio Unit, AAX, VST2 and VST3 – for 64 bit OS only, Windows 10 and Mac OS X 10.8 or later. Requires iLok or online PACE system connection (free) for activation, but runs without a connection after that. Supported sample rates are 44.1 kHz to 192 kHz. In my test system (PC Audio Labs Rok Box PC with 4-Core Intel i7-4770K, 3.5 GHz, and 16 GB RAM) a single instance of smart:reverb used from 0.8% to about 1.2% cpu resource depending on the settings used. Latency was zero.
Another unique and fascinating “smart” plug-in from sonible that delivers excellent sound, a wide tonal range of reverberation, and has some really unique functions. Even without the Learn feature it is a fine audio tool and the Learn mode provides an effective time-saving mode.
Excellent reverb quality from an algorithmic sound system with unique and very effective controls.
Easy and quick adjustability of reverb tone and time domain characteristics with a very wide range of effects.
Unique Learning mode that automatically “tunes” many parameters to best match a sound source.
The clarity control is brilliant and effective – provides ducking of the wet signal within smart:reverb which will save setting up an external ducking configuration.
Two modes, freeze and infinite, to obtain true “never ending” reverberation. Brilliant!
Very reasonable price for such a fine and unique audio tool.
My only concern is for those who use reverb as an insert (which I do not suggest doing!) since the level of the wet signal is always lower than the dry signal and there is no auto-gain to keep the overall mix level constant as you change the wet/dry ratio. For use in an FX bus this is not a concern.