Sonible smart:comp by Sound-Guy
Sonible are at it again, this time with a compressor, a “smart” compressor. Previously I found their smart:EQ2 to be a good surprise, actually helping balance tracks and mixes quickly and impressively, so can they do something as impressive with a compressor?
As with their other products, Sonible use the iLok PACE system, and since I already had the PACE module on my system, the installation and authorization went smoothly and quickly. You do need the PACE module (free), but you do not need an iLok dongle or even an iLok account. The iLok connection is good news for many - you can authorize smart:comp with an iLok dongle, but you need to contact Sonible to get the dongle license set up. And since you get two activations with each license, you can use smart:comp on two machines or one machine and one iLok dongle. You can also run a demo mode to evaluate smart:comp before you buy it, so you might want to give it a try.
What Is It?
Smart:comp is, like Sonible’s other smart processors, an “AI” assisted audio tool, in this case, obviously, a compressor. While smart:EQ 2 processes the spectral content of a track or mix, to balance the EQ for a better sound (which it does very well most of the time), smart:comp has a bit more to do. A compressor has more interacting parameters than EQ, with threshold, ratio, attack and decay times. And smart:comp has even more controls including spectral compression, band-pass EQ and a complex set of attack and decay adjustments.
If you are still trying to figure out how a compressor works or understand what one does but still find some tracks that don’t seem to respond properly to your adjustments, smart:comp may be very helpful. And if you are smarter than smart:comp, you can bypass its “AI” mode entirely and use your own settings, and will likely find its array of controls very useful.
As a compressor, smart:comp has an excellent set of controls. There is, of course, a threshold setting, from 0 dBFS to -50 dBFS rms (the detection circuit is not a peak detector, but responds to short term average), and a ratio control from 1:1 up to 10:1, which is pretty much a limiter mode. The compressor “knee”can also be adjusted from sharp to gently rounded. Attack includes a “hold” time, up to 60 msec, which is in addition to the attack time of up to 200 msec. You can vary the slope of the attack envelope, and even adjust its “curve” from fast-start-slow-end to slow-start-fast-end (i.e., convex or concave). There are similar adjustments for release times with a hold up to 100 msec and release time up to 1.5 seconds. Hold times are an excellent feature, something we don’t see enough on compressors.
But that is not all for the detection system - there is also a “detection focus” control that is very clever. It is a filter for the detection “circuit”, but not just a high pass as many compressors have. It provides a band-pass control that you can set graphically in a window while viewing an overlay display of the real-time spectrum. So, for example, you can quickly see a snare hit graphically as you hear it, and focus in on it if you want it to be the main aspect of control, or place it out of the focus area if you want it to punch through. I’ve not seen a control quite like it, although I have compressors with band-pass filters that can be set with frequency values. This is much more intuitive.
In fact, all the controls have both “dials” to set values and graphic views with handles you can drag to adjust: for example, in the case of threshold and ratio you can visually set the threshold as you watch a graphic plot of the short term loudness, and shave off as much or little as you wish. The loudness plot also shows both the original signal level and the compressed level as you can see in the screen shots.
And maybe the most impressive, and pretty well a unique mode is spectral compression, seen in the lower panel of the smart:comp window. This analyses the input track in real-time, across more than 2000 frequency bands, making smart:comp an intelligent, ultra-high-resolution multiband compressor. Not only a multiband compressor, but one that adjusts compression based on the spectral content of the music and on a “profile” using one Sonible has developed for different styles/tracks. These include a “standard” profile for overall mixes, and profiles for both female and male vocals, bass, drums, kick, snare, keys, and guitar. And if you want to “fine tune” how spectral compression is applied, you can even target or exclude a range of frequencies to be treated with spectral compression by moving a couple of blue bars at the top and bottom of its spectral display!
Let’s Get Smart
Speaking of profiles, smart:comp uses the same profiles to set up all of its compressor parameters using its Learning mode. The learning mode sets up attack, release, threshold and ratio, as well as the spectral compression settings. You can choose not to use spectral compression after a track is “learned” just as you can change any the the normal compressor controls to your liking. Although smart:comp doesn’t have the “weighting control” of smart:EQ 2 (I’m not sure how a “weighting” adjustment could work for all the compressor parameters anyway), it does have a sensitivity control for the spectral compressor that works very effectively controlling the strength of the multiband application.
There is more. Smart:comp has a sidechain input that I confirmed works fine in REAPER using its extra track-channels, and in Studio One using its sidechain routing. This enables “ducking” one track with control from another, and the Learning mode and detection focus can be applied to the sidechain control signal, while the spectral processing can be applied to the controlled track. This provides more than you might think at first - ducking of the controlled track can be frequency-selective based on the spectrum of the controlling track! This enables, for example, a bass track to duck the kick track only at the current frequency of the bass, rather than duck the entire kick level. Really a flexible dynamics tool.
Does It Work?
Absolutely! The smart aspect is often “right-on” for good normal production compression needs, and is a great starting point for more characterful compression. I was very impressed with what I could get with vocals - they could be transparently smoothed out, especially with the spectral compression cranked up in the 100% range (as you can see, it actually goes to 150%!). Drums could be smoothed nicely, or elements brought up to a wonderful crunchy thumping. I was “blown away” with what I heard on drums tracks! I tried smart:comp on a kick track that had high levels of snare and cymbal bleed, and was able to almost “extract” the kick sound - at least it was brought more clearly into the foreground, enough to greatly improve the drum balance in the mix.
I was impressed with every track and mix I tested - used with the profiles and normal learning mode, smart:comp provides very transparent level control. Yet I could easily take things to the point of extreme tonal change with a few adjustments. On a dance mix I cranked up the ratio, cut the attack and release, focused in on the kick and really changed the feel of the whole song, with about 30 seconds of work. A truly fine and useful compressor, and I believe it could be a great learning tool for “neophytes“ .
Tested using a PC Audio Labs Rok Box PC, Windows 7 64 Bit, 4-Core Intel i7-4770K, 3.5 GHz, 16 GB RAM. A single instance of smart:comp, measured in REAPER, used 0.20% cpu resource, which is very low for a compressor. However, latency is 2048 samples, so it’s not a compressor to use for live work or while tracking. But in a mix project you could easily use a dozen or more.
A truly surprising take on a compressor, that even without its “smart” mode brings new capabilities to a studio. Definitely worth a look and listen. Highly recommended!
The smart:comp Learn function works well for any style of music I tried, and the profiles help individual tracks and buses to stand out clearly in a mix.
Graphic controls that are superimposed on the input loudness plot, as well as dial settings, make adjustments visual and easy.
Extremely flexible attack and release controls with a “Hold” mode available for both!
Detection “Focus” control with graphical feedback of input spectrum to easily target specific frequencies/instruments.
Unique Spectral Compression mode that analyses the input track in real-time with more than 2000 frequency bands making smart:comp an excellent multiband compressor.
Sidechain input for detection system with the ability to use the instrument profiles, detection focus, and the spectral compression function to provide frequency selective ducking.
Auto Gain mode helps prevent loudness bias when comparing pre and post compression.
Very low cpu usage, under 0.2% per instance on my system.
Not much I could complain about, though latency is a constant 2048 samples, so not good for live use or tracking.
Does not include built-in mid-sides processing like smart:EQ 2; but it does so much so well, I can’t really complain. You can always use a tool like MSED to do mid-sides processing, and since you always process the mid and sides signals differently (or why would you bother with mid-sides!) you can use two instances of smart:comp since it has such a low cpu requirement.