You are here

Dbx AFS2

Feedback Suppressor
By Mike Crofts

Dbx AFS2

Is this Dbx’s most user–friendly feedback suppressor to date?

I remember when automatic feedback suppressors first appeared and were hailed (by some, including me) as a wonder of their age. Imagine having an invisible, unpaid extra pair of ears and hands constantly monitoring the PA for feedback problems and instantly applying selective EQ corrections as necessary, all without interrupting the performance or affecting the overall sound quality... It all sounded too good to be true at the time, but it actually worked pretty well in many applications, and the detection algorithms have now evolved and become increasingly sophisticated

The new Dbx AFS2 updates and replaces the well–established and successful AFS224, which has been around for a few years now. As a user of the previous model myself I was interested to have a look at its replacement in the product line.

I like to have a feedback processor available for every job I do, in the same way as I like to have extra sets of batteries and every kind of mains power adaptor — I may not use them every time, or even all that often, but every so often a situation will arise where that extra capability can make all the difference between a horrendous, stress–filled experience and a successful result. Much has been, and will continue to be, written about the reasons for and against using feedback processors, but I have always regarded them simply as automatic outboard EQ designed with a particular task — the control of feedback — in mind. One of the main criticisms of such devices has been the fact that they ‘alter the sound’, and of course this is true — that’s precisely what they’re for — but they are designed to alter the sound so as to eliminate feedback issues, and as such this has surely got to be a good thing.

With the AFS2 the team at Dbx have essentially taken the existing functionality of the AFS224 processor and developed new detection and control algorithms, along with an enhanced degree of user control and the ability to save and recall settings.

What It Does

In brief, the AFS2 has the ability to detect when feedback begins to occur at any particular frequency, and will insert a sharp notch filter into the signal path at that frequency. As with most devices of this type, it works by assigning either ‘fixed’ filters, which are determined and applied during the setting–up process and remain in place until manually reset, or ‘live’ filters, which are applied automatically during the performance if feedback is detected, without any action necessary on the user’s part. You could think about the two types of filters as addressing different aspects of a live show, in that fixed filters are addressing major issues relating to the venue and stage layout, and live filters will cope with anything that happens post–soundcheck, such as increasing monitor levels, or physical changes like singers moving around close to monitors, or even taking the mic out front so that the audience can assist with the choruses!

The huge advantages that the AFS2 offers over manually applied EQ — including traditional graphic equalisers — is that the centre frequencies of the filters are calculated to be at exactly the required frequency. Also, the filter width can be very narrow so as to apply attenuation at precisely those frequencies and not wipe out the ‘good’ programme material either side. Consider what a traditional third–octave filter actually covers in terms of a musical scale! The AFS2 can apply filters in three width settings, according to what the audio material contains and/or the severity of the feedback problems being addressed. The three types are described as speech (the widest setting), music (much narrower) or ‘music/speech’, which is of course somewhere in between and offers a good compromise. Although the ability to apply very narrow filters is desirable in more critical applications, it’s worth experimenting with the different settings as it may, for example, be better to apply one slightly wider filter than use up several narrow ones close together. It all depends on the application, and it’s always worth finding out what works best. The AFS2 does allow different filter types to be set for different filters — if you change the setting, it will only apply to new filters without changing any already set.

It’s worth noting two other clever things about the AFS2 filter types: the fixed filters will automatically adjust their centre frequencies and width during setup, and will only ‘permanently’ fix their settings when you move on to live mode, and the live filters have the ability to remove themselves from the signal path when no longer needed, thus freeing up filter slots for new filters and restoring the original fidelity of the material. If all live filter slots have been used, and new feedback is detected, then the AFS2 starts to reuse the live slots starting with the ‘oldest’, so it’s always going to address the current feedback issues no matter what has gone before — that’s neatness for you. These parameters and more (such as the ‘virtual high-pass’ function, which leaves lower frequencies alone, if you wish) can be accessed via the AFS2 option menu, which allows control of most aspects of the AFS2’s operation.

Let’s have a tour of the unit itself before getting on to using it in anger...

Racking Up

First off, the AFS2 is a well–built unit, with a neat and sturdy steel single-rack case, and the front and back layout is practical, simple and neat. There are cooling vents on top (above the power-supply area) and at both sides, but no fan. Although I didn’t mount the AFS2 in a rack, all I can say is that it didn’t exhibit any signs of getting warm after several hours of use. It’s quite a shallow unit so I’d want to mount it at least one unit below the top of a rack in order to leave some air space around it and be able to easily access the connectors.

Feedback processors (suppressors, eliminators, destroyers — whatever the makers prefer to call them) are functionally fairly straightforward and have but one goal in life, which is to detect and control unwanted audio feedback. The connectivity, display and control capability is therefore fairly uncomplicated and the AFS2 has a simple, focused appeal, which — like the very best managers — provides everything you need and nothing you don’t.

The back panel has left/right balanced input and output connectors, presented in both XLR and TRS jack formats, and there’s a switch for each input channel to select input sensitivity between +4dBu (‘pro’) and –10dBV (consumer) levels. The channel 1 inputs and outputs are grouped together, as are the channel 2 connectors, which is actually useful in itself — there’s much less chance of accidentally swapping the channels (for example in a left/right main mix) when making connections in low light or where access is restricted. Even if the connections have to be made completely ‘blind’ (surely we’ve all found ourselves groping around in the back of a rack at some point!) then all you have to do is start at the connector nearest one end of the panel and work along in a sequence of in, out, in, out, and everything will be correctly channelled. There’s a standard IEC mains inlet at the other end from the audio I/O, but no on/off switch, so the unit will power up (and start passing audio) as soon as it’s plugged in to a live supply.

The AFS2 has two inputs and two outputs, which can be used independently or in linked stereo mode.The AFS2 has two inputs and two outputs, which can be used independently or in linked stereo mode.A new feature of the AFS2, compared to its predecessor the AFS224, is the provision of a USB socket for applying firmware updates — generally a comforting sign that the model will be staying in the current product range for a good long time, as new features and upgrades can be applied without replacing the hardware.

The front panel is neat, simple and effective, the traditional rows of filter LEDs and dedicated button array complemented with a small LED display screen and rotary selector, similar in size and style to those found on the 1U DriveRack models. A dual input–level meter indicates five levels from ‘signal present’ up to digital zero (max), with a clip segment above that. The two horizontal rows of red filter indication LEDs are large and bright, and easily visible from several meters away. On this new model there is only a single bypass/clear button for each channel, with all other parameters and preferences being accessed via the data wheel and screen. Bringing the AFS2 in line with the DriveRack family is the inclusion of a new automated setup routine, and there’s a big, inviting illuminated button smack in the middle of the panel bearing the legend ‘wizard’, which might as well say ‘press me, you know you want to’.

Wave Your Wand

The setup wizard first asks if the unit is to be operated in linked mode, where the same settings and filters are applied to both channels, or as two completely independent processors (for example when used on bus or channel inserts), in which case you have to go through the wizard twice. Once that little formality is out of the way you can get on with the actual setup, which guides you through every step and is very easy to follow. I like this feature as I know of several potential users of this unit who would appreciate this level of hand–holding.

A quick gallop through the wizard process gives an insight into what the AFS2 is doing and how easy it is to use, so here goes: after deciding to stereo link the two channels (or not), the next screen directs the user to “perform soundcheck and set up rough mix for all mics”, which is obvious and sensible, and of course you’d assume that you leave all the performers in place if possible as this will affect the potential direct and reflected feedback paths. The wizard then reminds you to bypass any active noise gates, and fully lower the main mixer outputs. The next step is to choose how many fixed and how many live filters will be used, and to make things simple the wizard asks if it should set the default number of 12 fixed filters. Choosing ‘no’ allows you to dial in the required number so is only one extra step, and then the filter width is chosen to suit the programme material and/or the severity of the potential feedback issues.

Now it’s time to gradually raise the mixer output level, and the AFS2 catches the offending frequencies and assigns filters to knock out the ringing as it starts to happen, and will assign up to the number of fixed filters you chose a moment ago, or you can tell it to stop when you’ve achieved a reasonable level, even if all the fixed filters haven’t been used. When the fixed filters have been set and confirmed as all done, the AFS2 switches on the live filters, which will swing into action when feedback is detected from that point on. The live filters behave in a different way to the fixed set, and are designed to differentiate between sustained musical notes and unwanted feedback. To demonstrate this, you can play a recorded track through the PA whilst the fixed filters are in setup mode, and even at low levels with no open mics the AFS2 will start assigning fixed filters as it ‘hears’ the music programme. If you try this when the unit is in live filter mode, it won’t mistake wanted musical content for unwanted feedback (at least I couldn’t make this happen) because it is using a much more sophisticated detection algorithm. Finally, when you’ve completed the wizard, everything can be saved in the user preset library and recalled for future use.

Live Test

I connected the AFS2 into our rehearsal studio PA and went through the setup wizard with four open mics to see both how effective the feedback reduction was, and what effect it had on the overall sound. I used the default setting of 12 fixed filters and, as expected, I could set a significantly higher level in the room with the AFS2 than without it. I tried connecting the unit as an insert and in line between mixer and powered speakers, and I found no noticeable difference in performance between the two setups; the AFS2 seemed to be tolerant of varying input levels, although the manual sensibly recommends using inserts as these will deliver a suitable and fader–independent drive — so long as you have a couple of input-meter segments lit up it should work happily. The increase in available volume was impressive, and even with all 12 fixed filters set the resultant sound was fine. Using the wider ‘speech’ filter option, playing recorded music through the system and switching in and out of bypass mode will reveal some tonal differences, but these are slight and would be insignificant in a live situation where you’re using feedback–approaching levels. What you have to remember about feedback suppressors is that, yes, they will alter the sound, but that’s exactly what you want them to do, and boy do they sound better than feedback! The other thing I noticed about using the AFS2 was that even at volumes well below the feedback threshold, using it made the PA sound better, as all the worst excesses of the room response had been filtered and the result was a cleaner, clearer output.

Our rehearsal hall for my 21–piece band is a bit of a nightmare for live sound, especially with no audience in place, so I tried the AFS2 on our vocal mics and was very impressed with the results. I used one channel of the AFS2 on mono mains and the other on a single floor monitor just in front of the two singers, and not only was it possible to achieve much louder levels, but by pushing the volume during setup to force the system to ring out the worst frequencies, the whole sound became more pleasant, cleaner and less fatiguing. It’s good to know that not only have the essential filters been set and will remain set, but the AFS2 is ready to deal with anything that might crop up during rehearsal or performance.

A neat feature is the ability to look at the main parameters of all the filters in place — a single press of the data wheel displays a representation of where they all are across the whole frequency range, and their individual information is displayed as each is selected with the wheel, so you can see exactly what the AFS2 is doing for you!

Overall Impression

Over the years I have tended to make much use of automatic feedback processors, as I find them a convenient and effective solution in the varied nature of live work I take on, and I’ve installed and specified them for a number of clients, who have always been delighted with the results. I’ve used the older Dbx AFS224 units on monitor buses and always found them to do a great job, and my DriveRack 260 incorporates feedback suppression, which I don’t hesitate to use if the situation requires it. I’m in the process of building up a new rig for larger indoor functions, which has to work for bands, conferences and A/V, and I will seriously consider putting three AFS2 units in there for monitors — the one thing my current digital mixer doesn’t have is feedback suppression! I know that not everyone swears by it, but it’s been my get–out–of–jail–free option on many occasions, and I like the flexible way these new Dbx units work, and I certainly like the new user interface. Given that the Dbx AFS2 so obviously does exactly what it’s supposed to — automatically detect and suppress feedback without knocking a big hole in your audio — what’s not to like?


Various stand-alone products with similar overall functionality are produced by manufacturers including Sabine, Shure, Peavey and Behringer.


  • Good setup wizard for inexperienced and occasional users — no manual necessary.
  • Good live filter discrimination between music and feedback signal components.
  • Very narrow filter widths available.
  • Fast and reliable in operation.
  • Flexible operating modes.
  • Great filter parameter display screen.


  • If you’re convinced you don’t like using feedback processors then you probably won’t like this either.


The AFS2 is easy to use, especially with the setup wizard, and buys you considerably more level before feedback sets in, without drastically altering your sound.


£379.99 including VAT.

Sound Technology +44 (0)1462 480000


Harman International +1 203 328 3500

Published July 2015