The elysia xfilter 500 is a stereo linked equaliser in the ever popular API 500 series format. Analogue stereo equalisers are relatively rare, even more so in this compact format, so it’s nice to see this German audio company putting their expertise into such a useful audio tool.
Straight out of the box, the xfilter 500 slotted smoothly into my API lunchbox with no fuss. Extensive use is made of surface mount components and this is undoubtedly what makes a full feature stereo equaliser possible in such a compact format.
I was a little concerned at first about the lack of any metal shielding plate between the pcbs and any other adjacent 500 series modules mounted into the rackframe, however elysia have taken the precaution of dedicating one of the four layers of the printed circuits boards to be a dedicated ground plane. This goes a long way to ensuring lack of pick up from either stray hum within the rackframe (a known problem on some rackframes with integral PSUs) or crosstalk from adjacent modules.
Inputs and outputs are electronically balanced and the audio is 100% Class-A which is obvious in the complete lack of any measurable crossover distortion in the signal path.
The equaliser is presented in a 4-band semi parametric format with an additional passive sheen EQ and some clever switchable options within the 4 bands.
My notes say that the xfilter powered up fine, no problems. Very attractive back-lit logo which I was convinced had to be the power switch … but it’s not. (Still great to look at though!) Large rotary knobs are very easy to set, pushbuttons smooth, white markings on a dark blue background nice and clear but quite small.
With the unit powered but not switched in, a true passive bypass is offered. I really like that on an equaliser because with so many controls affecting response it’s nice to know when it’s ‘in’ and when it’s ‘out’.
Let’s deal with measured audio performance ‘on the bench’ first.
The xfilter 500 is a low noise, low distortion, wide frequency bandwidth equaliser.
On my Neutrik test set, output noise measured below -96dBu unweighted across the audio band, so well in line with the quoted A-weighted spec of -98dBu.
Distortion was very low, measuring at around 0.002% @ 0dBu rising to 0.005% when working at +10dBu levels.
Bandwidth goes from near to dc right up and beyond 200kHz. I measured the -0.5dB point at 135kHz. I have mixed views about very super extended bandwidths as they can at times encourage the transmission of interference within the audio chain, but let’s just say that those of you who work at digital samples rates of 192kHz won’t find the xfilter to be the limiting factor within the bandwidth of your recording system! To be clear, the noise spectrum of the xfilter was completely free of any nasty spikes or erroneous high frequency garbage. A very nice result.
Turning our attention to the equaliser bands;
The Low and High bands have a really nice feature. As standard they have a shelving response with selectable turnover frequencies and boost/cut of ±16dB. However they can be switched to become low and high cut filters respectively with a slope of 12dB/octave. That’s a nice feature, however there’s more to come. In this mode, the boost/cut control now affects the resonance, or damping factor, around the selected knee frequency. This allows you to dial up a variable resonance peak just before the usual fall off response comes into play.
The two mid bands provide sweepable peaking filters, with a good range of overlapping frequencies covering the useful audio band, and offering a boost/cut range of ±13dB. The mid bands also have switchable Q, between 0.5 and 1.0
Finally, coming to what elysia refer to as the ‘Passive Massage’, we’ve got a nice sheen filter with a very gently rising response from around 3kHz, peaking at just over +2dB at 12kHz and then falling away fairly quickly (-4dB at 20kHz) out of band. This is a classic LC (Inductor/Capacitor) combination and should prove very popular across many stereo mixes.
Let’s remember this is a stereo equaliser with a single set of linked controls. If we’re going to use it across a stereo subgroup, or the final mix buss, then it needs to track well between the left and right channels. So how does the xfilter 500 perform in this regard?
All the rotary controls are detented pots with 41 positions. Although there is no actual ‘0’ point marked on the boost/cut controls, I worked out that there were four ‘clicks’ between the -1 and +1 marks on the front panel. So I carefully set each of these to the assumed ‘0’ position, did a frequency sweep, and was very pleased to see a flat response within 0.5dB across the whole audio bandwidth on both left and right channels.
Next test was to boost at 1kHz setting as close to +6dB as the stepped pots would allow. Left channel measured +5.7dB and the right channel +6.1dB, so again within a 0.5dB tolerance. Boosting the gain up to +12dB actually reduced the left/right mismatch to within 0.1dB, perhaps because the pot was near the end of its travel.
Then I set the frequency and boost/cut controls randomly to a number of different settings and did a set of test measurements of level match between left and right channels. I found one instance where the mismatch was 1.1dB but in the majority of cases it was under 0.4db. This is a great result for a potentiometer based design. Elysia have told us that this is achieved by a combination of computer aided selection of potentiometers, to get good matching between the gangs of the 2-gang and 4-gang potentiometers required in the design, and in the use of high tolerance capacitors.
I set all controls back to ‘flat’.
In my opinion, the best run out for a new stereo EQ is to put it across some stereo submixes and across the 2-buss, and see how it handles a variety of unmastered mixes which haven’t yet been polished either with EQ or compression.
A very beautiful female vocalist (I’m talking about her voice!) with four piece backing band. Whilst the vocal has been nicely recorded, it was a little bit (sorry for the adjective) ‘honky’ in the mid range. I thought previously it was just a bit too dynamic but compression didn’t work at all. So in came the xfilter and I found that a 2.5dB cut at around 1.7kHz, with the wide Q setting, worked really well. This brought the vocal forward in the mix whilst vastly improving tonal quality. Good result!
Aggressive stereo drum submix. For all of its drumming energy (you could hear this guy having a great work out!) it was definitely lacking in dynamic range and may have been (unknown to me) ‘out to a long lunch’ with a compressor. So what could be done with EQ?
First I focused on the snare. With a bit of experimentation, applying the xfilter’s low mid band, I was able to bring the snare back to life with a 3dB boost at 400Hz and a narrow Q. This gave a much better balance against the bass drum which I now turned my attention to. The very bottom of the low end was already a bit too extended yet the bass drum sounded a bit thin. I switched the xfilter’s low shelf section into the alternative Low Cut mode and set the ‘gain control’ to +8 which, in this mode, brought into play the resonance boost feature. Then I swept the frequency knob till it brought about a dramatic thickening of the bass drum (the frequency setting which worked best was 100Hz) whilst at the same time low cutting the very bottom end, which was much improved. Job done!
So this time, a very nice Jazz quartet with Guitar, Bass, Drums and Organ. Even on my extended bass monitoring speakers, the double bass was felt rather than heard and I expect it would just about disappear in a typical listening set up.
I dialled up a massive (for me) 8dB of low shelving boost starting at around 90Hz. Still a bit over clean for my liking but better definition and now better balanced in the mix. I found nothing to be gained by boosting above 90Hz within the mix, so in this case I would have gone back to the multitrack and worked on the double bass on its own. However it was improved.
At the other end, the ride cymbal sounded a bit plasticy. This was a really nice application for the xfilter. I found the frequency sweet spot to give the ride cymbal some body (around 4.5kHz and high Q) and it needed only 1.5dB of boost to totally transform the sound of the ride cymbal within the mix. Nothing else in the mix was at all bothered by this small level of boost but I was very impressed how the xfilter 500 had really worked some magic.
A real problem mix. Hugely busy soft rock mix with doubled lead vocals, stereo flanged guitars, slide guitar, Fender Jazz bass guitar doing its rock thing, distorted lead guitar … you get the picture!
Well surprise, surprise, the mix was so busy that it just sounded incredibly flat. After a couple of listens the vocals seemed to disappear more and more into the mix and the words ‘wallpaper music’ came to mind. BUT it was a really good tune.
When I’m not sure where to start, I turn up the mix real loud and see what’s hurting most. Sure enough, the vocals were on life support and fighting for survival against the rhythm guitars.
I set the xfilter’s low mid at around 6dB cut with wide Q and started sweeping the frequency knob. I settled on 1kHz and backed off the cut to 4dB. Now the vocal came forward in the mix, lots of improvement in definition (I could actually understand the lyrics for the first time!) and a lot of the guitar grunge was brought under control. So now the whole mid/bottom gelled nicely.
What we had now was a nicely balanced mix just lacking a little bit of high end sheen. I tried a little high shelf boost but even at low gain settings I couldn’t find a turnover frequency which was subtle enough. I reached for the ‘Passive Massage’ button. In this bass/mid dominated mix, it worked well.
First thing to say is that the elysia xfilter 500 is a very classy sounding equaliser. The filter slopes and the two mid band Q settings are well chosen and work very well across a very wide range of musical styles.
The switchable high and low pass filters with adjustable resonance are a real bonus and I began to understand how useful they could be as an alternative to the high and low frequency shelving filters.
Passive Massage needs to be used with care. It’s tempting to punch it in and assume that every track or mix will be benefit. I found it worked best on tracks with limited high frequency energy. I would certainly consider using it to put back high end lost during compression.
I don’t know when I would ever need the range of boost and cut available in the xfilter, but with the stepped controls and large control knobs, it was easy enough to operate over the typical 3 or 4 dB of range needed.
I consider that the very low colouration of this equaliser is a major reason for its ability to tune into a particular element within the mix without disturbing other mix elements. The ability to bring clarity to vocals within the mix was particularly outstanding.
Although the front panel layout is simple and logical, it would have been nice to have some better visual clues as to each knob’s function. Yes it’s all labelled but you will know how important it is to be able to reach and adjust, as the thought of what might work comes into your head. To an extent this is one of the compromises of small module formats like the 500 Series.
If you need a really high quality stereo EQ for stem and 2-buss work then you should audition the elysia xfilter 500. And of course it will work its magic across a whole range of mono and stereo sources as well. At the typical selling prices I’ve viewed on-line, I also consider it fantastic value for money for such a fine equaliser. Highly recommended.
Measured Audio Performance (Neutrik TT402 Test Set)
Output Noise 22Hz to 22kHz
Left -96.2dBu unweighted
Right -96.1dBu unweighted
Maximum input level
Maximum output level (fed +10dBu input then used boost until onset of clipping)
Crosstalk (one input fed +20dBu, the other input terminated)
Left to Right -94dB at 1kHz -92dB at 10kHz
Right to Left -96dB at 1kHz -92dB at 10kHz
THD+N @0dBu I/O, 22Hz to 22kHz 0.003%
THD+N @+10dBu I/O, 22Hz to 22kHz 0.005%
-0.5dB at 15Hz and 135kHz
Stereo Left/Right tracking
All boost/cut controls set to ‘0’ better than 0.5dB
At all other settings of controls typically better than 0.5dB, worst case 1.1dB