DEQX v. passive transparency?

[SashaV]

I have seen some posts on other forums claiming that DEQX kills low level signal and does not sound as transparent as well executed passive cross-over. Apparently those low level spatial clues are lost with DEQX and according to some it is better to have MF/HF crossover in a 3-way as passive. Anyone with experience in DEQX transparency?

[Norman Tracy]

I have been a DEQX user for several years now, the 2.6P. My opinion remains that once you go DSP active with a multi-way speaker's crossovers there is no going back. I find it almost shocking that writers “on other forums claiming that DEQX kills low level signal and does not sound as transparent as well executed passive cross-over” because my experience is the exact opposite. In my forty years as an audiophile transparent true to the source reproduction has always been among my highest priorities in music playback. Passive crossovers are basically placing a bunch of complexly nonlinear filter elements between amp and drivers vs. the mathematical precision of a DSP filter feeding amps directly connected to the drivers? And in the case of DEQX's implementation we also get the phase correction option that allowed me to finally get to my goal of the coherency of a full range electrostatic with the dynamics of multi-way moving coil driver based system.

As to why some would claim “DEQX kills low level signal and does not sound as transparent” I would speculate the following:

1. Analog bigots, their belief system is if its digital it cannot be as good as analog. My experience is one of the better sources I can use with my DEQX is one of my turntables+phono amp plugged into its analog inputs.
2. Poorly programed DEQX. The setup learning curve is steep and long, as is designing your own passive crossovers. A poorly programmed DEQX system will of course be a little or long way behind a competently designed passive cross-over system.
3. Lackluster digital source. I have observed my 2.6P when used with its digital inputs is rather sensitive to the quality of the S/PDIF or AES/EBU data source. For example there is a loss of transparency when using my Oppo BDP-80 Blu-Ray player compared to the lower jitter sources like a Denon DVP-2900 universal player and my computer USB-to-S/PDIF converter. This same mistake is so common in the countless LP vs. Digital debates that will always rage on the Internet. We often see OK but hardly top rank digital sources being trotted out as the straw men to 'prove' their argument that analog rules.

[analog_sa]

Originally Posted by SashaV

Anyone with experience in DEQX transparency?

There cannot be a universal answer to this question. It will depend on the particular implementation, your system and your ears. Can you reliably hear opamps in the signal path? Capacitors? Difference between dacs? If yes then maybe you won't find DEQX completely transparent. And this does not even take into account any potential issues in the digital domein.

Does it even matter? Audio is all about compromise and this is a good example where compromising may actually bring better sound. Or not.

For me it is clear that DEQX is an essential tool for anyone with serious interest in audio. It is also clear that good results cannot be had cheaply or easily, it's just not possible.

[gornir]

Hi SashaV,

I would say that’s true with the older version of the DEQX (PDC-2.6P), but the new HDP-3 pre-amp processor is a different story. They share the same functions and features but I think the PDC-2.6 sounded mediocre and definitely not high-end, but the HDP-3 on the other hand is a true high-end product where you don’t lose any detail or transparency. However its sound is very neutral to me.

To get the best sound with the DEQX, my advice is to use as low order slopes as possible e.g. 48db/oct it sounds better as the default 96db/oct or higher. Another important factor to get a success with the DEQX is that you have to treat the loudspeaker driver possibilities and weakness just as when building passive cross-overs. You still have to consider baffle step, baffle diffraction and off-axis behaviors when building the active cross-over as well as the passive. Look at this post to see what I mean. SU551-RS28F – a 2-way DEQX system loudspeaker measurement study

Regards

/Göran

[ShinOBIWAN]

Originally Posted by gornir

They share the same functions and features but I think the PDC-2.6 sounded mediocre and definitely not high-end, but the HDP-3 on the other hand is a true high-end product where you don’t lose any detail or transparency. However its sound is very neutral to me.

I've had all three over the years. My first DEQX was a 2.6 then later the HDP3 and more recently I've had the opportunity to try the Express alongside the HDP3.

FWIW, the Express is based on the 2.6 with some tweaks such as different opamps and dac's. It still however has the SMPS that a lot of folks claim really held back the old 2.6.

My experiences have shown, for me, there's very little difference between the Express and HDP3 which is a downer considering the HDP3's price tag. The only thing I can really hear a definite difference between the two is the sound of my partners cry's about costs

Seriously though if you can absolutely pick out a definite difference between these two then you've got better ears than my 32year old ones. I'd like to think I heared differences but being honest its tough. When I think I can hear one doing something better than the other I'll repeat the test again and the difference is questionable. If pushed I'd say the HDP3 can sound a bit more spacious on acoustic pieces and maybe the highs are smoother at times.

My advice is if your looking to buy a DEQX then absolutely go for the Express over the HDP3 if your even slightly worried about budget. However if you like a better built unit, the HDP3 gives you that at a price.

Regardless of this and going back to the original question I'd say once setup optimally any losses through digital conversion and processing are more than offset by the additional frequency and phase accuracy offered by the DEQX.

In short I'm secretly wishing passive would decline in popularity to allow for surge in digital filtering research and DSP related products but like the immortal vinyl I doubt that will happen in my lifetime

[bigfishe]

I've become one with my DEQX

[ShinOBIWAN]

Originally Posted by gornir

To get the best sound with the DEQX, my advice is to use as low order slopes as possible e.g. 48db/oct

Goran its possible to use as low as 12dB/oct filters(not sure why there's no 6dB/oct option) with the DEQX but understandably these aren't FIR based. The minimum you use with those are as you stated - 48dB/oct. Personally I'm somewhat on the fence about high vs. low order filters. They can both sound terrible so you've got to take it on a per design basis. Generally I'd either go 2nd order if I have smooth extended and low distortion drivers and steep filters for almost everything else - I've found there's no real in-between there for the designs I've done. Its either all or as little as possible.

[gornir]

Originally Posted by ShinOBIWAN

Goran its possible to use as low as 12dB/oct filters(not sure why there's no 6dB/oct option) with the DEQX but understandably these aren't FIR based. The minimum you use with those are as you stated - 48dB/oct. Personally I'm somewhat on the fence about high vs. low order filters. They can both sound terrible so you've got to take it on a per design basis. Generally I'd either go 2nd order if I have smooth extended and low distortion drivers and steep filters for almost everything else - I've found there's no real in-between there for the designs I've done. Its either all or as little as possible.

For crossing between mid and tweeter I’ve never used anything else but FIR filters. However between subwoofer and mid I’ve used conventional filter topologies.

Yes, I agree each system is unique and it pays off to extensively test different cross-over frequencies and slopes when building a speaker with the DEQX and fortunately it’s a very easy task to do with it.

In my experience, if you have well behaved drivers with no or little cone break-ups they sound better with low order slopes. If you have e.g. metal cone drivers with nasty cone break-ups you must obviously use steeper slopes. To be honest it’s not a huge difference in sound quality between 48db/oct, 96db/oct or >96db/oct, but it’s noticeable. I think the DEQX Team officially recommending slopes between 48-60db/oct.

For best results with the DEQX, avoid drivers with nasty cone break-ups.

ShinOBIWAN, have you tested hybrid filters with the DEQX, a hybrid filter where perhaps you have some contour shaping passive filters, but where the final tuning is done with the DEQX?


/Goran