I'm new to this forum, but I've been a semi-pro musician for almost 40 years. I own the miniDSP 2x4 and the DCX2496, and have used both with my 'desktop PC audio system.' I say that in quotes because it's a bit beyond the typical computer audio system: a pair of Fostex NX-5A powered studio monitors and an SVS SB12-NSD sub mounted on an Auralex SubDude isolation platform. I use it for audio mixdown and personal listening, both at 'office' volumes and pretty wide open. My PC also includes an Echo Audio Mia PCI card, which provides balanced -10dBV and +4dBu analog outputs and an S/PDIF digital out.
My system is especially demanding, because I generally listen at -25 to 30dB levels, sitting 6-7 feet from the speakers, yet often crank the system all the way up when moving around between my office and lab. So, little noises that tend to disappear in the latter take center stage during the former.
I used posts from this forum as part of my decision to buy the miniDSP and SVS subwoofer, so I thought I'd give back a little to the cause, so to speak. I spent several days connecting the miniDSP and DCX to the system in various ways to get the best sound in the smallest/most convenient arrangement. Here are my thoughts, for what they are worth.
My starting reference was an unbalanced connection from the Mia to the stereo inputs of the sub, then out through its high-pass outputs to the Fostex monitors. The HPF outs are filtered at 80Hz, with an unknown filter type and slope; I adjusted the LPF on the sub to roughly correspond to the same frequency. Although the sub is arguably a -10dBV device, I found I could use +4dBu levels, with an appropriate cut with the Level knob. The result was pleasing, but just a little off in the mid-bass range, even with adjustments of the sub's Cutoff knob.
One of the nice things about the DCX is its real-time capability. While it certainly doesn't take long to reprogram the miniDSP, it's long enough to make objective comparisons of small changes problematic. (I can't speak for others, but my ability to hone in on minor changes with near-certainly only holds up for 10 seconds, and then only if I have no other distractions in between.) Since I wasn't worrying about noise floors and transparencies just yet, I decided to use the Behringer box to dial in the proper crossover points and coarse room EQ.
I focused on two songs for my 'coarse' tests: In the City
from the Eagles' live album, which has nice drum fills to evaluate the 'snap' of the sound; and Christina Aguilera's Make the World Move,
which has a boomy, intentionally obnoxious bass line that seems to love room resonances. A couple of quick tests showed the internal filters on the SVS sub to likely be LR-2 or LR-4 design. LR-8 filters tightened up the bass nicely at higher volumes. (Like many 5" studio monitors, the NX-5As get weird when driven below about 70Hz, even with some rolloff.) I also determined that while my office isn't perfectly flat at the low end, there were no glaring issues that the sub platform hadn't tamed. So, there was my starting point: 80Hz LR-8 crossovers, summed mono subwoofer audio, no additional EQ.
This is where the miniDSP shines. Setting it up in 'Rev A' configuration -- a strange name, as it only involves moving jumpers -- with the Mia set to -10dBV levels, I was able to obtain a great sounding 2.1 system. By comparison, even dialing up the input gain on the DCX by +15dB left the box running at halfway open, so to speak: -10dBV = -8dBu (roughly); add 15dB and we're at +7dBu, which is 15dB below the DCX's +22dBu ceiling; thus, we give up ~3 bits of digital data. While recognizing the target marketplace (live PA), it would have been nice if Behringer provided a -10/+4 switch on the DCX, as they do on many of their other boxes.
For an all-consumer level system, the miniDSP is the clear winner. However, I found some snags in my computer-based system:
- Running the miniDSP off a USB port produced an annoying digital -based background whine, similar to the sound of a data modem on a phone line. It was especially bad when I moved my trackball around the screen. This was a shame, since I would have liked to leave the USB plugged in to make changes over the next few days/weeks. If I didn't regularly use the system at -30dB, maybe this wouldn't be such an issue.
- Using an external adapter (Motorola RAZR charger I had lying around in the drawer) cured this issue, but left a 60/120Hz background hum in the Fostex monitors. I'm not sure if this can be blamed on the miniDSP -- or on the -10dBV inputs to the monitors -- but it went away when I unplugged the RCA inputs to the miniDSP.
- I had concerns (below) if consumer levels was compromising the output of the Mia.
I then turned the Mia up to +4dBu levels, to better utilize its capabilities. (Echo claims that the -10 soft switch doesn't result in any lost DAC bits, but it certainly moves the audio levels closer to the noise floor of the computer.) At this point, things got complicated pretty quickly ...
The miniDSP 2x4 doesn't truly have a 'pro' level setting. Rev A is 0.9V (max.) in, 0.9V out; Rev B is 2.0V in, 0.9V out. So, I could use a higher input level, but the output was still set to consumer levels. Then, we run into the second limitation: headroom --
-10dBV = 0.316V
0.9V = -0.9dB
So, there is roughly 9dB between the nominal output level and the miniDSP's maximum input, which covers transient peaks in the program material nicely. At Rev B, though --
+4dBu = 1.23V
2V = +8.2dB
Rev B only has about 4dB of headroom. I found that when at +4 settings, I had to drop the Mia's master volume by 4-5dB, or I could hear digital clipping artifacts in really loud passages. So, really, I was running at 0dBu, not +4dBu.
On the other hand, 10dB input gain on the DCX provides +14dBV nominal, with 8dB of headroom. For professional levels, one either needs the balanced version of the miniDSP 2x4, or to stick with the DCX. In my case, the built-in XLR outputs of the DCX made for a simpler connection to the monitors, and the sound cleaned up nicely -- whether from balanced wiring or higher output is unclear.
Now, the balanced miniDSP is still only $165, shipped + $10 for the software license. Right around the NAMM show, though, there were lots of dealers offering the DCX for $240-250, delivered. This brings the decision down to less than $100, while delivering some extra features (real-time control, built-in digital audio) that the base miniDSP leaves out. For what it's worth, I found the DCX's sound at high volumes slightly more 'open,' particularly on guitar-heavy tracks, although this could simply be rationalization on my part. (It's difficult to listen closely to really loud music for extended periods without losing some focus, not to mention frequency range!)
DIGITAL IN, PRO ANALOG OUT
This was probably my end goal all along, since the Mia provides S/PDIF while the rest of the system is (mostly balanced) all-analog.
To achieve this goal, one needs to add the miniDIGI board to the miniDSP 2x4 Balanced, and then to be willing to build a case for the end product. I didn't want to go this route, so my results represent the DCX path only. I suppose one could also by the miniDSP 4x10HD, but that adds extra functionality at a significant price bump, which would have been overkill for my needs.
Although the DCX has AES/EBU inputs, it can accept S/PDIF data, if the cabling is rather short. I did some research on the Web and concluded that most low-end 'S/PDIF to AES/EBU adapters' are merely a decent-quality S/PDIF cable with an XLR connector substituted on one end. I decided to try a 3-foot 75 ohm S/PDIF cable and an inexpensive ($4-5) RCA-to-XLR adapter, which works very well. The Mia sound driver has a choice between 'consumer S/PDIF' and 'professional S/PDIF' -- in their words, the definition of the copy-protect bit -- and I chose Professional. The monitor outputs were direct XLR balanced at both ends, while the subwoofer used an XLR-to-RCA cable and sufficient dropping of the Level knob setting.
Other than an accidental first setting to analog inputs, the connection was flawless. Using digital data, the DCX maps it to the full range of its internal data path, so input level matching is no longer an issue. I found the DCX's internals to run *very* close to the 'rails,' and an input pad of -1dB cleared up occasional clipping artifacts in the system. The outputs to the studio monitors also run at -1dB, just in case -- I don't like to see the peak LEDs kick in, even occasionally -- while an extra -2dB of padding allows for a better match at the sub's -10dBV inputs. Overall, I'm giving up -2dB from the ideal case on the monitors, and -4dB on the subs -- 1-bit or less from the 24-bit data path.
SO, there you have it. My results won't necessarily align with some readers' needs, given my unusual analog-digital mix, but it highlights some of the pros and cons of the miniDSP 2x4 and DCX2496. Overall:
- For an all analog, all consumer (-10dBV) system with an unbalanced signal path, it's hard to beat the miniDSP.
- For a professional (+4dBu) all-analog signal path, the results aren't perfectly one-sided, but tend to favor the DCX.
- For a digital in, professional analog out system, the DCX really shines.