top of page

October 2019,

​

​

The MiniDSP line of products are difficult to explain.  (for picture of the 4x10hd internal, see into "project tab")

 

Depending on the range, they start from an entry level and run to mid-HiFi products, but never better than Mid-HIFI, so sorry to disappoint. They all come with average specifications (most of them aren’t published, like noise figure and distortion). But they have a great price appeal to the DIY’er, and due to their intuitive ease of use and graphical interface, MiniDSP’s  are selling like hotcakes.

 

A few years ago, I jumped on the bandwagon and bought a MiniDSP 2x4 (the higher grade 2x4HD didn’t exist then). My plan was only to use it with my subwoofer. Back then, this was exploratory and not much more. Willing to experiment for less than $200CAD, wasn’t a big risk either.

 

Poor dynamics, loss of resolution, smudge on the perceived sound, I could go on, but it was bad so no point to hammer it down.

Reality is that that the MiniDSP 2x4 sounded very bad, but it did allow the user to correct for bad room acoustics.

​

A tricky trade-off…Win on room acoustics and loose on conversion sound quality.

Which one would prove to be the most important compromise to accept?

Turned out that despite a dismal sounding DAC inside the 2X4, the DSP enabled subwoofer did sound significantly better than a passive crossover, with no room equalization! (and a $2000 high end DAC feeding it)

DSP correction DOES trump pure DAC conversion signal quality, hands down… every time. This lesson alone was worth the $200 invested, and immediately reset my course towards a fully DSP system.

 

I was instantly hooked on room correction and all the possibilities the 2x4 enabled me to do. The MiniDSP adjustments provided more uniform/smoother results in the bass than any passive solution, and its associated “hope it’ll work” acoustics.  Trial and error time was also greatly reduced. Why bother for uneven bass, when few clicks of a mouse correct that nasty summation node at 40 Hertz? If only the D/A conversion inside the 2x4 wouldn’t be that bad, it’d be so much better and retain the dynamics as well.

 

I quickly re-sold the 2x4 and bought a 4x10HD, which was top of the line for MiniDSP’s, thinking that it was enough for my personal nirvana. (Dirac* newer series didn’t exist then)

​

-----

*SIDE NOTE on Dirac:

Dirac option is a great selling feature in a sense that the DSP programming functions are semi-automated.

The user just has to follow the steps by step process and the software calculates everything automatically.

No mistakes, no learning curve, what a great idea that is. Simply trust the computer. I wouldn’t touch it with a 10 foot pole.

​

Here’s a simple car analogy to explain how I see it:

-Average car driver performances are enhanced by driving aids (traction control, Electronic stability control and ABS brake).

-The car almost prevents a crash... keeping you safely on the road.

-Professional race car drivers are faster with the aids turned off, because the aids are programmed well below the car’s actual limits.

-The car driving aids are slowing the professional driver down... keeping parameters within safe limits and having a safety margin on top of that.

 

-The danger is that the driver without any electronic aids, risks hitting a wall if he exceeds the car’s limits.

You just have to enter into a too fast once, to know what your real skills are.

​

Therefore:

1-If you are a beginner DSP user, chances are 100% that the Dirac software will obtain better results that you could on your own.

2-If you are an average DSP user, you could see your odds at maybe 40% to outperform the software results. 

3-If you know what you're doing, the computer predictions aren't cutting it, you will beat them every time.

I consider myself in the 3rd category.

​

If you are in the 1st or 2nd categories, nothing prevents you graduating to the 3rd. It just takes time and commitment.

"Practice makes perfect" 

----

 

The bad news is that the 4x10HD sounded almost as bad as the 2x4. Obviously, there is hope and in direct comparison to the 2x4, the 4x10HD is a tad better…but not much, as dynamics are still missing big time. I’m not making this up. A cheap $80 DAC on eBay completely destroys the stock 4x10HD. Conduct an A/B blind test against most DAC’s you might have in mind, and you’ll identify the 4x10HD every time. Most cheap DAC’s surpass the 4x10HD by a fair margin. This is a bit frustrating and probably the reason most people don’t keep their 4x10HD’s forever.

 

Most of my complaints are that the 4X10HD has four electrolytic capacitors in the signal path, which greatly colors the sound. Because they are all polarized, they tend to sound super dull or “polite”. Moreover, due to a cheap operational amplifier used inside, distortion is high as per the average DAC converter specs. The DSP chip used in the 4x10HD is also quite noisy…relative to most DSP’s on the market. Of course, it’s only $500 but still, I was expecting much better. Close inspection revealed the MiniDSP team prioritized cheap over quality on every part they used. Even the chassis is thin metal. Quite thin indeed!

 

The 4x10HD stock version does have high frequency noise (centered at about 11 kHz) and some hiss, which is often a major complaint from online reviews, which I sadly confirm.

The MiniDSP corporate response for the “noise issue” is that users have to “optimize the gain structure” which is explained in the link below.

https://www.minidsp.com/applications/dsp-basics/gain-structure-101

Of course, however…it’s not displayed on the product page. Once you purchase the product and then experience the noise, you’ll find supporting comments through online forums of how to reduce the noise and maximize performance. NEVER, EVER, EVER expect a dead silent system if you use a 4x10HD. If you ask me, this is a deception for the eager buyer.

 

Only when gains are all optimal, the hiss becomes tolerable. Not fully inaudible mind you… but quite tolerable.

In my setup, I could barely hear it at the listening position. I knew the system was on, but once the music played, it was drowned out.

All vinyl playback systems also exhibit noise. The 4x10HD, when optimized, will have less noise than most analog set ups, but much more than an average CD player:

  • Vinyl playback noise is constant. If you change the volume on your preamp, the noise goes up or down at the same rate as your music

    • the vinyl surface noise is built-in with the signal, so attenuating it is impossible without having detrimental sound quality consequences.

  • Digital noise on the MiniDSP is fixed. If you increase the output volume, the background noise remains constant, but the signal is increased (and is therefore less audible relative to the musical signal)

    • only when your digital volume is set very low, does the hiss become too strong (relative to the music signal)

    • this is why optimizing your gain structure is so important. 

 

If a MiniDSP user knows how to modify the feedback loop in their amplifier, thereby altering the amplifier gain, he could get the gain set perfectly for the right balance. This is exactly what I did. However, most people can’t do that.

This means they must resort to using “in-line attenuation” to do the job. See below:

https://www.parts-express.com/harrison-labs-12-db-rca-line-level-audio-attenuator-pair--266-244

 

There’s nothing wrong with RCA or XLR in-line attenuators, but it’s more of a band-aid than a permanent solution. They will raise the impedance the amplifier "see" deteriorating the impedance the match between a preamp and the amplifier. Ideally a preamplifier output impedance should be a as low as possible. An inline attenuator breaks that. Due to the resistive divider network inside these products, they raise the effective preamp output impedance seen by the amplifier and tend to kill subjective dynamics. This will render an already polite 4x10HD sound a bit more polite. Not good, but significantly better than hearing constant hiss.

 

Subjectively, as previously mentioned, the 4x10HD sucks IMHO. Buy an $80 cheap eBay DAC and use the 4x10HD as a DAC with the RCA output (no EQ, just use as a pure 2 channels DAC). Compare the two, and 99.999% of the time, the cheap eBay DAC will spank the 4x10HD all day long.

If you use the 4x10HD and implement room correction, the 4x10HD will quickly pull ahead of the DAC, but that’s true with just about any DSP vs “pure, uncorrected analog” systems. I know I repeat myself, but I want to be sure you get the point across that the 4x10HD is not HiFi. It’s an audio tool.

 

 

How to get the best of the 4x10HD?

1. Know how to use it.

2. Modify for better performance.

 

Tips for best use of the 4x10HD.

  • Input analog signal should be shorted if not used. This is because the 4x10HD always ‘listens’ to the analog input, and if you only use the 4x10HD as a DAC fed from a computer, some analog noise will enter from the analog input reaching the AD stage. Treated by the DSP, the incoming noise will be sent to the output as if it was part of the signal chain. You don’t want that. The best option is to use a physical wire to short inputs, or at least an RCA shorting device, and set input gain to minimum at the plugin level. Doing so mitigates the problem.

  • If never using the analog input, one can disconnect the analog input at the DSP, as it only requires removing two coupling capacitors and will prevent the noise from the analog section reaching the DSP chip. I did it on mine, but reversed the modification prior to selling it.

  • Never amplify the frequency or use any digital gain. This ALWAYS creates clipping distortion in the signal. The 4x10HD has no overhead, and if the input signal from a digital source comes in at 0 dB (100% signals from a CD for example) and the 4x10HD is set at +6 dB gain, guess what… You’re clipping badly and it’ll be very audible, resulting in possibly blowing your speakers, if you play this distortion long and loud enough. If the distortion is in the mid-range, it’ll be evident, but if it’s in the bass, it may not be detected by your ear, but the DC offsets the clipping sending it to your woofer, potentially causing problems as it heats the voice coil and creates thermal compression.

  • Same goes for the DSP function. For example, if you put a +5 dB boost @ 50 Hz, and your output gain (master volume) is set to 0 dB, a full signal at 50 Hz will clip. Use a test CD and check it out. It really does. That flaw isn’t published by MiniDSP, and the 4x10HD will constantly clip as most people set their gain higher that they should.

    • Note that most CD "test tracks" are at recorded at -20 dB. To experience the clipping I describe, one would have to set the gain at +25 dB, but then remind yourself that the music you buy on CD is recorded at a range varying from -1 dB to -5 dB, thus leaving no headroom ,and often triggering the clipping problem.

  • The way around these two problems is to configure the 4x10HD properly. If you need +5 dB gain at some frequency, then set the channel overall gain (of master volume) to -5 dB. This results in a net gain of 0 dB at your troubled frequency and -5 dB for the remainder, and you’ll be good to go. Of course, you’ll have to re-adjust your “gain structure” as your new overall gain will now be lower. But it’s the only way to do it.

  • My own test proved the 4x10HD already clips when a 0 dB test signal is fed into the DSP (from a digital source or a test tone) while the output is set @ 0 dB. Setting the overall output gain to -4 dB is the maximum you can safely use.

  • Never use the MiniDSP at more than -4 dB for any frequency, or you’re begging for clipping distortion.

  • At -4 dB maximum output, you’ll never clip, even if the source is reaching 100% signal. Then you can re-adjust your gain structure to play loud enough for your system’s overall gain and speaker efficiency. Too high of gain and will reintroduce the good old PITA hiss mentioned earlier, with the balance being thin… very thin.

  • This limits the MiniDSP adjustable gain flexibility, but it’s the only way to use it ‘clipping free’.

  • If possible, use the AES digital input first, as it sounds the best (its a balanced signal). The SPDIF (RCA, single ended signal) is next input with the TosLink being last (optical signal).

  • The quality of the USB/SPDIF converter (sometimes called a bridge) matters a lot. Try a few and swap out the cables, as they influence rendering as well.

 

 

2. The only way to get the best of the 4x10HD analog section is to modify it. Elsewhere on my site, I have pictures and few comments on previously owned 4x10HD’s. I’ve sold them since, but my modifications boiled down to the following:

  • Out of four capacitive coupling capacitors stages per channel, two can safely be removed by shorting the pin and allow a direct signal to pass. This restores much of the dynamics, reduces coloration and renders a more accurate sound.

  • Upgrade the opamp. The XLR output has 16 opamps (8 on top and 8 under the PCB). The XLR signal feeds the XLR output. The same output signal is then taken from the ‘+’ XLR signal and feeds to four more opamps which provide a buffer function for RCA output. I’ve never fiddled with the RCA ones, as I’ve only used the XLR outputs myself.

  • All 20 opamps used in the output stage have no real power supply bypass. Adding a quality 33µF or 47µF electrolytic capacitor to each PSU rail performs magic in the dynamics and increases the perceived weight of the crescendo. Going to higher values than 47µF isn’t really required, and the 1/8 watts surface mount 10Ω resistor on each opamp power supply line might blow on start up (as it try to charge the big capacitor).

    • If you insist on 220µF and higher value bypass PSU rail capacitors, then simply replaces the surface mount 10Ω resistor to 0.5 watts ones or bigger (keep 10 Ω as a value, as this resistor is needed).

  • SMPS internal to the 4x10HD has regular diodes which provide +15Vdc and -15Vdc that feeds all the analog opamps. Upgrading with Hexfred (faster diodes with slow recovery) diodes does help with the noise and minimizes EMI. And for only a $4 investment, it’s a no-brainer.

  • The general choice of surface mount capacitors is poor, for both the 15Vdc analog output section and the +12Vdc, +5Vdc and +3.3Vdc. Upgrading these to Panasonic FM (low ESR, long life type) is a great idea.

  • If an external power supply is to be used (highly recommended and the 4x10HD reacts superbly to such an upgrade), increase the +12Vdc input voltage to a ~12.6Vdc for even more power reserve. Note, that too high of a voltage may blow something inside, so I wouldn’t push it too much. I used +12.6 Volts myself (5% extra juice) and it worked great. If you go to +13Vdc and above, then it’s at your own risk. I did not try it.

  • Damping, damping and even more damping, inside the chassis. Either Dynamat or any equivalent will work. The sheet metal enclosure is thin and far from ¼” thick aluminum front plate seen on higher end products. Damp it and reap the sonic benefits.

  • Mounting the 4X10HD on aftermarket feet is also a good idea, I had good luck with DIY Bamboo feet, but depending on your equipment rack and floor type, you may want to experiment several feet style (rolling balls, spikes, wood or Sorbothane).  Isolation products aren’t a predictable science. Trial and error is your only solution.

  • If analog inputs are used (i.e. for a phono input), bypassing those four input capacitors is a great idea (stock are the average Nichicon KW series capacitor). The second stage of input capacitor is used to isolate the analog input stage to the ADC and a voltage differential of +2.4Vdc is present. Those capacitors must remain, but upgrading them is also a great idea. Use your favorite brand as those Nichicon KW aren’t especially great.

 

3. Advanced modifications that would make the 4x10HD even better (I was planning it but sold my units before I tried it)

  • Remove the internal SMPS and use external regulated linear power supply to feed directly the sub sections of the MiniDSP. This would be even better than an external power +12Vdc power supply, no matter how good a PSU you use.

  • You’ll need a +15Vdc, a -15Vdc for the analog stage

  • +12Vdc (not sure where this +12Vdc is exactly used, and whether its critical).

  • +5Vdc for most digital chips. (that one IS critical)

  • +3.3Vdc for some other logic chips.

  • Onboard USB/SPdif conversion. No more need for the external USB bridge, and saving one critical digital cable. A simple add-on board and that’ll also allow removal of one digital transformer to the signal (reduces jitter).

    • Quality USB/SPDIF kits exist on the market, and one simply has to connect it inside the 4x10HD.

    • Amanero Technologies would be such product candidate to investigate.

 

 

With the 4x10HD used intelligently (no clipping from boosting signal as per text), optimal signal gain structure and the modifications described above, the 4x10HD performs very good.  I’ve had mine for about four years before upgrading.

 

But I was on the hunt for upgrading as I need more than 8 analog channels. It was only a matter of time before I had to replace it.

​

***

​

Follow up:

I bought a $2000 CAD Motu 1248 to compare it to my modded 4x10HD. The modded 4x10HD just blew the stock Motu in about every aspect I could think of, except that the Motu had no hiss or noise. I knew the Motu is a bit of an entry level brand, but still a professional brand unlike the MiniDSP which is consumer type.  The 1248 was Motu’s flagship model at the time of introduction. It boasts a superb, but obsolete ESS DAC chip inside. And whilst its the lowest ESS grade available, it’s still an ESS. Those ESS chipsets aren't cheap compared to the competition.

​

The Motu1248 was supposed to sound good. It didn’t!

Internal inspection revealed a poor design with serious modifications required to harvest the best of the 1248. I really struggled not to modify the Motu on the spot, as I knew it had serious potential to perform great. And unlike the 4x10HD, it was dead quiet from the start. Despite significant potential for modifications, I decided to re-sell it and keep searching for a stronger start.

​

-It is hard to say if a modded Motu 1248 would beat a modded 4x10HD. I think that it has the potential to do so, but I never upgraded the Motu to discover its true capacity. It remains a mystery, and I'm not that curious.

I also needed more than 8 analog channels. The Motu 1248 only has 8 analog outputs. I was considering the Motu 24Ao (24 independent analog output) but the Motu 24Ao use the exact same design as the 1248 and due to lack of performance of the 1248, the 24Ao was quickly ruled out.

 

***

​

More Follow up:

Long story short, I bought a Lynx Aurora N 32 Dante and never looked back. Yes, it was 15 times more expensive than the stock 4x10HD, but it offers SO much more. It sounds tremendously better too. Much, much, much better than the modded 4x10HD was. A real eye opener here.

The Lynx is a top of the line brand used in the professional world, not the consumer world...

 

  • It has 32 analog output channels, not 8 like the 4x10HD but 32 independent, balanced channels. I’ll never run out of channel as I can add extensions if I ever want (64 channels analog is possible). It’s very “future proof”. 

  • Unlike the 4x10HD, the Lynx is fully DC coupled, with no capacitor at all in the signal path, unrestricted dynamics are there, I LOVE that.

  • It relies on an external computer to do the processing or is only used as a DAW (digital audio workstation) unlike the 4x10HD which does the processing internally.

  • The 4x10HD is limited in how much processing one can do as its internal DSP processor has a limit. (but it’s very easy to use)

  • The Lynx’s only limitation is how powerful your external computer is. The computer execute all DSP functions and then feed it to the Aurora.

  • The Lynx can then "process" as much as you want and this is huge bonus, as I always ran out of processing power on the 4x10HD. (However, the Lynx take lots of time to properly set up and is not user friendly at all)

  • It’s super reliable and has been field tested in studios around the world, customer service is top notch.

  • The biggest drawback setting up the system is that it’s an epic PITA. Not intuitive, and errors come easy, as one doesn’t know how to proceed. It’s a very steep learning curve, so buyer beware.

    • No graphical interface are available here, one must "program" every DSP function, one by one.​

 

 

Last word on sound reproduction:

"If you’re looking for financial advice, don’t ask what an analyst thinks of a potential investment. Instead ask him what he personally owns in his portfolio.

Most of the time, he doesn't even own what he's recommending you to buy. Skin in the game my friend."

​

That being said, never forget that U2, Metallica, Norah Jones, Michael Buble and most affirmed artists don’t record music on MiniDSP electronics. You kidding me?

They record in professional studios and use professional grade gear, such as treated sound booths, broadcast quality microphones and professional high-quality microphone preamps etc. These recording studios use professional interfaces (that’s what the Aurora is… a professional interface) for recording and mixing.

​

The Lynx Aurora N 32 stands among these top products used by studios around the world. If you listen to a Norah Jones CD, understand that the signal once passed through a professional grade interface.

Why audiophiles would use audiophile two channel DAC’s and "hope for the best" is beyond my understanding. Instead, use what the professionals are using. It’s their job, so they should know a thing or two about sound. Trust me...

​

 

Reproduce the music with the gear the professionals use to record it with. Not what Stereophile magazine tells you to use.

 

I’d recommend readers to look for the following brands of interfaces for their DSP needs: (no specific order)

  • Universal Audio

  • Lynx

  • RME

  • Burl

  • Forssell

  • Antelope

  • Merging technologies

  • Avid

  • Focusrite

  • Apogee

 

While these brands are not advertised in Stereophile, and never will be, I truly believe they’re the ones to use as a signal source for a tangible “state of the art”, multi-channel, fully active, DSP corrected systems.

 

 

I realize I’m not making friends in the audiophile community with these types of articles.

-Collecting friends is not my goal. Quality sound reproduction is my goal.

​

However, the reality is real. I hope audiophiles will one day wake up and realize realistic reproduced sound is more than possible in most home systems.

Sadly, mainstream information sources are sure keep them in the dark. It has for 40+ years.

The passive crossover loudspeaker industry will keep selling their 40+ years behind, obsolete technology for a long time still.

Change is hard. Much harder than staying put.

​

MiniDSP range of products does just that, by bringing a welcome change to audiophile perceptions.

I raise my hat to them, and KUDOS, as they allow the average adventurous user to see the tip of the iceberg, previously uncharted....and known by pros for many, many years.

They sure did it for me, as without the MiniDSP 2x4, I’d probably still be using a passive crossover. I did for my first 25 years in the hobby… sad.

Who care about the latest expensive passive crossover capacitor released, "this capacitor contains trace of pure gold and contains oil". LOL.

  • Best cap is no cap…bar none!

The only drawback is that audiophiles will one day awaken and realize that MiniDSP was just scratching the surface.

Then, he’ll upgrade his MiniDSP for something much more potent, like I did.

  • You don’t learn to drive on an 800 horsepower Nascar race car. You slowly move your way up, but the time spent driving the rookie car was invaluable.

  • MiniDSP is a formidable "learner car".

​

Thanks, MiniDSP.

​

Remember, I have a French mother tongue, doing my best to express myself in the Shakespeare language.

Thanks to a friend, Bevis, most of my website has been corrected for proper English grammar, much easier read now :)

 

Any suggestion/comments,  feel free to E-mail me.

bottom of page