3 Comments

Sound: *******1/2
Value: ********
(Read about our ratings)

Measurements can be found by clicking this link.

No matter where you might stand on the subjective vs. objective, measurements vs. “trusting your ears” debate, you can’t deny that the miniDSP IL-DSP DAC-headphone amp ($99 USD) can have a radical effect on the sound of your headphones. Like other roughly finger-sized DAC-amps on the market, the miniDSP plugs into a computer, phone, or tablet, and performs digital-to-analog conversion and powers a set of earphones or headphones. But while most of those products’ marketing claims are simply that they sound more awesome than other companies’ DAC-amps, the miniDSP incorporates a digital signal processing chip that can be user-programmed with different EQ curves.

Let’s talk about the EQ first, as that’s what most distinguishes this product. You program it using a PC or Mac app; after that, the EQ curves are all in the IL-DSP and they’ll still be effective if you plug the DAC-amp into a different device. Understand that this is no mere tone control -- it’s ten bands of parametric EQ, with the center frequency, Q (bandwidth), and boost or cut individually controllable for each band. Each of the bands can be a peak filter (centered at a certain frequency), or a shelving filter, affecting all frequencies either above or below the user-selected filter frequency.

miniDSP

This is a level of control comparable to what I’ve seen in the interfaces used to program the DSP chips built into almost all wireless speakers and many active headphones, so you’re getting pretty much the same amount of control that the pros get. So that’s cool -- but whether or not an audiophile without measurement gear can accomplish something useful with such powerful controls is the issue I wanted to explore.

For a $99 product, the IL-DSP sports an impressive DAC, capable of PCM resolution up to 24/192 and DSD64 or DSD128. The DSP function won’t work with DSD signals (it would have to convert them to PCM, so what’s the point?) or with PCM sampling rates higher than 96kHz. The amp isn’t so impressive -- it’s rated at just 30mW into a 32-ohm load -- but miniDSP’s website suggests, with admirable honesty, that it’s enough for earphones and most portable headphones. And if your headphones are rated at 100dB/1mW sensitivity, which is pretty typical, then 30mW will get you a sound-pressure level somewhere around 115dB, which should be more than enough for anybody.

In the box

The IL-DSP includes three cables: two Micro-USB-to-USB-C cables, one measuring 10" (15cm) and the other 28" (70cm), plus one 10" Micro-USB-to-USB Type-A. That’s fine for laptops and Android phones and tablets; iPhone and iPad users will have to find themselves another DAC-amp.

miniDSP

Use

The IL-DSP hardware is mostly easy to use. It has volume up/down buttons, and a middle button that turns the DSP on and off when held down, and also performs play/pause and track-skip functions when the IL-DSP is connected to an Android device. There’s also an integral microphone, which will let you take phone calls even if your earphones or headphones aren’t equipped with an internal microphone. I did encounter one problem with the hardware: when holding the multifunction button down to switch the DSP on and off with the IL-DSP connected to a computer, I found that the music sometimes paused but wouldn’t restart unless I disconnected and reconnected the device.

The IL-DSP software can be a little complicated to set up -- you need to download a ZIP file full of software, find the app that suits your device (PC or Mac), and install it, and if you’re on Windows, install a USB driver. It’s not the most intuitive installation process I’ve encountered, but I muddled through.

The control interface for the EQ is terrific. Just click a button to connect the software to the IL-DSP, then start configuring your filters. If you understand how a parametric EQ works, you’ll have no problem. If you don’t, read up before you go blindly messing around. With up to 16dB of boost and cut per band, this is an exceptionally powerful EQ, enough to get practically any headphones sounding pretty good or dreadfully bad. If you’re very technically proficient, you can even enter biquad DSP codes manually, which will allow more extreme curves than the provided interface offers.

miniDSP

However, unlike some of miniDSP’s other products, the IL-DSP doesn’t let you do EQ presets. Thus, you’ll have to readjust it every time you change to different headphones or earphones. It does let you adjust right and left channels differently, which you might want to do if your hearing is substantially different in one ear vs. the other (although your brain may have already adapted to that condition).

Just so we’re clear, the IL-DSP is not the amp you want if you have low-sensitivity headphones such as the HiFiMan HE6es or high-impedance models such as the Sennheiser HD 600s. If that’s what you need to drive, get an amp with at least about eight times the IL-DSP’s power (like the Periodic Audio Nickel) and use whatever EQ app you can find for your source device. In fact, the IL-DSP was just powerful enough to drive the AKG K371 headphones, which aren’t particularly hard to drive, so if your headphones or earphones are even remotely hard to drive, this ain’t the amp for you.

Sound

I don’t get too hung up on the sonic character of solid-state headphone amps; they sound similar enough that I find it a little silly to praise one while slagging a similar model. What I wanted to find out is what the IL-DSP’s powerful parametric EQ can do for you. And by you I mean someone who I assume has no measurement capability -- not even miniDSP’s $199 EARS test fixture. (If you do have an EARS, you’ve probably already got an IL-DSP, so go back to the top of the article and click on the link for the measurements, because everything I’ll say below will just confirm what you already know.)

Of course, you could tune the IL-DSP’s filters by listening, but unless you have the skill to accurately identify the frequency and bandwidth of sonic anomalies by ear, you could probably do a better job, and stand less risk of messing up the sound, with something like a six-band graphic EQ, which your source device already offers.

When I thought, “How would I use the IL-DSP if I didn’t have measurement gear?” I thought my best course of action would be to look at measurements that compare the headphones or earphones I’m trying to EQ with some known reference -- either the Harman curve or an established standard such as the Sony MDR-7506 or Sennheiser HD 650 headphones.

miniDSP

I started by seeing if I could, without the luxury of measurement gear, make the JVC HA-FW01 earphones sound like Harman curve earphones. I looked at the chart from my measurements of the HA-FW01s to see how they compare with the AKG N5005s (the earphones that currently best reflect the Harman curve), and made some notes about where to boost and cut, and by how much -- a little boost in the low bass, some cut in the upper bass, and a little more oomph in the upper mids and highs. Although it’s hard to know exactly what center frequency will give you the results you want, or what Q will give you the right bandwidth for each filter, the display in the IL-DSP’s EQ app shows you what the curve looks like so you can at least make educated guesses. Still, doing it like this is a challenge because the EQ bands may overlap in ways that are hard to predict.

I expected to hear slightly tighter and better-defined bass, as well as extra presence in the upper mids, with a little more vocal clarity at the possible expense of a less-rich sound. And that’s kind of what I heard; the sound was definitely treblier, and the bass tighter, but the balance was a little thin, and too edgy on voices. To compare the result with Harman curve headphones, I pulled out the AKG K371s, which are said to be within 1dB of the Harman curve. The K371s sounded more balanced, so I used them as a reference for tweaking the IL-DSP, and messed around with the EQ settings until I got the HA-FW01s sounding pretty close to the K371s. The biggest advantage was that the JVCs now had that neutral, clean, non-bloated bass that’s characteristic of the best Harman curve headphones and earphones. While the tonality was about as neutral and balanced as that of the K371s, the treble wasn’t quite as smooth; with more tweaking, I might have been able to improve the sound further, but I was pretty happy.

miniDSP

So unless you have measurement gear, the best approach with the IL-DSP is to (1) find measurements of your headphones or earphones and the headphones or earphones you’d like to EQ them to emulate, (2) set the EQ in a way that you think will get them pretty close, and (3) fine-tune the filters by ear until you get the sound you want.

Of course, that means you’ll need measurements of both of those headphones, preferably done on the same equipment. We have many such measurements here on SoundStage! Solo, and you may find them on a few other sites, too. There’s also the option of e-mailing your friendly neighborhood headphone measurement guy (i.e., me); I have measurements of hundreds of headphones and earphones and might be able to send you the chart you need. (You’ll have to figure out the filter settings yourself, though.)

Comparison

I’m not aware of a direct competitor to the IL-DSP. If you want to do something similar but for whatever reason -- such as wanting a stronger amp or wanting to use an iOS device -- you don’t dig the IL-DSP, you could use a third-party app for EQ. That could be a freebie included on your source device or downloaded through Google Play or the Apple App Store, or it could be something fancier such as Sonarworks True-Fi. You could pair that with a beefier DAC-amplifier, such as the AudioQuest Cobalt, and have a nice rig that could power a big set of audiophile-grade cans.

Conclusion

The EQ in the IL-DSP is a powerful creation with more than enough muscle to get a mediocre set of headphones sounding a lot better and a great set of headphones sounding a little better. It’s also easy to use, although if you don’t have measurement gear, you can expect to spend a lot of time figuring out what curve is right for your earphones or headphones.

miniDSP

However, the amp in the IL-DSP is not a powerful creation. It’s ample for typical earphones and headphones, but ill-suited for use with the more demanding models that audio enthusiasts typically use.

All things considered, if you have the right headphones or earphones and plenty of patience, the IL-DSP is a very affordable and effective way to get better sound on the go.

. . . Brent Butterworth
This email address is being protected from spambots. You need JavaScript enabled to view it.

Associated Equipment

  • Headphones -- JVC HA-FW01, AKG K371, Campfire Solaris
  • Laptop -- Lenovo Windows PC

miniDSP IL-DSP DAC-Headphone Amplifier
Price: $99 USD.
Warranty: One year parts and labor.

miniDSP
Phone: (+852) 2358 2066

Website: www.minidsp.com

Say something here...
Cancel
Log in with ( Sign Up ? )
or post as a guest
People in conversation:
Loading comment... The comment will be refreshed after 00:00.
  • This commment is unpublished.
    gzost · 7 months ago
    Thanks for the review - this is an interesting product. Since EQ is the easiest way to improve the sound of (most) heaphones, the easier it becomes to EQ the better. Something you can take with you which EQs a favourite pair of headphones everywhere is immensely convenient.

    Since the headphone amp has such low output, how about combining this with an amp? What is the output voltage it can push into a line input?
    • This commment is unpublished.
      Brent Butterworth · 7 months ago
      Great point, that hadn't even dawned on me. My Neutrik ML-1 is showing 1.06Vrms at 0.014% THD into the 20KΩ unbalanced input on the ML-1 with the volume on the IL-DSP maxed, so it should work fine driving a larger amp with an analog input.
      • This commment is unpublished.
        gzost · 7 months ago
        Thanks for measuring! Yes, depending on the application that should work OK. It's a bit low (should be 2V for unbalanced consumer gear AFAIK), but then -6 dB level would not be a problem with my desktop amp and preferred headphone.

        BTW, https://www.qudelix.com/ are about to release something quite similar, but with much higher output.
  • This commment is unpublished.
    oratory1990 · 7 months ago
    For those that have no measurement gear but want to use parametric EQ with their headphones, I'm compiling a list of EQ presets for headphones (based on measurements):

    Frequently Asked Questions:
    https://www.reddit.com/r/oratory1990/comments/arz0tn/psa_oratory1990s_list_of_eq_presets/


    Preset list:
    https://www.reddit.com/r/oratory1990/wiki/index/list_of_presets


    Start from the preset and adjusting the values to your liking.
    This won't make every headphone sound the same, but it will make every headphone sound better.
    • This commment is unpublished.
      Brent Butterworth · 7 months ago
      Great stuff, and I recommend people read your FAQ -- it's excellent.
    • This commment is unpublished.
      Mauro · 7 months ago
      Hi oratory1990!

      thanks for the work you do..
      two questions if you don’t mind
      - in your experience, given the same headphone model, differences from headphone to headphone would justify using high Q filters? My experience is that even the two drivers on the same headphone might have quite significant differences above 5kHz
      - do you think that the GRAS rig is representative of the perceived frequency response? Moving across the frequency spectrum with a tone generator I hear dips not present in your measurements (for ex. Philips X2 around 6khz)
      • This commment is unpublished.
        oratory1990 · 7 months ago
        Hi Mauro!

        1. unit variation is a problem, yes. Which is exactly why one should avoid using high-q filters with positive gain unless one has had the chance to have his/her SPECIFIC headphones measured. But unit variance is not as problematic for every headphone. For example the 6k resonance on an HD800 is pretty much the same on every unit I've ever measured.
        But on a DT770 or DT990, the peaks will be slightly different for every unit. I always try to get multiple units to do an average of them for that reason.
        It still starts to get really problematic above 10 kHz for other reasons, which is why I rarely use peak filters at all in that region, as the peaks will not only differ between units but also change a lot between individual ears.

        2. according to my experience with most available measurement rigs on the market, yes. A setup with a 60318-4 coupler and a set of anthropometric pinnae (very important) is the best thing we currently have. There are problems mostly with smaller earcups where there is potential leakage between the head and the pinna (because on most measurement setups the "fleshy" silicone pinna component does not cover the whole cheek), which is why one needs to take extra care when measuring those. Some things are more easily measured on a full manikin like the Gras 45BC Kemar, others are more easily measured on a 45CA or a 43AG.

        Listening to sweep stimulus with your own ear can be really hard, our ears and brains have lots of psychoacoustic tricks they play on us. I wouldn't trust myself identifying peaks by listening to a sweep. It's much more easy for me to listen to music that I know well. The Philips X2HR has a peak at 5k, do you mean this one? I could clearly hear that one when I had it on my head. peaks in the 5-6k range are very audible with cymbals, rock guitars and the like.
        • This commment is unpublished.
          Mauro · 7 months ago
          I move manually through the frequency response incrementing of 10/100hz every time I tap a button. I use an iOS app called Audio Function Generator.
          With the Philips X2HR I can hear two distinctive dips at 5.8kHz and 8.6kHz. Peaks are in between and on the sides.
          in the plots acquired with your GRAS setup the frequency response seem to be pretty smooth.. that’s why I asked you about the correlation with perceived sound levels
          • This commment is unpublished.
            oratory1990 · 7 months ago
            Hi Mauro!

            The plots shown are averages over multiple reseats, with slight positional variation in between them. I do this because I want to measure the headphone, not the pinna. Some of the peaks and dips in the treble are caused by the pinna shape and as such will depend a lot on how exactly the headphone is positioned on the ear. You will find that if you move the headphones by just a few millimeters, the position and/or height of the peaks you described will change!
            And this will be slightly different for everybody, which is why I do the average over multiple reseats, and base the EQ on that.
            • This commment is unpublished.
              Mauro · 7 months ago
              I’ll have to experiment more moving headphones around my ears. I knew about the theory but I am not so sure about the consequences on perceived sound. The treble in that specific headphone seems to be quite sizzling. I guess the answer is not the frequency response being so smooth. Maybe some other measurements like spectral decay can be more explicative? I would be curious if someone could push the boundaries of measurements to perceived sound correlation beyond FR.
              • This commment is unpublished.
                oratory1990 · 7 months ago
                it's also entirely possible that the unit variation of your model is high enough that they would require a very different EQ to combat the specific peaks you're hearing.
                And lastly it's also very much possible that your specific pinna shape creates different peaks, if your pinna is shaped very differently than the measurement pinna.
                • This commment is unpublished.
                  Mauro · 7 months ago
                  Then I would like to ask Brent or Sean Olive which are the headphones better suited to people with very small ears like me!
                  thanks oratory1990 for the kind explanations!
                  • This commment is unpublished.
                    Brent Butterworth · 7 months ago
                    Sean might know, but I have no idea. I don't have subjective test data that identifies any particular differences in preference between people with average or large ears and people with small ears. I also don't have measurements done with a smaller pinna. G.R.A.S. does make smaller versions of the simulated pinnae I use, but those pinnae were about $800 each last time I bought one. Not an unthinkable investment for me down the line, but there are other pieces of measurement gear for which my need is more urgent.
                    • This commment is unpublished.
                      Mauro · 7 months ago
                      I read once about different pinna sizes here https://clarityfidelity.blogspot.com/2015/05/introduction-measurement-procedures.html, but it seems to me a different setup than yours..

                      Audeze has an app that goes with their newer cipher cable/dongle that processes the ear shape using a picture, just like SoundBlaster was trying to do.. maybe they also might know something more about different ear shapes.

                      Anyway I love NAD HP50 and from what I have read your ears are very large and mine very small..this is not a good sample for a statistician but it is a starting point for me!
                      • This commment is unpublished.
                        Brent Butterworth · 7 months ago
                        I'm not sure where the biggest of those ears comes from, but the two flesh-colored ones look like G.R.A.S. ears. The larger one is the KB0065, I think -- the one I used before they came out with the new anthropomorphic ears. The new ears were designed to help eliminate the inconsistencies he notes with the older large ears. The unnatural stiffness of those older ears made seating the headphones more difficult. Makes me wonder if I could get one of the old version of the smaller ears at a "blow-out" price . I'm guessing not...
  • This commment is unpublished.
    Mauro · 7 months ago
    Very nice product and article! Much appreciated!

    If MiniDSP comes out with a lightning version, I would buy it right away!

SoundStage! Expert: Sonus faber Olympica Nova Speakers - 1) General Care (February 2020)

SoundStage! Expert: Sonus faber Olympica Nova Speakers - 2) Grille Care (February 2020)

SoundStage! Expert: Sonus faber Olympica Nova Speakers - 3) Cleaning (February 2020)

Latest Comments

Brent Butterworth 6 days ago Beyerdynamic T5 (3rd Generation) Headphones
@Joel LeeGood question. According to Beyerdynamic:
1) Both plates on the earcup look entirely different
2) The new ...
1 image(s)
How do you know you have the 3rd Edition if the cups only say T5?

Joel ...
Brent Butterworth 20 days ago Where Are We At With The Harman Curve?
@SianaIt's absolutely possible to do that. It's called measuring in phons, which is a unit ...
But when I’m in a car and it’s pretty quiet I don’t experience eardrum suck. ...
This is a super-geeky article! Loved it!