I thought DACs barely mattered. Then the D17 happened.
Thanks to Hifi-Passion for loaning the iBasso D17 Atheris for a few days. Unfortunately my time with it was limited, so this isn’t a full review but rather a collection of listening impressions. German/EU readers can get the device with full warranty locally from here: https://hifi-passion.de/iBasso-D17-Atheris/HL02728 (not an affiliate link).
Quick Navigation
I didn’t think DACs could sound this different
I have to admit, I have never been poorly equipped with DACs. Fortunately, I came into contact with Chord very early in my review career and later settled on RME products for both my desktop and living room. Ever since, my search for DACs has effectively been over.
Especially at audio shows I have to concentrate very hard before I notice meaningful differences between source devices. My ears are trained to hear subtle tonality changes between IEMs. Macrodynamics, depth and presentation are qualities I never felt particularly confident judging.
That became very apparent when I connected the D17 to my iPhone. I pressed play. And immediately assumed something was wrong. The presentation was so much more forward, immediate and contrasty than anything I was used to that my first instinct wasn’t “this sounds different” but “this can’t be working correctly.”
I even suspected a compatibility issue after my DX260 refused to connect over USB. Luckily, Ichos on Head-Fi reminded me that the DX260 has a coaxial output that I had never actually used before. Once that worked perfectly, I had to accept that nothing was broken. The D17 simply sounded unlike any DAC I had previously owned.
Why the D17 intrigued me in the first place
The D17 combines several technologies that I had intentionally avoided until now: an R-2R ladder DAC and a tube output stage. Most of my experience has been with delta-sigma DACs and solid-state amplifiers. Having both in one single portable device just made me very curious. iBasso doesn’t stop there, though. Users can also switch to an FPGA-based 1-bit conversion path, effectively turning the D17 into a dedicated DSD DAC. The line-out also bypasses the tubes so that makes the D/A conversion particularly interesting.
Modern delta-sigma DACs generally outperform R2R converters in objective measurements. Yet R2R remains fascinating because its design priorities are different. The closest analogy I can think of is comparing a handmade mechanical watch with a quartz movement. One is objectively more precise, while the other appeals because of the engineering itself.
The Test Setup
I have tried the D17 Atheris as DAC/amp combo with my HEDDphone D1, my all-time favorite Softears RS10 and the smooth Grand Oriveti Supremacy. I used a ddHifi BC44XLR to isolate the DAC performance and feed my Lake People G111 MKII and compare with my Chord Hugo 2 directly.
Initially, I struggled to make sense of what I was hearing. Yet over the following days something interesting happened: the presentation stopped sounding strange and started sounding intentional. The more I listened, the more I found myself adapting to it – and eventually I could imagine using the D17 regularly.
Living with the D17
After a few hours I slowly stopped wondering whether something was wrong and started asking a different question: What exactly was I hearing?
The DAC Comes with 3 Modes
Before we dive into the various DAC operating modes, I noticed that all of them had a few things in common. First of all, I perceived the sound as very immediate, direct and lively. I noticed no smoothing-over or anything that would arbitrarily soften the sound. That makes the presentation very detailed.
Over-Sampling
Regarding the first mode, OS, I would consider it the warmest and bassiest option. The low-end is very controlled, though, but also the midrange feels thicker and makes voices especially intimate. The treble can come off as a bit contrasty and splashy at times, further increasing the sense of directness. In this mode, the digital signal is processed as PCM.
Sigma-Delta
The second option, SD, bypasses the R2R ladder and instead converts PCM signals into DSD to function as an FPGA-based 1-bit DAC. I immediately noticed the drop in volume as DSD demands additional safety headroom. Assuming I managed to match the volume during my comparison, I would say this mode sounds softer and closer to the Chord Hugo 2, but still distinct and with more bite. To me, the midrange seems to have more room and draws more focus than in OS mode.
Interestingly, I later learned that the D17 also reduces its available output power considerably in SD mode. I didn’t notice this during my own comparisons, but it’s worth keeping in mind if you’re planning to drive more demanding headphones. Marcus’ excellent Headfonics review discusses this in more detail.
Non-Oversampling
Finally, there is the hybrid non-oversampling mode (NOS), in which PCM will use the R2R ladder and DSD uses 1-bit. Similar to OS mode, the sound is direct and raw. Without over-sampling, my expectation was that the tone will be even warmer. But to my ears anyway, NOS sounded tonally leaner with more midrange presence than OS. I would even say that this mode is the most detailed. My impression was similar to wearing glasses that might be a bit too sharp – you can’t pinpoint any flaw but it stresses your brain, at least for me. It felt slightly stressful when a lot is going on, but not in a way that anything is overemphasized but just that I was handed a lot of information at once.
While I initially preferred SD mode, it is because of the NOS mode that I wish I could spend more time with the D17. This kind of presentation is exactly what makes me want to go through my whole library again.
The NuTubes
Up until this point I had exclusively used the line outputs into my own amplifier. That means everything above reflects the D17 purely as a DAC. The headphone outputs always use the NuTube stage, so the following impressions include both the DAC and amplifier working together.
Pairings
Note: Users of ultra-sensitive IEMs should keep the output impedance in mind. At around 1 Ω single-ended and nearly 2 Ω balanced, it is higher than what I would normally consider ideal for highly sensitive multi-driver IEMs. Fortunately, I did not notice any hiss.
That said, the headphone output performed close to my expectation. The presentation is full and warm. Using the very clean and ultra-detailed Softears RS10, the D17 gave it warmth and richness with ample bass – depending on whether you consider the RS10 a precision tool or a method of enjoying uncolored music, can either be perceived as an enhancement or a less-than-optimal pairing. Either way, I found NOS to be the most eye-opening.
No surprise, the D17 matched well with the Sennheiser HD 650. After all, this headphone is known to pair well with tubes and the iBasso enhanced all its strengths, especially warmth, intimacy and micro-details. At times, I felt like it was exaggerating, but ultimately I really enjoyed the harmony, especially in SD mode. Revisiting my listening notes, I found the remark: “What more could I want?”
The HEDDphone D1 is the most transparent headphone to my ears yet. It thus showed strong reaction to the different modes and also to the tube amplification. I’d ascribe a slight overall gritty signature to the pairing, but not so much that it would be distracting – rather film grain that can make the image look more detailed. In this case, I actually enjoyed the OS mode with tubes a lot, which does color the D1 quite a bit, but arguably in a good way, giving it weight and more saturation.
TL;DR
Compared to my iBasso DX260 DAP, the D17 remained consistently more intimate, immediate and contrasty. The DX260 sounded smoother and more distant by comparison. Against the Chord Hugo 2, the gap became smaller, particularly in SD mode, yet both devices still approached music very differently. The Hugo 2 presented details effortlessly and almost silkily, while the D17 constantly pulled my attention toward the music. Neither approach struck me as obviously superior, but the contrast between them was fascinating.
I still don’t know whether I would buy the D17. I also don’t know whether I would ultimately prefer it over my own DACs after months of listening. What I do know is that it challenged assumptions I had held for almost ten years. I’ve often argued that DACs and amplifiers generally make subtle differences compared to headphones. The D17 didn’t necessarily prove me wrong – but it certainly reminded me that there are still designs capable of surprising me.
In a hobby where genuinely new experiences become increasingly rare, that alone made the D17 memorable.