Astell&Kern A&Ultima SP2000 - Flagship Digital Audio Player - Review

Review written by Ian Dunmore (@Torq)

Introduction

Astell&Kern’s SP2000 is the third release in their flagship A&Ultima tier of products - behind the SP1000 and the smaller, slightly revised, SP1000M. The new model features a number of hardware upgrades over the prior top-of-the-line model, including the latest flagship DAC modules from AKM, an all-new and more powerful audio/amplifier circuit and double the storage of its predecessors.

The unit is initially offered in copper and stainless steel chassis/bodies. If A&K observe their previous flagship release cycle, a lighter (and perhaps less expensive), aluminum bodied model may follow later.

The A&K SP2000 being reviewed here, and pictured above, is my personal unit and was purchased from headphones.com. This review will be longer than usual due to the broader mix of functionality that digital audio players exhibit vs. just a DAC or an amplifier; to get to the meat of what’s interesting to most sooner, I’ve put details on use and non-sound related features after the “Sound” section - and there are further details on sound in the “Comparisons” section.

The gear chain/equipment used in this review can be found here, and the music I use in my reviews and evaluations is listed here.

Specifications

The full set of specifications for the SP2000 can be found here, but the highlights are as follows (upgrades from the SP1000 are italicized):

• DAC: AKM 4499EQ x2
• PCM: 32 bit/768 kHz
• DSD: DSD512 (1 bit/22.4 MHz)
• MQA: Full Decoding
• Output: 3.0 Vrms (SE, 3.5mm TRS), 6.0 Vrms (BAL, 2.5mm TRRS)
• SNR: 124 dB (SE), 125 dB (BAL) @ 1 kHz
• WiFi: Dual-Band (2.4 & 5 GHz)
• Internal Storage: 512 GB
• Dimensions: 2.98” W x 5.2” H x 0.64” D @ ~387g

The output/amplifier stage of the SP2000 is all-new, and features separate, independent, channels for the balanced and unbalanced outputs, with a 50% increase in power over the previous models (output is equivalent to the SP1000 AMP module on it’s low-gain setting, and as such the external contacts for such a module are omitted here).

Features & Functionality

Astell&Kern, being the progenitor of the high-resolution digital audio player, tend to deliver extremely feature-rich models. The SP2000 is no exception. The only feature I can think of that is omitted here is the ability to operate as a Bluetooth DAC/receiver.

There’s on-board dual-band WiFi, which permits direct, on-device, streaming for the most popular streaming services, which is combined with an Android-based “open application service” that allows installing each service’ native streaming clients. Over-the-air firmware updates are also supported.

Dual output sockets, for single-ended (3.5mm TRS) and balanced (2.5mm TRRS) headphone drive, with optional cables available to provide balanced line-out for desktop (pre)amplifiers that uses both sockets and terminates in industry-standard dual 3-pin XLRs.

Finally, the unit supports fast-charging - with a 9V @ 1.67A charger (15W) reviving the unit from fully-exhausted to 100% in about 2.5 hours.

Build

This is a substantial player, coming in at 15.25 ounces for this copper unit - it has a very satisfying heft in the hand, though the size and weight may bother some. Buttons are tactile and responsive. Connections are solid with no apparent play or movement. Fit and finish are excellent, with the grain of the metal and the gorgeous copper finish adding a definite sense of opulent luxury to the proceedings.

A little play in the volume dial/multi-function selector results in the unit not quite exuding the same “hewn from solid” feel of the Sony NW-WM1Z, but this is a very minor nit and is true of other similar units that also have dials or rockers on them (e.g. the Cayin N8).

I’ve had my unit for a little over a month now, under extensive use, and I’ve not noticed any discoloration, tarnish nor patina on the copper finish - something that was a frequent occurrence with the AK380.

Package

The SP2000 package is relatively simple, if elegantly executed. Within the outer sleeve you’ll find a compact wooden box and within that a protectively wrapped SP2000 player. Beneath the player you’ll find a spare microSD card slot cover and a sturdy USB-A to USB-C cable with Astell&Kern branding.

In the top section of the box there is an Astell&Kern leather case, which has the usual verysnug fit, and then the multi-piece screen/chassis protector films - which includes coverage for all surfaces on the SP2000, and finally the warranty and quick-start guides.

 

Sound

It took very little time to determine that the SP2000 offers flagship-class digital-audio performance, comfortably in the same realm as players like Cayin’s N8 and, for me, ahead of other excellent units such as the Sony NM-W1MZ, as well as Astell&Kern’s own SP1000 line. In fact the SP2000, and players in its class, are making healthy inroads into higher-end desktop-class performance – albeit typically at higher prices than the same level of performance would require with desktop gear.

The general signature here is a straightforward “reference” delivery but without stepping into “clinical” or “analytical” territory - just very natural and realistic. It is extremely clean, highly resolving, fast and has a neutral tonal balance. The unit is not editorializing what it is fed. If you’re looking for a unit to color the sound you hear in any significant way, this probably isn’t it.

Performance is notably better with harder to drive headphones than A&Ks previous models, with the SP2000 having output equivalent to the external SP1000 AMP module on its low-gain setting. While not unusually demanding to drive, the ZMF Vérité definitely benefit from the increased power and authority of the SP2000’s new audio output implementation. Bass hits harder and faster, while remaining highly composed and well controlled, and large, staccato dynamic shifts do not exhibit any sense of compression at all - even from the single-ended output.

Low-level details are teased out of one’s material in a way that makes them more apparent, but still natural. This isn’t an “over-sharpening” or “fake detail” effect; more simply that either through a more agile amplifier, or a subconsciously lower noise floor, you can quite readily discern tiny inflective, dynamic and tonal changes. This was first apparent listening to some complex percussive elements with the realization that I could tell that drum strikes are impacting in subtly different spots on the skin (very evident with Timpani, but still quite audible with Snares). Similarly, when played with fingers and heel, instead of sticks, it was no work to hear changes resulting in different hand positions. And this extended into other instruments as well, with subtle elements such as changes in the musician’s hold or orientation levying audible changes in tone and timbre that were much harder to hear, if you could hear them at all, with other units. This level of rendering is something more commonly heard only in serious desktop components.

Transient performance is exceptional, with pizzicato and rapid staccato elements being impressively fast and clean, but without artificially muting or terminating the natural decay of more sonorous instruments. A rapid draw across the width of a harp lets you easily make out the initial bite and tone of each note and string, while still letting those notes blend and ebb away beautifully.

With high performance IEMs, and their tendency to be ultra-high resolution, especially at higher frequencies, the detail rendering and transient performance is even more obvious. Getting the best out of the SP2000 demands highly capable transducers. Many of these, of late, have been very low in impedance and extremely sensitive, and this has posed issues with noise and hiss for lots of high-end IEMs. The A&K players are typically silent in these cases, and this new flagship is no exception - backgrounds were absolute void, even playing digital silence with the volume raised (to avoid the trick some units employ of turning off the amp output entirely if music isn’t playing). Be it the notoriously hiss-prone Campfire Audio Andromeda or Shure SE846, or even more sensitive models, there was no noise or hiss at all.

Tonal elements are absolutely balanced with no reticence nor emphasis at any point in the spectral range. Bass elements are powerful and present, with no issues digging into near infra-sonics and delivering sub-bass growl and rumble in an almost effortless fashion while delivering tuneful bass lines (“The Jerk” - Drums of Death, “Red Wave) at the same time as slamming you with even lower tones. Another highly revealing track for this sort of test is Ken Ishii’s “Prodrome” (Garden on the Palm) - assuming you can find a good (lossless) copy.

Midrange performance is just as solid, with perfect balanced relative to the rest of the spectrum, a lucid, almost holographic, delivery, that is both light on its feet but full in body. Vocals are natural, nuanced, detailed and inflected, and so precisely resolved that even shifts in the performer relative to their microphone can be heard quite easily.

Treble issues are non-existent, with no sense of artificial tension or glare, no added sibilance (only what you feed it) just a buttery smooth delivery across the board. Cymbals and brass have superb bite and convincing aggression, but no exaggeration or splashiness at all. Extension is excellent as is the sense of space and air the SP2000 is capable of relaying.

Stage is stable with a natural sense of scale and ambiance. The binaural version of Amber Rubarth’s “Sessions from the 17th Ward” (Chesky Records), an engaging piece in any context, and a superb example of solid recording/mastering, provides an highly vivid means to explore staging and imaging abilities. Played here, either with IEMs or the LB-Acoustics MySphere, the effect is particularly pronounced.

This is an excellent, and completely coherent, delivery with no notable weak spots nor issues. Just pure, detailed, emotive music. The SP2000 won’t add “excitement” where there is none - that’s going to be down to the music and your transducers, but it won’t omit anything either - if it’s in the recording you’ll hear it just as it is.

Single Ended vs. Balanced

Using the balanced output, both with IEMs (where the additional power is not really relevant) and particularly with full-sized, higher-impedance, headphones, yields a noticeable step up in overall performance.

For a start, both low level detail and micro-dynamics are improved, with tiny inflections in a performers voice, resulting from subtle, rapid, changes in volume level being more readily audible – which helps convey their emotion (“Show me the Place”, Leonard Cohen – “Old Ideas”) more vividly.

Stage expands slightly and both imaging and spatiality become notably more solid and better able to delineate depth and position (with headphones that are actually capable of doing so). The MySphere headphones are particularly good for this, and binaural recordings (try “Dr. Chesky’s Sensational, Fantastic, and Simply Amazing Binaural Sound Show”), with good IEMs, also provide an excellent demonstration of the difference here.

The most notable difference here, however, is with drive authority and more challenging cans, where level, raw impact/slam, immediacy and bass coherence are all improved and the overall sense is one of “ease” and better control.

EQ

Engaging the EQ function on the SP2000 will immediately reduce the output level at all frequencies, so as to avoid digital-clipping when boosting any given frequency. I understand why A&K have taken this approach – it’s a safe way for those not fully-versed in the operation of digital EQ from running into digital clipping. However, this level reduction is fixed and significant and with some headphones, playing quieter recordings, it may result in being unable to get sufficient volume.

I’d really like to see an option for manual override of the pre-amp/gain level reduction for more advanced users.

Actual EQ settings can be saved as presets. And you can configure them either using a graphic-equalizer style interface or a parametric one (and you can flip between them with the unit converting your settings as you do so).

Just enabling EQ, but without making any adjustments will, as mentioned above, reduce the output level from the player by quite a bit (5 dB, or almost a 50% reduction in apparent volume), but overall sound quality is not, otherwise, negatively affected (once you adjust the volume to match). Adjustments work as expected, with no apparent unintended artifacts from the player-side of things. This is a nice change from many DAPs I’ve encountered, where simply enabling EQ at all has deleterious effects.

Copper vs. Stainless Steel

For as long as Astell&Kern have been offering their players in different chassis materials, I’ve seen various claims that they sound different – with some people going as far as calling the differences “night and day”.

To date, in level-matched blind comparisons, I’ve never personally been able to discern any audible difference between A&K’s players based on different chassis materials. Not with the AK380. Not with the SP1000. And still not with the SP2000 – though my opportunity to compare two SP2000 units was very brief and not in ideal conditions.

On the scope, I did find a small delta … but we are talking about a ~1 dB rise in the treble from about 26 kHz to 30 kHz – which isn’t going to be directly audible (unless you’re a different genus of mammal than would be a typical DAP user). I’d like to re-verify that as those measurements were taken in a hurry, using portable tools, and its possible settings were different on the players.

From my perspective, I can’t offer any encouragement to do anything but either audition them side-by-side for yourself (ideally level-matched and blind) and decide for yourself OR just buy the version you like the look of the best and/or that weighs the least.

USB DAC

Like most DAPs these day, the SP2000 can act as a USB DAC/amp. This is enabled via the “USB Mode” toggle from the drop-down quick-panel available from the notification bar. Note that you’ll actually need to have the unit connected to a USB source device before you can change this from its default of “Disk” to “DAC”.

I noticed no audible shift in sound at all when using the SP2000 in this mode, regardless of which USB source I used, vs. having it play directly off its internal storage. This is not always the case with digital audio players, though really should be when talking about units at this level.

Oddly, the unit gives no indication of the bit-rate, sample depth or format of what it is being fed when in the mode, nor is there any indication for MQA content. In fact the display only has a mode acknowledgment, and a button you can press to switch out of USB DAC mode when used this way.

Under macOS

The unit shows up under “System Preferences | Sound” as expect and will present 16, 24 and 32-bit integer bit-depths at sample rates of 44.1, 48, 88.2, 96, 176.4, 192, 352.8, 384, 705.6 and 768 kHz within the “Audio Midi Utility”. It switches rates correctly and cleanly (and without clicks, pops or other noises) under control of your player software (Audirvana and Roon in my testing).

Under iOS

While I can get the SP2000 to show up as a selectable output using both types of Lighting to USB adapter, and no “Accessory uses too much power” message is thrown, the unit disconnects again after a few seconds. When it is connected, if it stays that way for long enough to start music playing, no sound emanates from the unit. Thus if you want to use an iPhone to feed the SP2000, you’ll need to do it via WiFi hotspot and the on-device streaming capability of the player itself.

In-Use/Operation

Usability, and performance, of DAPS varies greatly. They’re a relatively niche market (within the already niche market of high-end portable audio), especially with the more expensive models, and this can significantly limit the kinds of processors available to a manufacturer. What might be an $8 state-of-the-art 8-core CPU in smartphone quantities (millions of units), might not be cost effectively available at all for the kinds of volume most high-end DAPs would require.

Happily Astell&Kern don’t seem to have any issue securing good SoC/CPU parts for their players and aren’t limited to the group-buy bargain chips that have been commonly featured in some other players.

User Experience

Despite having had a number of DAPs with a similar footprint to the SP2000, even recently, none of them have had the same active screen real-estate - 5 inches in this case. This makes for a highly readable display and allows for rich, visual, navigation. There’s enough space here that even browsing via the album-cover grid view is a pleasure. The screen itself has excellent contrast, vivid color, and enough brightness range that I can comfortably using it sitting on my deck on a brightly lit day.

The fluidity and responsiveness of the user interface on the SP2000 is excellent - the best I’ve come across so far, in fact. It’s not quite as smooth and snappy as current-flagship smartphone interfaces, but its quicker and smoother than anything else I’ve used.

Most common options are available from a pull-down menu via the notification bar at the top of the display. And when scrolling a long list, you can slide your finger along the right-hand edge of the display for faster navigation.

Search has some oddities, which I run into a fair bit with pushing 40,000 tracks on the device. The most notable is that when searching for an album, any variance in the fields can result in some tracks not being returned. And if you hit “Play” from that screen, it’ll only play the tracks that are in the album AND matched the search, not simply the tracks that are in the album.

Something that Astell&Kern have nailed, and which other DAP manufacturers still have not, in my experience, is handling of album art. All of my albums and tracks show the proper art on the SP2000. With the Cayin N8, I still have 50 or so albums I cannot get to display the art for (despite them working in Roon, Audirvana, all of A&K’s current players, and after changing the JPG encoding to “Baseline” with the Sony NM-WM1Z).

Internal Memory & Storage

Having 512GB of storage in the player itself, is a welcome change from the relatively cramped capacity present on most other high-end players - and one of a few reasons I decided to buy the SP2000. The microSDXC slot supports high-capacity cards and I’ve been using a 1TB SanDisk Extreme card with no issues.

NOTE: If using card storage, make sure you format (using the “Erase SD Card” function in under “System”) in the SP2000 itself before you use it for the first time, otherwise you will get a process warning message on every startup as well as some odd/slow behavior around library indexing.

Adding Music

Astell&Kern’s modern DAPs uses the ungodly-slow “Media Transport Protocol” (MTP) to expose their internal storage as well as providing access to the microSD card, if installed. This is natively supported under Windows, but requires extra software if you’re running macOS.

The Astell&Kern recommended, free, macOS application for this is “Android File Transfer”. This works, but is extremely slow (more than 12 hours for a full 1TB card), very primitive and has a nasty tendency to fail on even moderately large transfers - leaving you to figure out where it gave up. A better option, and the one I use myself, is “Commander One Pro”, though you will need the paid version to do MTP file transfers.

While you can access a microSD card that’s installed in the player, it’s much faster to copy to it directly using a high-speed card reader and with the card mounted as disk on your OS.

Library Indexing

With up to 1.5TB of music on the device, library indexing can take a while - a full 1TB microSDXC card of mixed lossless and high-resolution files, takes about 10 minutes. This occurs in the background; pull down the notification area and it will show “Scanning media …” while this is occurring. An actual progress indicator would be nice here. Powering up the unit results in a brief “refresh” (60 seconds or so) scan each time, though the unit isusable while this process runs.

One quirk to note here, is if doing a full library index, with a card not formatted in and by the SP2000, not only does it run unusually slowly, and result in much longer start-up refresh-scan times, but it sometimes regulated in audible artifacts. These went away when indexing was complete and didn’t return. But you’re just much better off avoiding the issue by formatting your card in the player before you use it.

As a Transport

The SP2000 supports both USB and TOSLINK digital outputs, allowing use of the unit as a pure transport. The only time I tend to use DAPs in this fashion is at shows and meets, but I know others like DAPs as their one-stop, consistent, source, even in their home rigs.

I did my primary testing here into a Chord Hugo 2. TOSLINK output, via a 3.5mm TOSLINK to normal TOSLINK cable from Lifatec was without issue and worked without dropouts or other issues all the way up to 24/192 kHz.

USB output is extremely clean, and did not result in any audible, nor measurable, artifacts at the analog output of any of the DACs I tested this with, even vs. my custom-built, fully isolated, ultra-low-noise USB source.

WiFi Performance

This is markedly improved over the SP1000, and every other WiFi equipped DAP I’ve had in my hands to date. It maintains a more tenacious grip on the signal, even at the extremes of range. With the SP1000M I could find spots in my little waterfront place that would be somewhat tenuous, especially with high-resolution streams from Qobuz. With the SP2000 I could go down four floors and walk across the street and still stream Redbook audio without interruption from my own access point.

Battery Life & Charging

Astell&Kern claim up to 8 hours of battery life (playing FLAC at 16-bit/44.1kHz, Single Ended, Volume 80, EQ OFF, LCD OFF), which seems to be a reasonable number and perhaps a bit on the conservative side. In my own use I’ve gotten as much as 9 hours play back, at higher volumes, and with a mix of streaming lossless content over Wi-Fi and local files, with some full-size cans. Low impedance transducers impact that some.

Coming off a full charge, I will often get about an hour of playtime before the battery indicator moves from 100%, and then generally see about 11-13% battery consumption per hour.

Charging is via the USB-C socket and supports normal and “quick” charging. You can see the charge level with the player switched off by pressing the multi-function control in. The claims of ~5 hours for normal speed (5V @ 2A) and ~2.5 hours for quick charging (9V @ 1.67A) from 0% charge have been reliably met for me.

You can, of course, charge the unit while it is playing and even the normal charging mode will see the battery level increase even with spirited levels of play-back and streaming.

Comparisons

At the time I received the SP2000 also owned the Cayin N8 and the SP1000M, as seen below, and as such these are the two units I was most interesting in comparing against. Partly because my intent was to upgrade from the SP1000M (as my portable-streaming player) to the SP2000, and also because I was intrigued to see if the SP2000 might also replace the N8.

SP1000M (Lapis Blue)

Usability and the overall operational experience is identical to the SP2000, excepting the higher weight, larger size, and the larger, slightly more easily navigable, screen on the newer unit; so I am not going to cover that again here.

The most noticeable difference when comparing to the SP1000M is found when driving full-size headphones, particularly those of higher impedance, where the extra drive/grunt/power available from the SP2000 results in a more convincing, dynamic, authoritative, and clean delivery with impressive control with very large swings in energy and level.

In my DAP usage, I am often driving high-impedance/more demanding headphones rather than IEMs, so this additional power, particularly the higher available voltage, allows more headroom and higher, clean, peaks, and without the overall sound becoming compressed vs. the SP1000M which could start to reach its limits here.

Very small dynamic fluctuations, for example with an edgy, emotion-filled, delivery of grittier male vocals (such as Leonard Cohen or John Lee Hooker) are more readily audible vs. the SP1000M, resulting in a richer and more finely textured/nuanced vocal rendering. Of course, this ability is retained with instruments and their timbral reproduction better exposes low-level resonances and uneven friction.

And there is, perhaps, a more stable projection of imaging and stage, but the difference here is quite small and was really only discernible when feeding headphones that can render proper dimensionality, such as the MySphere (fed directly), or when testing the output of the SP2000 into my SR1a setup via a power amplifier.

The differences are audible with IEMs too, but the biggest isn’t as much of a gain there as the SP1000M already has plenty of power for every IEM I’ve ever tried and on its own I don’t think is sufficient to act as an upgrade-driver -but is certainly worth having if just buying from scratch or moving up from a rather lower-tier product.

Cayin N8 (Brass Black)

Usability of the N8 is good, but not as fluid or flexible from a navigation perspective as the SP2000. The touchscreen is responsive, and similarly vibrant, but it’s smaller than it looks (the glass front-panel is 50% larger than the actual display) and is not as sharp as the A&K’s. This isn’t an issue for album or playlist-based play, but if you’re the type to keep picking and choosing songs as you listen, the SP2000 has a definite advantage.

Build is similarly solid and luxurious feeling, with similar caveats regarding dial-movement I have for the SP2000’s dial. The N8 package is rather more comprehensive; including multiple cables and adapters, for its greater degree of connectivity – including full size balanced XLR outputs. I do appreciate the 4.4mm “Pentaconn” balanced output on the N8 more than the 2.5mm TRRS of the SP2000.

Adjustable gain and power output (for higher impedance cans), is a useful feature here vs. the fixed settings on the SP2000. It makes it easier to keep things in a sensible, usable and safe range for whatever IEMs or headphones you are listening too at the time. With the A&K you simply keep turning the volume up until you reach the level you want, with no specific gain settings.

Sound-wise, the N8 and the SP2000 are not worlds apart when using their respective balanced outputs and with the N8 in solid-state mode; and depending on the music and headphones in use it can be hard to tell them apart at all at times. At a push I’d say the SP2000 has more of a “reference” signature, slightly better noise performance (not