Follow-Up March 2014

John Atkinson returned to the AK100 in March 2014 (Vol.37 No.3):

After I reviewed the Astell&Kern AK100 portable media player in August 2013, rather than return the review sample to A&K, I bought it. Yes, I very much liked what this little gem of a player did—much as I appreciate how much music I can store on my 160GB iPod, it sounds harsh and lacking in detail compared with the AK100. However, one thing the iPod does that the AK100 didn't was support gapless file playback, which is essential for such works as through-composed operas and symphonies, or even albums like Sgt. Pepper and Abbey Road, where there is no pause between movements.

Astell&Kern addressed this omission in June 2013, after I had submitted my review text for publication, with the release of the v.2.01 firmware. This is supplied as a binary hex file that you copy into the AK100's root folder with the player connected to your computer as an external USB drive; when you disconnect and reboot the AK100, it installs the new firmware. (Don't do what I inadvertently did when I installed the new firmware and set the default language to Korean—it took me a long time trying to decipher Korean messages in order to reset the default to English!)

The next upgrade, in July 2013, was v.2.10, which added the capability of using the AK100 as a USB DAC, offered as a menu item when you connect the player to a computer via USB. Then, in December 2013, Astell&Kern released the v.2.30 firmware, which allows the AK100 to play DSD64 files in both .dff and .dsf formats. I was somewhat puzzled by this new functionality; according to its datasheet, the Wolfson WM8740 DAC chip used by the AK100 doesn't support DSD input data, just PCM data with word lengths of from 16 to 24 bits and sampling rates up to 192kHz.

Nonetheless, I copied some DSD64 files to my AK100, which duly recognized them and added them to its library. When I selected one of these files for playback, "DSD" appeared on the AK100's display and I heard music through the headphones, as expected. However, when I connected the AK100's TosLink output to an Auralic Vega DAC, which I reviewed in February, it became apparent that the AK100 was outputting 24-bit PCM data sampled at 88.2kHz. So while I appreciated its new ability to play DSD64 files, the AK100 is actually transcoding DSD data to high-resolution PCM, in order to be able to present them to the Wolfson DAC.

Is this anything to be concerned about? Not from my auditioning. Acoustic Sounds now offers DSD downloads, and I purchased some favorite albums in the format, including Shelby Lynne's Just a Little Lovin' (DSD64, Lost Highway). Played through the AK100, that little sob in Lynne's voice in the title track was just as heart-wrenching, and the sense of space in this superb recording was just as convincing as they'd sounded with the Auralic Vega, whose ESS Sabre32 9018 DAC supports native DSD decoding. The space around Joe Morello's drums in his solo in "Take Five," from the Dave Brubeck Quartet's Time Out (DSD64, Columbia), was palpable with the upgraded AK100, even though, in this early stereo recording, Morello remains firmly planted in the left channel.

One thing I did notice was that the touchscreen was slower to respond when the AK100 was playing a DSD file than it was with PCM files, presumably because the conversion of DSD to PCM uses more of the player's processing power. Astell&Kern warns that playing DSD files will shorten the time between battery charges, which also suggests that greater processing overhead is involved.

The v.2.30 firmware upgrade brings the $699 AK100 closer in its capabilities to Astell&Kern's top model, the AK120, which costs $1299. The AK120 offers twice the internal memory (64GB vs 32GB), and the ability to accept 64GB Micro-SD cards; it uses two WM8740 DAC chips rather than one to increase resolution, flanks the exposed volume-control knob with flanges to provide it with some physical protection, and offers a significantly lower output impedance: 3 ohms vs the AK100's 22 ohms. (In December 2013, A&K introduced the AK100 Mk.2, with the same 3 ohm output impedance as the AK120.)—John Atkinson

Company Info
Astell&Kern, Korea
US distributor: iRiver Inc.
39 Peters Canyon Road
Irvine, CA 92606
(949) 336-4540
Article Contents
Share | |
Comments
gld3gld3's picture
"Ten years later, while the

"Ten years later, while the current top-model iPod features a 160GB hard drive, it still can play only files with sample rates of 48kHz and below and a maximum bit depth of 16. Those of us with a growing library of high-resolution files are therefore restricted to playing them in our big rigs at home."

Not exactly. There is at least one option for those wanting to play higher sample rate/higher bit depth files on their ipods.

My 5th gen ipod is currently playing "Happlessness Blues" by Fleet Foxes at  sample rates of 96kHz  and bit depth of 24 (FLAC). How? Rockbox. 

John Atkinson's picture
Rockbox

gld3gld3 wrote:
My 5th gen ipod is currently playing "Happlessness Blues" by Fleet Foxes at  sample rates of 96kHz  and bit depth of 24 (FLAC). How? Rockbox.

Rockbox does indeed add a lot of functionality to iPods (and other portable players). However, it can't overcome the hardware limitations of the player. Even if you can play a 24/96 files, as you report, the iPod's DAC chip is limited to sample rates of 48kHz and below, so the file will be downsampled.

In additon, while Apple's AAC and Lossless codes will preserve 24-bit resolution, the iPod's DAC is a 16-bit part, so will truncate the word length by the 8 LSBs. For example, if you look at the spectrum at www.stereophile.com/content/ipod-classic-160gb-1khz-90dbfs-16-bit, this shows my 2010 iPod Classic playing a dithered 16-bit tone at -90dBFS. Everything is correct, and the noisefloor is free from harmonic spuriae.

By contrast, the spectrum at www.stereophile.com/content/ipod-classic-160gb-1khz-90dbfs-24-bit shows what happens when the iPod plays back a 24-bit dithered tone at the same level encoded with the ALAC codec. The appearance of the odd-order harmonics is due to truncation of the 24 bits to 16.

So yes, it may be possible to play 24/96 files on this iPod running the Rockbox software, but the file will be downsampled and truncated. The AK100 doesn't suffer from these hardware limitations.

John Atkinson
Editor, Stereophile

jcborges's picture
iPod 24-bit

How were your measurements of the iPod taken? I ask because looking at the specs of the iPod DAC at http://cirrus.com/en/products/cs42l55.html it is definitely capable of 24-bit, although it is limited to 48khz sampling frequency.

Why would Apple put a 24-bit capable DAC in the iPod and have it truncate to 16? And how could it be done? And couldn't Rockbox bypass that to access the full bit depth available on the chipset?

Thanks and regards

John Atkinson's picture
Re: iPod 24-bit

jcborges wrote:
How were your measurements of the iPod taken? I ask because looking at the specs of the iPod DAC at http://cirrus.com/en/products/cs42l55.html it is definitely capable of 24-bit, although it is limited to 48khz sampling frequency.

I took a 24-bit AIFF file representing a dithered 1kHz tone at -90dBFS and prepared from it WAV, ALAC, and AAC at 320kbps files. The WAV file wouldn't play in either my iPod Classic 160GB or my iPhone 3GS, but from the other three versions I expected to get a spectrum that looked like the red and blue traces in fig.5 at www.stereophile.com/content/astellampkern-ak100-portable-media-player-measurements.

Instead, from both players and with AIFF, ALAC, and AAC versions of the data, I got the spectrum shown at www.stereophile.com/content/ipod-classic-160gb-1khz-90dbfs-24-bit. The characteristic pattern of odd-order harmonics indicates truncation of the 24-bit data to 16. By contrast, the AK100 playing both the AIFF and ALAC versions of the data gave the correct spectrum shown in the review.

jcborges wrote:
Why would Apple put a 24-bit capable DAC in the iPod and have it truncate to 16?

I have no idea. This was not what I was expecting, as the new "Mastered for iPod" protocol recommends using a 24-bit master to prepare the AAC version, which in turn implies a 24-bit datapath..

jcborges wrote:
And how could it be done? And couldn't Rockbox bypass that to access the full bit depth available on the chipset?

That is possible, but as you point out, the chipset is still restricted to sample rates of 48kHz and below, contrary to what the original poster claimed.

John Atkinson
Editor, Stereophile

jcborges's picture
iPod DIGITAL playback

First, thanks a lot for your answers. But you left me wondering about something. 

If understood correctly, your measurements were made in the analog domain. What about digital?

If you play a 24bit/48khz ALAC file through a digital dock, what is the format of the PCM stream that comes out? 24-bit or 16-bit truncated? Have you ever checked or know someone who has, for example by connecting it do a DAC that shows the input format?

In the other words, if you have 24/96 files that you want to put in an iPod to playback digitally, should you just resample them to 48khz or should you apply TPDF dither to avoid truncation? Does it matter? 

 Thanks and regards

John Atkinson's picture
Digital Output

jcborges wrote:
If you play a 24bit/48khz ALAC file through a digital dock, what is the format of the PCM stream that comes out? 24-bit or 16-bit truncated?

If you play a 24-bit AIF or WAV file, the AK-100's digital port outputs 24-bit data. But I will check with ALAC data.

John Atkinson

Editor, Stereophile

John Atkinson's picture
A clean bill of digital health

John Atkinson wrote:
jcborges wrote:
If you play a 24bit/48khz ALAC file through a digital dock, what is the format of the PCM stream that comes out? 24-bit or 16-bit truncated?
If you play a 24-bit AIF or WAV file, the AK-100's digital port outputs 24-bit data. But I will check with ALAC data.

Okay, played back a dithered 24-bit 1kHz tone at -90dBFS, encoded both as AIFF data and as ALAC data. In both cases, all 24 bits were active in the AK100's optical output and the digital-domain spectra of the two codecs were identical, with a noisefloor at the 24-bit level.

So the answer to your question is yes, when the AK100 plays a 24-bit ALAC file, the digital output is true 24-bit as it should be.

John Atkinson

Editor, Stereophile

JohnnyR's picture
What a Bargain

*eyeroll*

So having a player that has crappy high output impedance and doesn't have gapless playback is a good thing? For the money I would expect a lot better. What's wrong with using FLAC in other players that cost less and have gapless playback to boot?

Another laughable product that Atkinson likes.cheeky

John Atkinson's picture
Re: What a Bargain

JohnnyR wrote:
So having a player that has crappy high output impedance and doesn't have gapless playback is a good thing?

As I wrote, the most recent firmware upgrade allows gapless playback. I examined the effect of the 22 ohm output impedance at length in my review - "crappy" is a mischaracterization of the relatively small changes in frequency response that result from this impedance, especially when you consider that many headphones and in-ear monitors have boosted low frequencies.

John Atkinson

Editor, Stereophile

JohnnyR's picture
Quit Defending High End Crapola

For $699 you should get a lot better response and features. Sorry but changes from 1 to 2 dB would be scandalous in a preamp or power amp but since it's this THING then it's "okay". oh brother.blush Seriously, lower priced players don't have this problem so it's a FLAW easily avoided if they had designed it right.

ChrisS's picture
Because Uncle Joseph says so!

Comrade,

There must be something about your Stalinist upbringing that makes you believe that the perfect audio product exists somewhere out there, when really that kind of perfection only exists in your imagination.

Site Map / Direct Links