Lumin P1 streaming D/A preamplifier

I start this review with a confession. I have consistently found that when I play CDs on a transport and feed the digital data via AES3 (AES/EBU) to a D/A processor, the music has more drive, particularly at low frequencies, than it does when I send the same 16/44.1 data to the same D/A processor via my network. However, these days I almost always stream my music rather than playing physical discs. Not only is it much more convenient; the wealth of metadata available with Roon 1.8 is addictive. I also have convenient access to much more music: my growing library of hi-rez files and hi-rez music streamed from Qobuz.

I don't understand why there should be an audible difference between AES3 and network audio. The data are the same in both cases, and the reclocking of those data when fed to the processor's DAC chips should be identical. And if the reclocking isn't identical, the timing accuracy should favor the network connection, where the clock doesn't have to be extracted from the datastream, as it does with AES3. In any case, for the reasons stated above, I always keep my ears open when I try new streaming D/A processors, like the one under review here, the $10,000 Lumin P1 from Hong Kong–based Pixel Magic Systems.

The P1
The elegant-looking P1 offers a complete set of digital inputs—AES3, S/PDIF (coaxial and TosLink), plus USB and Ethernet (electrical and optical) with full MQA decoding—adding to that balanced and unbalanced analog inputs, one HDMI 2.0 input, and three ARC-enabled HDMI 2.0 outputs with 4K video passthrough. The P1 offers balanced and single-ended analog outputs and S/PDIF (BNC) and USB digital outputs. The P1's gently curved front panel displays the volume setting or metadata for streamed music being played, including bit depth and sample rate, though not the album cover. The Roon Ready P1 can be controlled with Roon or with Lumin's own app for iOS and Android devices. It comes with a slim, solid remote control.

322lumin.bac

The elegance continues inside the case. The linear power supply for the analog circuitry is based on a shielded toroidal transformer. The digital and analog input circuitry is housed on one printed circuit board; the dual-mono, solid state output circuitry (balanced and unbalanced) and the D/A converter chips—a pair of ES9028PROs capable of processing PCM data up to 384kHz and DSD data up to DSD512—are on a second board. The analog outputs are coupled to the outside world with a pair of premium-quality Lundahl transformers.

322lumin.bac2

The USB input uses an XMOS controller, and the DAC chips are clocked with what Lumin refers to as "Femto Crystal Oscillators" with "precision FPGA distribution." The P1 has a volume control based on Leedh processing (footnote 1), which was invented by Gilles Millot of Acoustical Beauty, a French high-end loudspeaker manufacturer. Leedh minimizes the number of additional bits introduced in mathematical operations to reduce or eliminate the truncation-related loss of information. The P1's Leedh volume control operates with 32-bit precision. Leedh processing is the P1's default setting; it can be disabled with the Lumin app for with/without comparisons.

322lumin.ins

Installation
When I connected the P1 to my network, it was recognized by Roon 1.8 as "Lumin X1." (The X1 is the P1's bigger brother.) The small image of the P1 on Roon's Audio Settings page confirmed that it was a legitimate Roon Ready device. I enabled the P1 as a Roon endpoint, setting the volume control to "Device control" and MQA decoding to "Decoder and renderer." Lumin recommends burning in the P1 for 400–500 hours; with outputs not connected, I streamed audio to it for five days. At the end of that time, I connected the P1's balanced outputs to my Parasound Halo JC 1+ monoblocks, reduced the volume control to a relatively low value, selected a file with Roon, and pressed "play."

Nothing.

I told Lumin and Source Systems, the company's US distributor, about the problem (footnote 2), and a second sample was shipped tout de suite. I set up the new P1, made sure its firmware was the current version, and made the necessary connections. Success! I had sound via both digital and analog connections. I repeated the burn-in and started my critical listening.

322Lumin-App-Setup-Screen

The L1
Along with the second sample of the P1, Source Systems sent this diminutive, network-attached UPnP server, styled to match the P1 and powered by a 12V wall wart. The L1 is priced at $1400 for the 5TB version; a 2TB version is also available. The L1 is a network-attached storage device, but not a conventional one. To load it up, you must copy files from a computer via a USB 3.0 port. Once it's loaded up, the L1's network port can be connected to the user's router and the P1 can see it and access music files.

One wrinkle for this Mac user was that the L1's 5TB hard drive is NTFS-formatted "due to drive size and Windows recognition," and Macs don't recognize NTFS drives. Lumin says that the L1 was intended to be populated by simply dragging-and-dropping files from a PC—"it's not designed for anything else." To copy files from a Mac, the L1 can be reformatted as a FAT32 drive (for the 2TB version) or exFAT (for the 5TB), or third-party apps could be used to allow the drive to be recognized by a Mac. For simplicity's sake, I copied my music library onto the L1 using a USB connection from a Windows 10 PC that lives in the test lab.

Once the L1 has been populated and connected to the network, it scans the files to create the metadata it displays with the Lumin app. During scanning, its status LED blinks blue. 100 albums take approximately two minutes to scan.


Footnote 1: See processing-leedh.com/, especially this AES paper.

Footnote 2: After this first sample had been returned to Pixel, they discovered that the linear power supply had been wired for 220V operation instead of 115V.

COMPANY INFO
Pixel Magic Systems Ltd.
US distributor: Source Systems Ltd.
San Clemente, CA 92672-6000
(949) 369-7729
ARTICLE CONTENTS

COMMENTS
rwwear's picture

Rather expensive but at least it has HDMI inputs unlike most other modern DACs.

Kursun's picture

I am disillusioned by some of the fancy and expensive equipment Stereophile reviews.
A $12.000 phono preamp with only 3 db overload margin...
A $10.000 streaming preamplifier with only 1.2 kOhm input impedance (really?)...
These are all very serious design flaws.

C'mon, your readers surely deserve better equipment, without getting ripped!

Jack L's picture

Hi

Any publishers need advertising sponsors to keep their journals running so the readers can read them. They are running a business not a charity.

So they would not & could not turn down any vendors' new components for review. This is their business 'law'.

Stereophile is a much better journal as it got a technical dept to provide its readers the bench measurement data vs the specs provided by the equipment manufacturers.

Yes, you are right: 1,240 - 1,260 ohm is rediculously way too low for any decently designed audio inputs. No wonder the Chinese manufacturer does not publish it. Who would spend 10 grands to buy such lousy design & built cheap in China ????????????

Jack L

Jim Austin's picture

So they would not & could not turn down any vendors' new components for review. This is their business 'law'.

We routinely choose not to review components offered to us for review, by advertisers and nonadvertisers alike. (There's no difference.) This should be obvious when you consider that far more components are released than that Stereophile reviews.

Jim Austin, Editor
Stereophile

Jack L's picture

Hi

I see. I stand corrected.

Jack L

TJ's picture

Thanks John, beautifully engineered beyond a doubt, and I always look forward to your intriguing analyses of the sound quality. But on the topic of SQ, what about digital room correction? That's becoming a core component of a modern, transparent audio system. In comparison, a network streaming DAC with DSP takes SQ to a new level, eg the miniDSP SHD with Dirac 3 and much else at almost a tenth of the price. A paradigm change?

Glotz's picture

JA's measurements, but not his ears... huh.

Sounds like there is no trust.

His review is important; these comments above are not.

I would really love for JA to review the latest Wadax Network Streamer and its 'wave-shaping' controls for getting closer to a more perfect analog waveform. I think it might address some of his consternation regarding the AES3 and network streaming comparisons he brought up in the review.

rex's picture

Peter Wklie, Project Manger at Lumin, who seems to have the job or answering all enquiries about Lumin
on AudioShark: answered Why their might be sonic benefits of the L1 over the Nucleus.

He writes:
"Unlike a typical NAS, the L1 eliminates all non-essential network services, so there is less network traffic that is unrelated to music transfer, and the L1 itself will not be affected by background detection of network file (and other) services by computers. Since digital audio is heavily affected by timing jitter, I have always wondered whether this aspect of L1 can result in audio improvement, however minimal it may be."

I have both the L1 and the Nucleus with ssd drives and I do hear an improvement of the L1, wether playing files or streaming from Roon. However it's minor and not as easy as to detect the bump in quality as compared to playing anything from a computer.

The fact that Mr. Atkinson can hear a difference is no surprise.

bradchaus1960's picture

you dont suppose you are actually tossing yourself when you listen to this thing or writing the review, and waiting for the advertising contract to be signed?

John Atkinson's picture
bradchaus1960 wrote:
you dont suppose you are actually tossing yourself when you listen to this thing or writing the review...

No.

bradchaus1960 wrote:
...and waiting for the advertising contract to be signed?

Sigh. Although I have the title of Technical Editor, I am a freelance contractor and have no connection with Stereophile's advertising department and no insider knowledge of which companies advertise. And even when I was the magazine's editor, there was no connection between advertising and editorial content. See this essay I wrote in 1996: www.stereophile.com/asweseeit/366/index.html. Editor Jim Austin practices the same policy.

John Atkinson
Technical Editor, Stereophile

X