Tips for setting up an Audiophile NAS system with Aurender’s John-Paul Lizars

The crushing basslines and powerful vocal overdubs torturing eight-inch transducers stopped so suddenly that there wasn’t even an echo of a note left hanging in the air between my speakers.

I hit play on the app interface residing on my iPad from the comfort of my sofa and it was back. The song pounding into the room as I controlled the ones and zeros and sat, scrolling through the thousands of album and song titles available to me from the touch screen cradled in my palm. Such is the experience of music playback in this modern computer-audio era.

Using computers as part of the delicate circuit path in a hi-fi setup is a relatively new phenomenon in this hobby and getting the most, sound-quality wise, out of one is a tougher proposition than most when it comes to a source  component because computers aren’t designed with high-fidelity in mind.

There’s a lot of (sonic) compromises to be made when your music source is what many digital-audio experts consider electronically noisy PCs or laptops accessing cloud-based or local-area network drives for music be they mp3s, 16/44 FLAC or high-resolution PCM files.

As such, there’s a growing market segment of manufacturers who are focusing solely on digital, audiophile-centric source components (servers/streamers) that are all about removing the PC from the equation. Source devices with specialized chassis, RF and EMI isolation casework, processors, circuit design and hard drives all holistically researched, developed and built with SQ as their primary priority. Companies like Roon, Innuos, dCS, Bel Canto, Simaudio and Aurender (to name but a few) come to mind as those who have invested heavily in the latest technology, production tooling, digital and mechanical engineering to bring products to consumers that don’t want that noisy computer anywhere near their ones and zeros.

So, with all that in mind, I decided to talk to someone who has knowledge and experience when it comes to hard drives, network-attached (NAS) storage devices and some of the how-to when it comes to a proper, sonically-sound methodology of implementation for storage, access and retrieval of data of the audiophile sort. Not spreadsheets, Word documents, Youtube videos or photo-correction software, but simply the retrieval of music files.

So I hit up John-Paul Lizars of Aurender to go over some of the basic and not-so-basic questions that surround NAS drives and in particular how Aurender approached these problems, as that is his particular field of expertise. Someone from dCS or Innous may have varied answers and I welcome any responses from other companies who specialize in this type of technology. Lizars’ response? “You want to talk SQ of HDDs? You my friend, are opening a can of worms!”

An apt response no doubt as many consider bits-are-bits and any cable or digital interface will work and should not affect sound quality in the least. A line of thinking which seems to enjoy vigor among audiophile-related Internet forums. But, in the real world and in my personal experience (and the experience of many others whose opinions and education I respect and know in the high-fidelity industry) this is simply not the case. I put forward that digital-music files will, and often do, have sonic-reproduction differences when played back from varying source media. Who has heard differences between DSD, FLAC, WAV or ALAC file types? (I’m raising my hand, you just can’t see it) And that’s just the file type, (provenance of where/how the file was sourced plays a huge role in how it sounds IME) never mind whether it is being accessed from a cloud-based source (Tidal, Qobuz, Spotify for example) or an internal HDD (be it solid state or spun-up), a NAS or even a self or bus-powered USB HDD or thumb drive. Do these differences add up to the analog equivalent of sonic impact regarding the type of stylus, tonearm and platter/drive-system one utilizes on a turntable?

Here is my Q&A with John-Paul Lizars

Rafe Arnott: Can you talk to me about where you’re coming from in the industry when it comes to hard drives and their implementation and what it is you’ve learned about them?

John-Paul Lizars: “I have experience with the sound differences of hard drives. Many years ago I actually did a pseudo-scientific test of my own on SQ differences between several brands of hard drives in a Synology 213+ NAS. I discovered radical differences in sound not only between brands, but between two different/same brand/model drives. This all dates back to my “computer audio” days of repurposing IT stuff to make music and I’m sure glad that’s all in the past.”

RA: Why is knowledge of how circuit architecture involving data retrieval via HDD (Hard Drive) or SSD (Solid-State Drive) important to SQ and how does Aurender approach this key aspect of design in its products?

JPL: “Harry Lee who developed and conceptually designed the Aurender music server must have known about this phenomena because key to Aurender’s approach is caching which addresses this issue head on. All Aurender music servers have both a conventional spinning HDD and an SSD. The music library is stored on the HDD and when a command to play music is given, the music selected is moved from HDD to the SSD and playback begins. The benefit is twofold: little wear and tear on the HDD and no inconsistent sound quality because the HDD is not used for direct playback. Every Aurender uses a playback SSD which we feel sounds best. The size of the SSD does vary depending upon the model: 120GB, 240GB or 480GB SSDs are employed and all are large enough to hold numerous albums or massive playlists in the queue. 

Internal drive arrangement of Aurender N10

“Also, even if you use a USB thumb-drive, USB hard drive or NAS with an Aurender, you still get caching. Aurender doesn’t care where the file-based library is stored, internal or external, everything gets cached. This is one of the very special features employed to thwart the deviation of sound found with HDDs. We choose drives based on their reliability grade. For the models which have 3.5-inch drives we use enterprise-grade Western Digital Red and for 2.5-inch drives we use Seagate. In the new Aurender Content Server – ACS10 – we use Seagate's Iron Wolf for the 24TB model and WD Red for the 16TB and 8TB models.”

RA: How does configuring a NAS drive for the varying types of RAID (Redundant Array of Independent Disks) affect performance and most importantly, sonics?

JPL: “Frankly, I don’t know for sure, but theoretically it shouldn't. Being an Aurender user, it’s of no consequence because of, as already explained, Aurender's caching system. When caching is employed it doesn’t matter where the file library resides or how it’s configured because the playback you hear is from the SSD. The Synology NAS I use is strictly a scheduled data backup device to the Aurender and is configured using their Hybrid RAID (SHR).”

RA: With some types of RAID implemented there is a level of backup, or redundancy for your thousands of carefully-curated songs, albums, album artwork and playlists. How important is this to take into account when setting up a local-area network? Also, many network-attached storage devices can be tied-in with an UPS (Uninteruptible Power Supply). Is this type of backup more in the domain of information storage networks for hospitals, law firms, etc. than for the burgeoning digital-based playback audiophile?

JPL: “You can never be too careful when it comes to safeguarding your music library, so I’d say it’s very important. And cloud services or offsite storage can provide added security from disaster as well. UPS protection from power outages is critical to prevent hard drive corruption and therefore should be considered mandatory for NAS drives. Among its many features, the Aurender ACS10 Content Server functions as a music-dedicated NAS and it incorporates a kind of low-noise UPS using super capacitors that will gently shut the unit down under abrupt power interruption.”

RA: How much of a factor, in your opinion, is noise when it comes to NAS-drive setups for audiophiles? I can’t think of anyone who would want to add noise to their system-playback chain. Is this a real concern in implementing a NAS? Is the trick to simply have the drives in a separate room?

JPL: “Most NAS systems are quite loud, especially the ones with four bays or more. The constant-spinning drives and fan noise preclude these devices from the listening room. Being networked-attached devices does make it quite easy to relocate them remotely.  “

RA: What about the ability to remotely access your drives via an app, an FTP (File-Transfer Protocol) server, like FileZilla, for example, your smartphone, laptop or a PC elsewhere in your home. How key is that to dealing with possible file corruption issues, data damage to a drive, or just routine HDD or SSD maintenance, or to add or remove music files at one’s discretion?

JPL: “Once again, best practice in backup protocol dictates you should have three copies of your library. One, the library you use daily. Two, a locally accessible copy kept on a NAS or USB HDD. The NAS is useful here for its ability to perform scheduled backups you don’t have to think about it when adding new content to the library. Unlike a USB HDD which would have to always be manually updated. The third copy should be in the cloud or on storage located remotely. With hi-res streaming services so accessible on mobile devices, accessing content from drives via FTP doesn’t seem worth the trouble to me. For hi-res music on the go, you can’t beat the convenience of Tidal or Qobuz.”

RA: What about HDD or SSD format options? Some say it doesn’t matter, as the file system is only ‘seen’ by the the NAS OS (Operating system). But there are Windows-based networks, others are Mac, Linux or UNIX-based systems. Is there a sonic difference in your experience associated with formatting options and if so, what did you experience when you switched between different drive-formatting options? Which OS seems to work best in an audiophile, SQ-first network ecosystem?

JPL: “I don’t believe the formatting of an HDD (Mac, Windows NTFS, FAT, FAT32, exFAT, Linux Ext3, Ext4) or others has any appreciable affect on the sound quality. But operating systems can because the OS typically does have other add-on programs along with the OS core. Aurender does not want any other processing unrelated to music playback running in the background as that has negative consequences to the sound quality. With Linux, engineers with enough knowledge can remove unnecessary applications and system programs thereby creating an OS optimized for music playback which, results in much better sound quality than either a Windows or Mac platform.”

RA: How crucial is the choice of 100/1000 network switches, or CAT5 or CAT6 Ethernet cabling and their inherent price differences, for sound-quality playback in your experience? I’ve seen switches range in price from $50 to well over $500, does the cost factor outweigh the benefits, or is it about striking a balance within the network’s architecture?

JPL: “I consulted Aurender’s Tech Support Manager Ari Margolis on this issue as he is a networking professional and he suggested that there is the Law of Diminishing Returns at work here. Quality ethernet cables and, even more importantly, better ethernet switches and routers can make a significant difference in a highly-resolved playback system. The ethernet cable is especially susceptible to carrying RF noise and using high-quality shielded cables from a reputable supplier can significantly lower the noise floor. We are starting to see highly modified “audiophile” switches and routers with very sophisticated power supplies and very precise clocks on board all for the same reasons. Surprisingly, this category of cable, switches, routers and LAN isolation upgrades has been largely ignored until recently. None of this affects the accuracy of the transfer of packets of data across the Local Area Network, but noise from that network can be introduced to the system at several points.

“Whereas the goal of high-end audio is to lower the noise floor as much as possible, it follows that network cables and accessories are important links to better sound quality. Aurender’s newest model, the ACS10, offers a double-isolated LAN port for connection to the router. This eliminates the necessity of using an outboard filter or optical isolator for LAN noise suppression.”

RA: Lastly, is NAS going the way of the Dodo? Are its days numbered as more and more companies switch to internal storage options on their servers/streamers? Is a USB-based drive network a more viable option for future-proofing your network investment, or is Ethernet-based storage the way to go in the long term?

JPL: “Aurender is a strong believer in using built-in storage as it can be optimized to work as a part of the ecosystem as opposed to a using third-party external storage. This methodology allows for Aurender to be aware of and be in control of the total system behavior.  This results in appliance-like predicability, stability and reliability which I think music server users find appealing. Although I wouldn't say that the NAS drive days are numbered, just less relevant for purposes of audio playback. NAS drives are best when used for the IT purposes they were designed for, which is mainly data backup. That said, setting up a NAS can be a daunting experience for the neophyte and it’s for this singular reason Aurender developed the ACS10 to make audio-centric NAS functionality simplicity personified.

mentt's picture

Aurender claims that caching to SSD makes sure that files from NAS/internal hdd/USB hdd sound the same, but this is not my experience. Files from local hdd sounds better than files from NAS. Tested on Aurender A10