The reality is that 1tb is an unattractive proposition. I have more than 1tb. I don’t listen to it all, I just don’t have to worry about thinning it out. My Innuos has a 4tb hard drive, so I will never have anything to worry about. The downloads I purchase tend to be 24/192 and about 3gb each.
Innuos produce very sophisticated servers, Auralic very sophisticated streamers and Roon very sophisticated music management (multi-room) because they have been doing it for 10 to 15 years and are on 4th and 5th generation products. Moreover, they have been doing it funded by successful businesses and lots of user feedback. The have also had very low priced entry points, the Auralic Aries Mini cost me £350 new, I think it was $500, a killer product, had it about 6 years and still in daily use.
PS Audio has, for better or worse, decided to take on server, streamer and software in a single first generation device. My feeling is that is high risk because it puts them up against seasoned competition and there are very few companies that get it right first time. One exception is the dCS Bridge, but it is not a server, but optimised for network data. It is also Roon Ready.
My view is that the make or break will be the software/user experience, not the hardware. It’s got to be as good as what people are already using, most of which is already very good indeed.
As a counterpoint, I thought I’d mention that Channel Islands Audio (Dusty Vawter’s outfit) has chosen to go the other direction. They recently entered the market with their DMC-1 at $1995 US, a digital music player/streamer with their own controller app for mobile devices, but they’ve chosen not to take on the server role at all. There’s no onboard storage of any capacity, though they’re going to offer a companion CD transport as an extra-cost option, so an owner can rip files to their own storage device(s), which the unit can access over the network.
I guess my point is simply that every manufacturer makes their own choices about the market’s wishes, needs and price points, and goes from there.
When a device has no on-board storage it’s called a streamer. Loads of companies have gone that way - Linn, dCS, Auralic etc. The whole idea of a server is that it stores data, moreover it stores data and serves it to whoever asks for it (with permission). If a device cannot offer that data to multiple users, and if the storage is not effectively limitless, it’s not really a server. Perhaps this device should be called the Octave Streamer and there would be no confusion.
I agree with you, Steven. The make or break will be the software. And yes, to your point, if we’re making a server it has to be big enough to actually serve enough media. Appreciate the input.
I wonder what, in reality, they do to make an SSD “audiophile quality”? Is there any information presented on the particulars to this or is it just from their website?
For what it’s worth, Innuos feature mechanical isolation of the drive in a damped chassis, EMI insulation, three asymmetric feet with damping, vibration material lining the inside of the top plate and the unit his of course fanless. The drives use a proprietary type of format and a large chunk of RAM is used for buffering so the disc is rarely if ever active when music is playing. Innuos buffer streaming and internal storage. If you connect a usb drive and play from that, it is played direct without buffering. Don’t think anyone does that.
I know Innuos uses “special” HDD/SSD disc formatting and algorithms for loading files. No details are provided as the methodology is proprietary. A hint as to what may be going on is slower than normal rip times.
Innuos has optional fast and slow rip speeds, you can happily listen to music with the slow rip active, not the fast one as it is quite noisy. The quality is the same and the fast one is very fast. Mainly because they use ultra-reliable TEAC drives.
Yes the Teac is very quite at slow rip speeds, and having the 2Tb HDD is fantastic, especially with the RAM in the playback loop. As I am a hoarder, according to some, I am pursuing a HDD upgarde with Innuos, hopefully to increase capacity to at least 4Tb. As I am involved with a group that show cases jazz bands in the area I get quite a bit of digital product to review. Typically it gets loaded on to my ZENith Mk3, which is full at present, 1.8Tb mol.
Most of the bands we work with offer very limited release items in digital format only.
Below is a summary of information about Melco’s audio grade ssd’s as far as I have found it.
With Buffalo Technology as parent company Melco is one of the largest computer peripherals manufacturers in Japan. With their longterm experience and R&D (being big enough for funding/personnel) one can assume they know a thing or two about routers, switches and storage devices.
Melco states they re-engineered ssd’s to reduce speed with precision slow data writing and self-integrity check operating in the background for greater stability. They are produced in low volumes (hence the price), use higher-grade flash memory and are electrically isolated via a separate PSU that includes audiophile grade film capacitors just for the ssd’s. Audio grade ssd’s should address what they state to be limitations of conventional ssd’s and differ by removing some strategies usually present on common ssd’s (all these are not needed in audio applications):
wear leveling; ssd memory has limited durability. To limit write access to individual memory cells, the drives‘ control units scatter data across the entire media and move it regularly which leads to data fragmentation, so the server has to work hard to put everything back in the correct order. No need for this in the write once/read many audio environment (music data is written infrequently and rarely modified, rendering the usual ssd lifetime optimisation algorithms as inappropriate with impact on data integrity, as they operate in burst as required). Audio grade ssd’s use low level and consistent management algorithm that has zero impact on data integrity;
compression; used to boost capacity and cover damaged elements via redundancy. This happens without any loss but requires computing power to decompress when required again;
acceleration; in common fast ssd’s this pushes the flash memory, thus creating more power supply noise.
the N1ZS has mechanisms -especially faster processors- to circumvent weaknesses and take advantage of ssd’s. In-/external data communications are differentiated with separate isolated power supplies.
The above may be partly marketing nonsense but given the experience with my Melco I’ve only reason to believe these audio grade ssd’s are not just common ones provided with different stickers.
These and the way they’re handled really do miracles to my music library
I use a streamer first and foremost for optimal sound quality and ease of use. Direct accessibility from pc and tablets (UPnP), Roon-, Qobuz- and Tidal-ready, minimserver for library management, user-friendly (incremental) backup facilities and the ability to format / restore to factory settings in case of emergency is enough.
To me there is absolutely no reason to bother with partition management on the 1T audio grade ssd in the Melco.
