REFERENCE LEVEL DIGITAL AUDIO PERFORMANCE
Digital playback systems must have precise timing for the resultant analog signal to be accurate. But the precision of the ‘clock’ used in a digital playback system is undermined if the digital signal suffers noise interference. Whereas in analog playback, noise is heard as noise, in digital playback noise distorts the very sound of the instruments and voices, and in a way that is very unnatural to the ear/brain. Typically it adds glare and grain, and robs the music of its solid bass foundation to the music.
Antipodes Music Server designs are focused on optimising timing accuracy, and a big part of that is minimising electronic noise interference with the clock data. Unlike higher powered servers, Antipodes Music Server technology reduces electronic noise interference to such a low level that noise filtering (which adds its own set of problems, in particular robbing the music of energy) is no longer needed.
- Fanless, silent operation
- 2.5" slow-spinning HDDs using customised firmware
- Custom scripts & firmware to optimise the handling of the digital signal
- Two stages of digital signal buffering and re-clocking in the server
- High-end USB audio output card with selectable 5v power
- Custom-designed internal linear regulated power supply, with one rail
dedicated to the USB audio card (superior to our ears to any battery
solution) and, most crucially, fed from a uniquely designed power transformer.
Antipodes Servers do not include a wifi connection as wifi would generate large amounts of electrical noise inside the server and compromise audio performance. However you can connect via Ethernet cable to a wireless Access Point if desired without sonic penalty.
Headless (no screen, mouse or keyboard needs to be attached)
VortexBox Linux OS for easy remote maintenance from PC/Mac/iPad browser.
Custom scripts simplify setup to plug and play.
Simple click or tap to play with iPad/iPhone/Android/PC/Mac remote applications.
Easy remote management of music files using you choice of tools on a PC or Mac connected to the same network - music files on the music server are as easy to manage from your computer as if they were on the computer
MEDIA LIBRARY MANAGMENT
Automatically rips the main feature and language of movie disks, and stores the movie as a single playable file (MKV format) - to activate this feature you need to purchase a license (50 Euros via the VortexBox GUI) to use the pre-loaded software.
Automatically manages your music library:
- Rips & tags CDs, just insert the disk and it ejects when ripped
- Automatically gets coverart from internet databases while ripping
- Automatically organises ripped music into folders by artist & album
- And use any music management tool on a PC or Mac to adjust tags
(eg. iTunes, JRiver, dBpoweramp, MediaMonkey, Picard, etc).
Easily copy music files to & from the music server, or download directly
One-click to get coverart for music files that you copy/download
Simple backup and restore to a USB hard drive (not included), or use your favourite backup program from a PC or Mac
Rips in ‘paranoid’ mode for the best possible rip, to uncompressed flac, indistinguishable from wav files in listening tests, superior to compressed flac.
NAS access - store your music files internally and/or on your NAS to expand storage for a very large library.
Control playback through a USB DAC with MPD client applications, or with SqueezeBox client applications
Stream via Ethernet to Streamers or SqueezeBox products
Plays all music file types (unless DRM protected) to your USB DAC:
- Bit-perfect and gapless within the capabilities of your USB DAC
- PCM files up to 32bit/384kHz
- DSD64 & DSD128 (Double DSD) music files using DoP.
Video rendering is not performed on Antipodes Servers. We recommend streaming video from Antipodes Servers to video rendering devices, as this makes it simple and easy to avoid the audio and video getting out of 'sync'.
Antipodes Servers come with Plex Media Server and a DLNA server to stream to UPnP compliant video rendering devices, such as a Smart-TV, networked Blu Ray player, Roku, Boxee Box, WD Live etc. In this mode the Antipodes Server acts as a high quality NAS, with no setup hassles, auto-ripping, fanless/silent operation, superior sound and video quality, and rock-solid stability.
WHY BIT_PERFECT IS NOT ENOUGH AND TIME DOMAIN ACCURACY IS SO IMPORTANT IN DIGITAL
When you store a digital file you are storing 1s and 0s. But when you are transporting a digital file it is done using an analog carrier. The data is 1s and 0s but the method is either an electrical or optical waveform, and any such waveform is subject to potential noise interference. In any high-precision digital application, whether audio or guided missiles, being bit-perfect is not enough.
Being time-perfect is also important. Any clock irregularities or noise interference with the signal affects the precision with which the downstream stages can process the stream, and therefore the precision of the end result. Time-domain distortion in digital is often referred to as jitter. The downstream stages can usually extract the correct 1s and 0s but in a real-time system, such as in playing digital files, noise distortion can obscure the timing of the recognition of a change in the analog waveform that indicates the 1s and 0s, and the resultant analog output is distorted. Even very minor levels of jitter can affect the ability of the ear/brain process to make sense of what it hears, instruments and voices don't sound natural, and musical enjoyment is diminished.
Some digital engineers will disagree with the previous paragraphs, because in their abstracted view of reality digital is resilient to noise interference and able to reject moderate jitter. Other digital engineers will endorse what we are saying because in many fields, reducing noise interference with the waveform carrying the digital data is a major design challenge and critical to the precision of the end result. The argument, if there is one, is really about whether you can hear it.
These issues don’t matter if you are simply going to record these 1s and 0s on a hard disk at the destination. When stored, digital files have no jitter in them. Based on this, digital is argued to be ‘fixable’ by buffering and reclocking the signal, late in the journey, close to the DAC. That is, first write the 1s and 0s to a buffer (which does not store jitter information) and then clock it out with a good clock for a short journey to the DAC chip. This and similar stories have often been told to assert that a DAC is immune to jitter, or totally eliminates it.
But it is simply not true. One key reason why it is not true is that a buffering and reclocking step is by definition a digital process with two sets of clock data in it (even if a common clock was used), and that in itself generates noise interference. Because the writing to the buffer and reading from it occur at the same time, it re-generates at least some of the problem that it purports to solve. Buffering and reclocking is a good idea, and we buffer and re-clock twice in our Music Servers. But it only reduces the noise in a digital stream. It does not eliminate it. Listening tests in a good system make it clear that reducing noise in the digital stream improves the analog output.
We pose the analogy that the buffering and reclocking steps used to clean up digital are similar to the concept of suspension on a car. And the distortions in the waveform carrying the digital data are similar to roughness in the road’s surface. Some suspension systems are more effective than others, but with any suspension system, rougher roads feel rougher, despite the benefits of the suspension system. We contend that it is important to not molest the signal at any point on its playback journey.
This is a key way that Music Servers differ from Streamers. Streamers, using the car analogy, put all the emphasis on the suspension system and none on the smoothness of the road. Transfer over Ethernet using TCP/IP has no timing integrity approaching what is needed for high-end audio, so the money goes into the buffering in the Streamer. Whether this is better than the Music Server method or not is a matter for the ears, and heavily influenced by implementation skill.
Another misleading claim is that the modern DAC chips themselves are resilient to jitter. It isn't easy to generalise about all DAC chips, but simplifying things, DAC chips these days tend to up-sample to a very high rate and then convert to analog at this high rate. This is done in order to use a gentle filter on the output. This is not to be confused with DACs that have separate upsampling stages before the DAC chip, which reduce the upsampling work that has to be done by the DAC chip. DAC chips will perform with greater precision when they have to do less work, and pre-upsampling helps, but successive upsampling has its downsides - so it is a trade-off. The misleading claim that is often made is that this upsampling by the DAC, whether before the DAC chip or in the DAC chip, removes jitter. Technically, the claim is correct but it does not tell the whole truth. The upsampling step results in a low jitter output but the jitter on the input is mapped to broadband noise - effectively meaning some of the bits are changed, so you are trading away bit-perfect data for reduced jitter. Jitter in the input still impacts the sound but in a slightly different way. The immunity to jitter is a myth, and the only value of the upsampling is on the filter needed to get rid of extraneous digital noise in the analog output.
In Antipodes Music Servers the process is about minimising the influences that create jitter at every stage of the journey, AND buffer and reclock at each stage:
• Get a bit perfect read from the CD by ripping in paranoid mode.
• Eliminate any jitter at the first step by storing it on a hard disk for playback later.
• Store the files with zero compression (see below for an explanation why).
• Read from the hard drive, and transport the data with timing integrity all the way to the DAC, with obsessive emphasis on eliminating noise interference
with the waveform carrying the digital data.
• Buffer and reclock, using precise clocks fed with heavily regulated power supplies, at each stage of the journey.
There are many misunderstandings on this topic, but we can be concise. Lossless compressed files, such as ALAC and FLAC do not sound as good as AIFF and WAV files. The simple reason is that a compressed lossless file breaks down the music into frames and potentially compresses adjacent frames at different rates.
This means that compressed lossless files play at a rapidly changing bit-rate. This is a recipe for high levels of jitter and such files have to expanded into a buffer and clocked out of the buffer just to play. As explained above buffering and reclocking does not put 'humpty dumpty' back together again, and so there is more jitter in a stream played from a ALAC or FLAC file than from a AIFF or WAV file.
But beware of using WAV files. Apart from there not being a tagging standard for WAV files, the biggest problem is that there is no crc check (almost all file formats on your computer have a method to recover data automatically when it is lost due to small bit-level errors on your hard disks). This means your WAV files will acquire ticks and pops and eventually become unplayable. Therefore we rip to our version of FLAC, which is 100% compatible with the FLAC format, but which has no compression whatsoever. In blind testing the difference between uncompressed FLAC files and compressed FLAC files was reliably recognised, and no difference was recognised between WAV files and uncompressed FLAC files. You can convert your compressed FLAC files to uncompressed FLAC files using dBpoweramp on a PC or XLD on a Mac. Because FLAC is lossless, there is no damage done by having your files in a compressed form and then uncompressing them. They will sound the same as if you ripped to uncompressed FLAC initially.