Antipodes Audio High-End Audio Products
Design Approach

Design Approach



Antipodes Audio has been designing high end music servers since 2009 and launched its first music server for worldwide distribution in 2011.  In the period between 2009 and 2011 we focused exclusively on research and development, and many of the conclusions we drew in that period remain part of our design philosophy today.


The most central of these conclusions is that elimination of high frequency noise is the most important objective in a music server.  Other things are important but are comparatively trivial to achieve.  Elimination of high frequency noise is not at all trivial and presents the greatest challenge.

Reading from a computer drive is a much better place to start than reading from a CD.  The read is simply cleaner, faster and more accurate.  But the problem is that you are using a computer.  Standard computers generate a lot more noise than a CD player.  When that noise is high-frequency, then that noise will get through to mix with the digital signal sent to the DAC.  It typically results in sound that is mechanical and overly controlled, and sometimes aggressive and fatiguing.

Regardless of your reading or understanding of digital audio theory, it is easy to hear that reducing the high frequency pollution of the digital audio signal improves the naturalness and musicality of the audio coming out of any DAC.  There is a massive misconception that because you can move digital data perfectly between two points, that this is the same thing as sending a perfect digital audio signal.  It is not the same thing at all.  There is also the misconception that jitter is the culprit (variation in the clock rate).  Where you need a great clock, is in the DAC.  Good DACs do a good job of cleaning up the clock rate, but they all suffer reduced sound quality when high-frequency noise is increased.

There is no need for magic.  The magic is already in the recordings.  If we could just output a perfect square wave representing the data in the recording then our job would be done.  But noise that is at a frequency above the bitrate of the transmission creates problems for any DAC.

With analog audio, noise is separate from the music signal.  With digital, you don’t hear the noise separately, but you hear the distortion of timbre and space.  If you like, in digital the noise becomes part of the sound of the music, which is why eliminating it is the objective.


We observe that many of our competition appear to use electronically noisy components to complete the core functions and then adopt noise filters and add-on output cards to attempt to address this.  This is a bit like offering you filtered sewage to drink. 

Our approach is different because we design from the bottom up to minimise high frequency noise and to manage the frequency range of any residual noise.  This means no noise filtering and no add-ons.  In one sense, it is as simple as that, but that is not to imply that it is easy.

The three key things we do are:

  • Manufacture transformers that minimise parasitic capacitance between primary and secondary windings.  This is critical to stop high frequency noise entering the system via your mains power lines.  Our power supply designs are a big part of why our music servers sound more vibrant, natural and musically engaging.
  • Employ state of the art circuit design and layout to minimise high frequency noise generated and high frequency noise contamination between sections of the circuit.
  • Manage and tune global high frequency noise by the selection of all active parts and for some parts tuning the speeds at which they run. 

The tuning of global high frequency noise is not easy to explain.  It is not just about reducing the noise generated by individual active components, but is also about minimising overlapping frequency peaks.  If the frequency spectrum of noise of every active component had a peak at the same frequency as the others then the noise would add to create a very high noise floor.  If all the peaks do not overlap each other at all then the noise floor is dramatically lower.  Tuning the speeds of the chips and maintaining unconditional stability of the total system is not a trivial exercise.


The result is seen in better parts and a lower parts count. And it is heard in a more vibrant, natural and musically engaging presentation of the music.

Leave a Reply

antipodes audio


© 2019 Antipodes Audio Limited