Interview with Mark Jenkins, CEO of Antipodes Audio New Zealand
A detailed article about Antipodes Audio OLADRA technology from the new Antipodes music servers. It's a long piece, but definitely worth reading...
Antipodes Audio started in 2009 researching music server design. We discovered that simple circuits, high-quality components and a focus on reducing noise within the digital audio signal produced the best results. After spending 25 years working in companies developing digital audio transmission systems before founding Antipodes Audio, we had also learned a few tricks.
We launched our first Antipodes music servers in 2011 and then realized that it wasn't just the amount of noise that mattered, but also the frequencies at which the noise peaked. Around 2014 we developed advanced techniques to go beyond just noise reduction, but also to shape that noise. When designing motherboards, an important technique called 'spread spectrum' is used, which is used to reduce noise by spreading the peaks over a wider frequency band. This is used by all motherboard manufacturers to ensure they comply with RFI and EMI emission regulations. We discovered that we could use these techniques to completely tune a motherboard to achieve an accurate timbre, and this was our focus until the end of 2016.
It was October 2016 when I realized that focusing entirely on minimizing noise was a double-edged sword. It became clear to us that many techniques to reduce and manage noise could improve timbre, but at the same time would destroy the 'life' in the music. Real live music is explosive and our music servers didn't reproduce this quality. We were very pleased with the sound of our music servers when compared to competing music server products, but when I listened carefully to them compared to turntable systems costing €80,000, it was clear that the transients simply weren't coming through properly. While the best vinyl systems would achieve a pure bell-like clarity for every transient, that same transient on music servers would collapse into mush or a crunch.
You see, a music server is in many ways the ideal way to present music, except for one thing: you need a computer to do it and computers are inherently very noisy. So it was easy to become completely distracted by this pursuit of ever less noise. But this ignores the terrible twin of noise, bandwidth limitation. And bandwidth limitation is just as important. The problem is that many of the things music server manufacturers do to reduce noise impede bandwidth.
Our first step in this direction was to examine all the design decisions we had made in the past to reduce noise and remove the decisions that were limiting bandwidth. The result was very interesting. The music gained life and directness. With each of the tuning decisions we had previously implemented, we hadn't noticed this impact. It wasn't until we removed several of them at once that we realized what was going on. This led to the Antipodes CORE, EDGE and the DX3.
Our second step in this direction was to re-examine motherboard tuning, realizing we needed to make a trade-off between noise and bandwidth; and to design new linear power supplies with greater bandwidth (or in this context you could use the word 'speed' instead). This resulted in the Antipodes CX and EX, models that were widely praised for their musicality.
But we did more than that. In 2017 we started Project OLADRA, which was entirely focused on designing music server technology to jointly optimize noise reduction and bandwidth expansion, and to discover how the trade-offs between these goals can be best managed. A simple example: if you filter noise out of a signal, you limit bandwidth. So you need to design from the ground up to prevent noise generation in the first place, rather than trying to filter it out later. Since we started in 2009, we have not devoted so many resources to research and development.
By late 2019, almost three years later, this project had delivered on its first promise. The first commercially viable prototypes offered sound quality significantly better than anything we had previously heard from a music server. The greatest and most obvious gain was pure clarity and brightness. There seemed to be virtually no noise floor at all, along with excellent detail resolution and dynamics. Most importantly, the timbre and musical expression were excellent.
The new K-series and S-series are the first implementation of the OLADRA design and are designed using two key technologies:
Project OLADRA provided new insights into how noise generated by the motherboard can be addressed while maintaining bandwidth. The science suggested the possibilities, but it required careful prototype development, burning in the prototypes and listening, repeating this over and over, to arrive at the conclusions we use today. To explain it a bit further, we don't develop motherboards. Companies like Intel make chipsets and publish reference designs. Companies like ASUS develop circuit board layouts for those chipsets and reference designs, and make many component choices for the circuit. Antipodes selects motherboards that use the components we like and we tune from there, and reject those that don't meet our expectations. Almost all chips on a motherboard have parameters that can be adjusted, which can shift or spread the noise peaks generated by the chip. This can also be used at a global level to eliminate noise nodes between the noise spectra of multiple chips. It's possible to do this purely by measuring the resulting noise, but the real challenge is to listen and make good judgments about how to optimize sound quality as best as possible while choosing from a large number of possible permutations.
Project OLADRA also examined the optimal power supply for each different section of the music server. Antipodes was the first major music server manufacturer to use linear power supplies and many others have followed this example. It's an easy way to reduce noise interference and results in a smoother sound and tone. But it also limits bandwidth because even the fastest linear power supplies aren't fast enough for the job. The ideal power supply for a digital circuit would have the low-noise power of the quietest linear power supplies and the speed of the fastest switching power supplies. Some parts of a motherboard circuit sound better with a linear power supply and others sound better with a switching power supply. And of course it's much more complicated than that, since the topologies and components of power supplies, whether switched or linear, vary greatly. So we don't subscribe to the view that you should power as much of the music server as possible with linear power supplies. That approach delivers a smooth and easy sound, but misses the liveliness that makes real music exciting. Instead, Antipodes music servers use a complete power supply scheme that incorporates elements of linear and switched power supplies to get closer to the ideal power supply for digital circuits. The interesting result of this is that it's not a compromise to get more musical expression at the cost of increased noise. If you get this right, perceived noise is drastically reduced, clarity is increased and the life and emotional expression that are important to all types of music are retained.
The application of these technologies has resulted in new hardware building blocks, and this section gives you some insights into the different models as well as the differences with previous models. The main building blocks of the current models are the following new hardware components: V5.6H motherboard. This is physically the same board as the V4H used in the Antipodes CX, but audio performance has benefited significantly from new ISM tuning. The V5.6H is used in the K50 and K40 and is the main reason why the K50 and K40 are our best-sounding models. In terms of power, this board can upsample CD resolution music files and transcode to DSD512 using Roon DSP, using a single CPU core. It has an additional 5 real cores to handle other functions simultaneously with playback, making it very versatile and giving the main room to run the Server Apps excellently. V5.2H motherboard. This is a new board with the new ISM tuning that has the same single core performance as the V5.6H board, but with fewer cores. If the playback solution is kept simple, this board will give you some of the performance of the V5.6H, but it can't handle the range of simultaneous activities as easily and this will affect audio performance to a greater or lesser extent, depending on what you ask it to do. The V5.2H board is used in the K30 and S40.
V5X motherboard. This is a new board with the new ISM tuning that is slightly more powerful and sounds better than the V4X board used in the Antipodes EX. In terms of power, this board can upsample CD music files and transcode to DSD256 using Roon DSP, but requires Roon Parallelization to be enabled to use all its cores to achieve that. Like the V4X, the V5X performs well with server apps, but excels with player apps like Roon Ready, Squeezelite, MPD and HQPlayer NAA. R1i Reclocker Card. This is a new reclocker board that completely isolates the input electronics from the reclocker and output electronics. This means a very substantial improvement in sound quality over the Antipodes P2. The R1i board is used in the K50 and S20. HSL80 Power Supply. The HSL power supply scheme used in the HSL80 and HSL50 is responsible for much of the sound quality improvement over the CX and EX that preceded it. The HSL80 is heavily over-specified and is used in the K50, which uses three HSL80s, and in the K40 and K30 which each use a single HSL80. HSL50 Power Supply. This power supply has the same design as the HSL80, but is not as over-specified. The HSL50 should prove to have enough headroom for use in the S-series. The HSL50 is only used in the S60.
Server app as the Player app and can be used as Server app only or Player app only. The K40 is designed to run the Server app only. S30. The S30 uses the V5X board, offers USB and analog outputs and is powered by a standard smps power supply block. The existence of the analog outputs does not mean that an audiophile DAC is supplied. Analog outputs are provided for convenience, allowing a low entry point for a user who doesn't have an audiophile DAC. Sound quality increases significantly when a high-quality USB DAC is added. The S30 can be used to run Server Apps and/or Player Apps, but the hardware is optimized for the Player App function. S40. The S40 uses the V5.2H board, offers Direct Streaming via Ethernet and USB outputs, and is powered by a standard smps power supply block. The S40 can be used to run Server Apps and/or Player Apps, but the hardware is optimized for the Server App function. S20. The S20 uses the R1i Reclocker Board, which provides S/PDIF, AES3 and I2S outputs, and is powered by a standard smps power supply block. The S20 gets its input via USB. The S20 can be used to upgrade the sound quality of an S30, S40 or K30. S60. The S60 uses the HSL50 power supply, which can be used to dramatically improve the sound quality of the S20, S30 and S40. A single S60 can power two of the S20, S30 and S40. The ultimate solution from the S-series is an S60 powering an S40 running the server apps; the S40 feeds the S30 via Direct Stream Ethernet and runs the Player Apps; the S30 feeds the S20 via USB; and the S20 and S30 are powered by a second S60. K30 The K30 combines the V5.2H board running the server apps with the V5X board running the player apps, both powered by a single HSL80 power supply and USB audio output. It's similar to combining the S40, S30 and S60 in one chassis, but with superior interconnection and powered by the larger HSL80 power supply.
K40. The K40 uses the V5.6H board, powered by a single HSL80 power supply, to provide the ultimate device for running Server Apps. The only output is Direct Stream Ethernet, which can send an incredibly clean signal directly to the Ethernet input on any DAC. K50. The K50 uses the V5.6H board to run Server Apps, the V5X board to run Player Apps and the R1i Reclocker board. Each of the three cards is powered by a dedicated HSL80 power supply. Outputs are Direct Stream Ethernet, USB, S/PDIF, AES3 and I2S. K10. The K10 is a highly precise ripper in a housing cut from two pieces of solid alloy, which can be attached to any Antipodes music server and can be used to automatically rip your CDs. It connects to your Antipodes music server via a dual USB cable. Signal Processing To optimize the output signal from the Player app, only one Player app can be used at a time and only one signal pipeline can be used at a time. Therefore, the user selects a player app, such as Roon Ready, Squeezelite, MPD or HQPlayer NAA. Similarly, the user selects one of the analog outputs, a single USB output or the reclocked digital outputs. Such a restriction is not necessary for Server Apps. All installed Server Apps can be used simultaneously for streaming over the network, such as Roon Server, Squeeze Server, DLNA/UPnP, SONOS Server, Plex Server etc. General Design Focus The 'terrible twins' of noise and bandwidth are mentioned above, and we have tried to explain below why these are important in designing a music server. To do so we have omitted many details that are important in digital audio, but are not necessary to make the point, in the hope that we can make this understandable to as many readers as possible. This is not intended as a technical white paper or academic document, so excuse the simplifications made. What follows not only addresses the effects of noise and bandwidth, but also makes clear that you should not assume that higher resolution files (bit rates) will necessarily sound better. From a sound quality perspective, the goal of a music server is very simple: send a perfect square wave representation of the music file to the DAC. Unfortunately, the difficulty in achieving this goal is underestimated by many and can never be perfectly achieved in practice. The concept of jitter misleads people into thinking that all you need in a digital signal are the correct bits (which is relatively trivial to send) with great timing (low jitter), and thus all you need is a great clock. This simplistic view is very misleading. At least three things are important: the clock, noise and bandwidth. The difference between a great clock and an ordinary one can be erased if noise and bandwidth are problematic. Using a great clock is good marketing, but is not always very relevant in general design considerations. This article describes as simply as possible how noise and bandwidth limitations cause jitter, regardless of the clock used.
Noise In the image below of a perfect square wave, the horizontal axis is time and the vertical axis is voltage. We assume that the clock is perfect - that is. the vertical signal lines occur at perfectly spaced intervals (the bit rate). When the signal represents a binary 0, it is at 0v. When the signal represents a binary 1, it is at 1v. And we assume that the receiver of this signal decides that the transition between a 0 and a 1 has occurred when the signal rises through the 0.5v level, and that a 1 has transitioned to a 0 when the signal falls through the 0.5v level. The image below shows exactly the unattainable goal of all music server manufacturers, and it really is as simple as this image. Now imagine that noise is added to the signal. If the frequency of the noise is below the bit rate, this perfect square wave floats on top of a longer and smoother wave. The interesting point to note is that the timing between the data transitions (where those vertical lines pass through 0.5v) is unchanged. So no problem yet. If the frequency of the noise is above the bit rate, the horizontal lines become fuzzy. And if we combine the low-frequency noise with the high-frequency noise, the effect is combined. Again, the interesting point to note is that the timing between the data transitions (where those vertical lines pass through 0.5v) is unchanged, provided the noise is not extremely high. So again, no problem. Noise by itself (as long as the deviations caused are materially less than 0.5v) is not a problem. The reason it's not a problem are those vertical lines, because noise doesn't change the space between them. Bandwidth Now imagine that there is no noise. No noise is impossible, but something else that is impossible is the vertical line on the square wave, because it requires infinite bandwidth. The vertical lines indicate that the signal can reach 0v and 1v at more or less the same moment. Whatever tools we have to send a signal, the demands of high bit-rate signals go far beyond what the available tools can provide. Think about how analog cables can mess with sound up to about 20 kHz, and then think about the enormously larger frequency range required for a digital cable (and optical cables just have a different set of problems, mostly related to reflections). The higher the bit rate, the more difficult it becomes. When we account for limited bandwidth, instead of transitions happening immediately, the signal goes up a slope at the transition from 0v to 1v, and goes down a slope at the transition from 1v to 0v. To illustrate this, if the bandwidth were the same as the bit rate, the signal would be a sine wave. To make the signal reasonably square, you need to add various harmonics of the bit rate (say 7 or more) above the bit rate, and that's a lot of bandwidth - even more for higher bit-rate signals. The image below is how a sine wave starts to square when you add one harmonic. Interestingly, in both examples of limited bandwidth, the transitions through 0.5v are still perfectly spaced apart - even with the sine wave. So still no problem. As mentioned earlier, a higher bit-rate signal (for those of you who think high bit-rate files must always sound better) requires even more bandwidth to control the wave, and thus in a system with a finite bandwidth limit, a lower bit-rate signal is represented more accurately than a higher bit-rate signal. Food for thought? Furthermore, if you ask anything in a music server to work faster, it will work with less precision and this is an important trade-off to watch for when assuming that higher bit rates must be better, just because the numbers are bigger.
Noise and Bandwidth So why did I say that noise and bandwidth are important? The attentive reader will realize that from the examples above we can only conclude that there is no problem if we can achieve zero noise or infinite bandwidth. But each of those goals is unattainable and the problem becomes clear when there is both noise and limited bandwidth. In the image below I've added a low-frequency noise component to a bandwidth-limited digital audio signal. I hope you can now see that the 0.5v points are shifted right or left by the addition of low-frequency noise that lifts or drops the signal between bits. By shifting the slopes up or down, the 0.5v points are shifted left or right. The greater the amplitude of the noise and the greater the bandwidth limitation, the greater the effect on timing (jitter). In the image below I've added high-frequency noise to a bandwidth-limited signal. I hope you can see that the transition time point at exactly 0.5v is now difficult to distinguish for a digital receiver. If the signal is vertical at the transition, noise has no effect on it. But as soon as the transition is not vertical, noise changes the transition point. Conclusions So this is the point. Combining limited bandwidth and noise inevitably leads to jitter (variation in data transition timing), regardless of how great the clock is. What we often see nowadays are music servers designed using very simple computer components, which produce a lot of noise, with filters added to reduce that noise. And we see music servers powered in the wrong places by slow linear power supplies. However, there seems to be very little understanding that filters reduce bandwidth, which means that reducing noise doesn't improve timing, and that very fast power supplies at certain points in the circuit are essential. The goal of Antipodes' OLADRA project is to design music servers from the ground up to have both low noise and high bandwidth. Because zero noise and infinite bandwidth are unattainable, there is also a need to find the best trade-offs between noise and bandwidth in different parts of the circuit, which leans more toward the art of objective listening than science. By that I mean that many of the trade-offs made in the OLADRA project were informed by listening rather than by theory or electronic measurements. This results in our ears in music that is not only significantly more transparent, but that is definitely more musically expressive, as if more of what's important in music comes through.
Ohm-Audio BV 2021-09
Source: Mono & Stereo
