Interview: John Kenny of Ciunas Audio

John Kenny, based out of Dublin, Ireland, has been producing audiophile grade computer audio products that are affordable, feature proprietary power schemes, and sold direct to consumer. In this interview, we discuss many topics including his design philosophy, his new line of ISO products, and USB audio.

1) Can you tell us a bit about your background and how Ciúnas Audio came to be?

At an early age I was building crystal radio sets, taking transistor radios apart and building/improving amplifiers, constructing new types of kites, model rockets and model hovercraft. Later, at University, I completed a science degree but at the same time, computers captivated my interest and I eventually set up a computer company specializing in bespoke software. Returning to my electronics hobby in later life I found a trove of information on the web regarding building & designing audio electronic devices and forums dedicated to DIY which I avidly digested.

My growing interest in the impact of power supply noise in DACs led me to an investigation into the use of a relatively new battery chemistry. I premised that this battery had characteristics which were important for powering digital audio devices. Apart from ultra low noise, they offered huge current output (120Amps instantaneous) which I believed would make them a very stable PS. I considered the latter hugely important, particularly in the sensitive circuitry of digital audio devices, even though they represent a light load albeit at high frequencies. Another appealing aspect was that their 3.3V voltage output held up during their discharge cycle. Normally batteries drop in voltage as they discharge but these batteries remain close to their 3.3V output until near their fully discharged state.

In 2009, I tested the batteries by modifying two popular digital audio devices at the time; Musiland 01US DAC and the original Hiface USB/SPDIF converter. The results exceeded my expectations and I posted details of the modifications on DIYAudio. A number of people took up the challenge and reported great success. However, not everybody wanted to do the modifications themselves and I was inundated with private messages requesting that I make the modifications for them — a business was born.

At that time my products were referred to as JKDAC or JKSPDIF and JKDAC32 but I eventually settled on a name which I felt encompassed what I was about — Ciúnas Audio. Ciúnas is the Gaelic word for calm, stillness, serenity and, as my design focus is on noise reduction and stability, I felt it represented this as well as introducing some Gaelic words into the audio world.

2) What about John Kenny designs, philosophy, business model, and customer relationships make it unique in the marketplace?

Early on, I decided to specialize in computer audio and USB input in my devices as I felt that it was capable of truly excellent sound and it was becoming a serious platform for audiophiles. I reasoned that designing a single input excellently (USB) was much better than designing a number of inputs adequately such as SPDIF, Toslink, I2S, HDMI. USB was often an afterthought input on many audio devices, at the time.

I found that with low noise, highly stable battery power to crucial areas of digital audio devices, (like USB receiver chips, low jitter clocks, D/A converter chips), great improvements in sound could be achieved. The reasons for these improvements only became clear when I found answers in the workings of auditory perception and ASA.

 Auditory Scene Analysis (ASA). research, instigated by Al Bregman, arose from trying to understand the well-known “cocktail party effect” which is our ability to focus on a particular conversation in a noisy room full of other conversations. This is achieved by the brain’s processing of the auditory signals from the ear, not by the ear, itself.

The significance of this research to me is this — audio playback is attempting to achieve an audio illusion — an illusion that the listener can believe is ‘natural’ sounding. The more natural sounding, the more believable and engaging is the playback. But ‘natural sounding’ isn’t just ‘preference’, it’s explainable by ASA’s research findings. From infancy, we absorb how sound behaves in the world & this teaches our fledgling auditory perception as to what is natural. ASA is the study of exactly how auditory perception creates an auditory scene from the nerve signals on the auditory nerve — much the same as visual perception creates a visual scene from the signals arriving from the optic nerve.

My clarity about how power supply stability affects auditory perception is explained by ASA. Low levels of power fluctuation are ground noise fluctuations which ultimately affect sensitive circuitry in DACs, for instance. (We even see new ways of hacking what is being typed on computers by sensing ground leakage fluctuations on power outlets powering the computer).  Now these noise fluctuations on the DAC are not random, like the hiss we hear from analogue tape–they are correlated with signal processing. Random noise is categorized by ASA as background & easily ignored – it doesn’t intrude into our perception of the foreground sound. Fluctuating noise is analyzed differently – it has pattern characteristics (temporal, amplitude, spectral) which ASA tries to make sense of – group into any of the auditory objects it has already identified. If it can’t make sense of this signal, it is confusing to our perception & leads to a perception of a less clear auditory scene, less believable, less natural.

When this source of fluctuating noise is removed by using stable power from batteries, the whole auditory scene becomes clear, more realistic, more natural – the scene falls into place & makes sense. It’s a system-wide effect, not a change in certain frequencies or amplitude — all of which makes it difficult for some to understand.

Just to be clear – the concept of fluctuation of the PS during processing is well established in engineering and it is now one of the well established techniques for hacking called side channel analysis where unintended side channels give an indication of the signal processing activity and can be used to hack it . In Simple Power Analysis (SPA) key bits are seen directly in the power consumption of a chip using it in a multiplication operation.

There’s one important point to understand about the whole area of auditory perception and it is exemplified in this phrase, borrowed from linguistics, “poverty of the stimulus”. The meaning here is that there isn’t enough data in the signals coming from the eardrum to create an auditory scene we are sure of. Our perception has to use all sorts of other information to try to resolve this dilemma — the dilemma of achieving, in real time, a stable auditory scene which is trustworthy and useful for our interaction with the physical world. To resolve this we have to draw on other data, data from sight being a primary one but also our prior knowledge and experience of how sound objects behave in the world — for instance a lot of natural sounds have a fast attack and slow decay — the reverse, (slow attack and fast decay) is unnatural sounding. We know that a sound which has a slow decay & then abruptly stops that it is being damped. Through experience we have built a library of such ‘sound pattern information’ which we draw on to make inferences and predictions about the sounds we are perceiving.

How does ASA inform my designs? I realised that what I was focussed on in my designs was having an audible effect on the whole presentation, not just better bass or better treble. Other aspects that listeners commented on were a perception of increases in depth of soundstage, solidity of the individual auditory objects in the scene, naturalness — in essence a more realistic auditory scene. My current working premise is that, below a certain low noise floor, what is crucial is the stability of this noise. If this noise is fluctuating, it affects how we perceive sound but isn’t heard as an entity in itself. An exaggerated example of this is pre-echo in some low quality MP3s which blur transients but the pre-echoes themselves are not heard, just their effects (here’s a Youtube video of this effect)

ASA research, even though started in 1990, is still at a relatively early stage of investigation — 27 years is a short time in perceptual research. This makes for a difficult task in uncovering the exact ASA mechanisms in operation during auditory perception of complex signals such as music. ASA is still at the stage of using relatively simple audio signals for testing and is only now beginning to consider more complex signals. Until there are advances in this aspect, all we can rely on is intuition, experiments and knowledge of ASA to guide us.

My philosophy …

Doesn’t just apply in the scientific arena but in all of life. Remaining open-minded and inquisitive is a struggle but one worth trying to pursue.

Having established a relationship with those people who engaged me to modify their Hiface products, I have stayed loyal to this type of customer relationship — a personal interest in ensuring that they are getting the most from any product of mine. That includes giving advice about the surrounding equipment being used with my audio devices. In the past this has uncovered ground loop issues & sub-optimal playback settings in customer’s systems.

My direct sales business model allows me to stay in touch with my customer base and also ensure the best value for money that my products have been renowned for.

3) Can you tell us about how you test finished designs, and about your reference system, along with your musical preferences?

Throughout my product design I measure and test components, modules & sub-assemblies. All finished products are tested for their specific functionality and battery functionality — charging is tested. Finally, listening tests for all products, are performed before shipping.

I found that with low noise, highly stable battery power to crucial areas of digital audio devices, (like USB receiver chips, low jitter clocks, D/A converter chips), great improvements in sound could be achieved but this wasn’t showing up in my measurements. For instance, early on while I was experimenting on the ESS DAC chip I found it sounded much better when run in synchronous clock mode rather than asynchronous mode where it used it’s inbuilt & well regarded ASRC (Asynchronous Sample Rate Converter). But nothing I could measure showed why this sounded better. And although I use an analogue 100Mhz scope & digital signal analyzer to aid in testing and fault-finding individual components, I remain committed to evaluating my designs using audition — by me, and by a local group of audiophiles in Ireland.

Just to say something about measurements here. I find noise which fluctuates because dynamic music signals are being processed is particularly difficult to measure. I haven’t yet discovered, or seen anybody describe, a technique for measuring this  dynamically fluctuating noise.

For testing of DACs, I use headphones directly from the RCA outputs which gives me the best forensic insight into the sound. At other times, I use a Naska amplifier and some amplifiers of my own design outputting to Jordan JX92S fullrange speakers in speaker cabinet of DIY design — a 3D spiral horn design. My prototypes are also trialled in the local audiophile community who have a variety of audio equipment. Quad ESL 57s & 63s, Raidho & many other speakers; AirTight & a variety of solid state amplifiers and very heavily modified computers as source. DACs such as Lampizator, Meitner, Eximus and many others are used in auditions. We meet up via a local audio forum Tirnahifi

My music listening is varied, Swing Jazz, rock, some classical and some folk & country. Any music which I can connect to emotionally is my preference and I find good music systems make this far easier to achieve.

4) Can you explain, your view, how USB audio has evolved, issues that have been brought to light, and what the current state of the USB protocol for audio is?

USB audio came of age when Gordon Rankin introduced asynchronous USB protocol where the clock in the receiving audio device is the master clock timing the USB packets rather than the lower quality clock in the PC. This advance took USB audio out of the realm of an audio toy and into consideration for serious listening. It opened up computer audio to many and certainly began to appeal to audiophiles, particularly with the convenience offered by a music database, tagging and search functionality available.

There was great progress in USB audio and many subscribed to the belief that jitter was now under control with asynchronous USB and this meant the audio quality must be optimal. But many reported that their CD playback sounded more dynamic, alive and interesting — computer audio wasn’t matching this for some reason.

My path down the road of ultra low and stable power supply noise opened my mind to the psychoacoustic repercussions of modulating noise. I had already premised that ultra low noise is one thing but stability of this noise is the more important characteristic. We can tolerate and even become unaware of noise which is stationary, not fluctuating — this is easily explained in ASA. Noise which modulates, irregularly, is much more intrusive in our auditory perception and captures our attention because it is not being treated as background.

The linked Youtube MP3 demonstration gives an example of this — what is happening here is that quantization noise is generated in MP3, but we cannot hear this fluctuating noise itself, we can only perceive it by its effect on the perceived blurring of the transients. I believe a more subtle form of this, low level noise, is at the heart of why digital audio can sound uninteresting, lacking in perceived dynamics, unmusical, unnatural. One major problem, in all this, is that measuring this low level fluctuating noise is a very difficult task & has not been fully resolved yet — glimpses are seen, however.

My audio devices, running directly from battery, do not suffer from this issue of delivering uninteresting audio playback, lacking in dynamics — it’s always been one of the main characteristics that users mention when they first hear my devices. Computer audio now matched and often surpassed even the best CD playback. From my discussion with a few customers in which we resolved why my audio devices weren’t providing the sound I know they can, I was reinforced in my conclusion that noise intrusion can deaden the sound, make it uninteresting to listen to, rob it of life. My direct battery power is addressing the main cause of fluctuating noise but it wasn’t until I heard the Intona USB isolator with my Ciúnas DAC that I could appreciate the audible effect of removing another source of noise coming from the USB connection itself.

I had already experimented with this, trying to block common mode (CM) noise using CM chokes on the USB signal lines — no effect; using ferrite rings on USB cable — dulling of sound; using a USB cable ferrous sheath covering — a significant reduction in hiss but some drop off in sparkle (this one will be revisited when time allows). But the USB isolation of the Intona affected the sound in a different way — it solidified the soundstage, made the interplay of the musicians more understandable, making the whole presentation more believable and musical. I knew this was a different type of improvement in sound and decided to investigate further and find a way to use USB isolation with and in my existing Ciúnas audio devices.

Again,in his case my premise is that the USB receiver chip is generating some self-noise within the chip & ultimately on the ground plane connected to this chip. This may be due to USB processing itself or, more likely, some overshoot/ringing on the very fast USB signal waveform (USB high speed 500pS is the rise time spec). I believe this is where the digital design requirements & analogue design requirements clash. A Reasonable amount of noise at the digital receiver does not cause bits to be misread but sensitive analogue circuitry like clocks are affected by ground plane noise. We can see a USB signal overshoot in the following measurement as the spikes at the top of the rising edge of USB waveforms in the plot (even though this is a plot of the slower full-speed USB & may be due to the scope probe used although I would expect TI know this – click on pic for link to TI doc)

USB isolation of the USB signal lines removes any fluctuating noise, usually called common mode noise (CM noise), which is arriving at the USB receiver. Using a USB hub to regenerate the USB signals after this isolator is required to remove the jitter that all isolators add to the signal.

It took some time in testing, listening and ironing out the bugs but I eventually achieved success. My testing also revealed that USB reformatting/reclocking is needed after all isolation devices as they introduce jitter into the signal. My Battery powered ISO Hub was better sounding than the Intona when the group auditioned it side by side.

I recently released a number of products incorporating this technology:

  • ISO-HUB — a new device which offers 1 USB input and 4 isolated/reclocked high speed USB 2.0 outputs. Use before any USB audio device to obtain audible benefits of more sonic realism.
  • ISO-SPDIF — A world first — an isolated USB SPDIF converter. The Ciúnas USB SPDIF converter is integrated with the ISO-HUB technology providing a USB signal with optimal signal integrity to the USB/SPDIF converter.
  • ISO-DAC — Another world first — isolated USB DAC. The Ciúnas DAC again with integrated ISO-HUB providing an isolated and reclocked USB signal, internally, to the USB DAC.And lastly a product that many have been asking me to release, so they could avail of this battery power for non Ciúnas devices — a self-contained isolated battery supply which can be used to power many audio devices, not just my Ciúnas devices.
  • ISO-PS — an external battery power supply giving isolated, selectable voltages of 3.V and 6.V outputs and other voltages available on request

5) What do you for enjoyment when not designing digital audio devices and running your business?

I’m in the enviable position that one of my passions is actually my audio business. I believe most audio designers are similar and do this because of their passion and it’s no different in my case.

When I am forced to unplug my soldering iron and scope, I relax with painting. It’s similar to audio design in the way that one can get lost in doing a painting but it comes from a totally different part of the brain — the creative, intuitive part and yet the joy in a finished painting is similar to the joy in testing an intuitive design change & hearing the results.

Some paintings of mine…!