Marcin Ostapowicz - JPLAY

I first came across the name back when JPLAY was a stripped-down Windows program making waves among people willing to spend an entire weekend optimizing a computer and DAC, disabling system services, testing drivers and checking whether the latest change had made the sound even slightly better. I knew enthusiasts like that and understood that many of them considered JPLAY a real discovery, but I watched from a safe distance. I only returned to the subject recently, by which time JPLAY had become an iOS app. First one friend told me I had to try it, then a second, then a third... These were not casual listeners, but experienced enthusiasts and people from the industry, including distributors, consultants and designers of amplifiers and speakers. As a longtime Roon user, I was wary of their enthusiasm. But after hearing the same recommendation again, I started to think there might be something to it. Eventually, I installed JPLAY and understood almost immediately what all the fuss was about. It was not just the sound. I was impressed by the app’s appearance and the way it worked. It is fast, lightweight and polished, starts without requiring an additional server, and has surprisingly extensive capabilities beneath its simple interface. I have written a separate article about it, so I will not repeat myself here. What I did start wondering was who was behind JPLAY and how its creator had gone on to design computer cards, power supplies, servers, switches, LAN cables and other devices that make the minds of audiophile-voodoo hunters start to boil.

That person is Marcin Ostapowicz, a computer scientist, music lover and audiophile who has devoted much of his professional life to one seemingly narrow question: how far can the quality of a digital audio source be improved? He began by experimenting with computer audio and software, which led to JPLAY. Later, he concluded that software alone was not enough, so in 2013 he founded JCAT, a brand offering specialist USB and network cards, linear power supplies, clocks and accessories for building high-end computer audio sources. The next step was XACT: no longer a set of tools for improving a PC, but a brand of complete high-end components, including the S1 and S1 EVO music servers, the N1 switch, Phantom cables and Immotus anti-vibration feet. All of these projects grow out of the same philosophy: the belief that digital audio is not just about whether the correct sequence of ones and zeros reaches the end of the chain, but also about the entire process that gets it there - the power supply, clocks, interference, network, operating system and software itself. You can agree with that, view some of these ideas with skepticism or decide that a switch costing several thousand dollars is pure madness. But it is hard to miss the fact that Ostapowicz has followed the same path consistently for years, trying each time to push the boundaries a little further. That is why I wanted to ask him not only about JPLAY, JCAT and XACT, but above all about where this obsession came from, where it has taken him and whether there really is still room for improvement in digital audio.

Your father was a professional musician and teacher, and you attended music school through the first level, studying guitar and piano. But you eventually left music education behind and turned to computer science. How did that happen?

My father was very keen for all of us to have at least a basic musical education. I have two sisters, and they both graduated from music school as well. One followed in our father’s footsteps, while I was more interested in computers, so I went into computer science. Looking back, though, I am very grateful to my father for giving me that musical education. I think I would be doing something else today without it.

When did you first recognize a computer’s potential as a music source? Do you remember the moment when you realized that music from files could sound better, or at least different, from traditional sources?

I remember when I first came up with the idea of getting into this subject. It was when I was in high school. I already had internet access and spent a lot of time on a forum dedicated to PC tweaking - overclocking, modifications, water cooling and so on. Somewhere, I read about unofficial drivers for the popular Creative Labs sound cards of the time and started experimenting. I have kept experimenting ever since. Most of my products grew out of those experiments.

Back then, it was not easy to get people in audiophile circles interested in file-based playback, was it? I remember how much criticism I got for it. After reviewing my first DACs and streamers, I would get comments saying that by listening to "MP3s," I was choosing convenience at the expense of sound quality...

At the beginning of my journey, I had not yet had any contact with the strictly audiophile community. It was more a matter of PC audio and a forum where users were interested in tweaking everything. This was around 2000. Not long afterward, I had my first USB DAC, based on a DIY kit built around the Burr-Brown PCM2702 chip. That was how it started. The real revolution came a few years later with the first USB-to-S/PDIF converter to use asynchronous data transfer and support 24-bit audio at sampling rates up to 192 kHz. That product, from M2Tech, sparked a revolution in the hi-fi market. Anyone with a DAC featuring an S/PDIF input and a computer could experience hi-res files. That moment coincided with the launch of the first version of JPLAY for Windows. Despite its Spartan interface, the program offered such good sound quality that it quickly became a market success.

While experimenting with computer-based music playback, you discovered the alternative kX Project drivers for SoundBlaster cards and delved into foobar2000. Is it true that you asked so many questions on the foobar forum that you nearly got kicked off?

That is true. There are places where certain questions are not welcome and certain ideas are not welcome either. The foobar forum was one of them.

Was that how the idea for your own music player came about, eventually leading to the first version of JPLAY?

Essentially, yes. I did not do it entirely on my own, but the people who helped me do not want to be part of the public story. Let’s put it this way: I met another user on one of the online forums, and we worked on developing the program together for years.

What was your original goal with JPLAY, and what did the first steps in its development look like? The early versions did not look particularly impressive...

Sound quality was always the top priority. That is why the Windows program looks so Spartan. In the current version, you could say it does not look at all, because it runs in the background as a system service. You need an external UPnP app to control it, such as JPLAY for iOS.

What was the reaction from users and the audio press? Did listeners notice an improvement in sound quality straight away? There were some enthusiastic reviews. The Absolute Sound praised JPLAY for delivering "the highest sound quality ever achieved from a digital source," while 6moons wrote that the app brought digital playback close to vinyl. How did that recognition affect you and JPLAY’s further development?

At first, we had a small group of highly engaged customers who also served as beta testers for successive versions of the program. That group grew quickly. People from the industry, manufacturers and reviewers began to discover the program. Very positive reviews and awards soon followed. I am proud of that success, and it only motivated me to keep working.

Many people believe that if a software player provides bit-perfect playback, all such apps should sound identical. From the beginning, though, you have argued that you can hear differences and that JPLAY can improve the sound of music files. What is the secret? What makes one music player sound better than another, even when both deliver the same data?

I noticed differences in the sound of different digital transports very early on. Bit-perfect and good sound are not the same thing. Many factors affect the sound of a computer-based transport, both in the software and in the hardware it runs on. The secret to JPLAY is that we optimized the code as much as possible and tested every change by listening. There are many ways to optimize digital audio playback, and those methods differ between operating systems. It is particularly difficult with commercial operating systems such as Windows, because they are heavy, full of bloat and were not designed for this kind of application. That is why manufacturers offering computer-based music servers running Windows strip out everything they can, leaving only the processes and services needed for audio playback. I did that myself, too. But this approach has limits that cannot be overcome, so at a certain point it became clear to me that if I wanted to go further, I would have to build a new system around a lightweight version of Linux with a real-time kernel, running on my own hardware platform.

JPLAY did not stand still, of course. New versions and features continued to appear. In 2018, JPLAY Femto was released, adding, among other things, the Femto UPnP server for streaming. Which changes in JPLAY’s history do you consider the most important?

That was the most important change in JPLAY’s history on Windows. A computer running JPLAY Femto became a dedicated audio server/streamer that could be controlled remotely through an app on a tablet or smartphone.

After years of success on Windows, you decided to make a mark in the world of mobile devices as well. In 2022, the JPLAY app for iOS made its debut. What led you to create it? What features does the app offer, and what problems is it designed to solve for music lovers?

Creating the JPLAY app for iOS was the best decision I have ever made. There are several reasons for that. First, I was already working on my own audio streamer/server, XACT, and I knew I would need an app. Customers using JPLAY Femto on Windows also needed a good app, and what was available on the market was not impressive. What is more, there are hundreds of thousands of devices on the market that support UPnP AV, so I saw a lot of potential there. I had plenty of experience, knowledge and industry contacts, and now I know it was a great decision.

That raises an obvious question: why iOS only? Many Android enthusiasts feel left out. Can you explain why the app was developed exclusively for Apple devices? Is there any chance of an Android version in the future?

JPLAY is a native iOS app. That is why it is so fast, smooth and stable. We specialize exclusively in iOS. We could have made a cross-platform app using some framework, but it would not work half as well as it does now. An Android version would basically mean building the app from scratch, so unfortunately, there is no chance of that. The Apple market is large enough that I do not see this as a problem. We have customers who use Android in their everyday lives but buy a small iPad just for JPLAY. That says something.

In an interview several years ago, you said you had long been skeptical of streaming, believing that streaming services could not match the quality of local files. Recently, though, you have softened your position and said that streaming from TIDAL or Qobuz now sounds good enough to compete with local files without having to make excuses for it. What has changed?

It is not that I have stopped listening to local files. I still listen to them, but streaming now sounds good enough to enjoy without any reservations. Streamers, software, network infrastructure, cables and power supplies - when all of these elements are properly optimized, they let you enjoy high-end sound whether the file comes from a hard drive or the cloud.

The market for streaming players and music systems is highly competitive. There is Roon, Audirvana and HQPlayer, for example. Users can use their streamer manufacturer’s app for free, or even play music directly from their favorite streaming service’s app, thanks to features such as TIDAL Connect and Qobuz Connect. So let me put it bluntly: why on earth would anyone need something like JPLAY as well?

Because JPLAY combines the advantages of all those solutions and can play music from both streaming services and local files. And it does it better than the competition.

JPLAY is a Polish product that has earned recognition in the global audiophile market. There are plenty of opinions from experts, but I am interested in how much real interest there is among audiophiles. Can you put that in numbers?

Yes. JPLAY has tens of thousands of customers around the world.

Do you plan to expand cooperation with other audio companies, for example by increasing the number of devices compatible with JPLAY?

Of course. We are working on that all the time, and the list of JPLAY Certified manufacturers keeps growing.

In 2013, you took things a step further and launched JCAT, a brand focused on hardware accessories for computer audio. What led you into the hardware segment? Did you feel that software alone could no longer improve the sound and that the physical components in a computer audio chain also needed attention?

As with JPLAY, the motivation has always been to improve sound quality. I experimented with hardware alongside software. I already knew where the weak points were and what needed to be improved. I started with accessories, and a few years ago I brought all that knowledge together in the form of the XACT S1 music server.

JCAT’s first product was its own USB card for PCs, followed by the world’s first audiophile network card. Over time, the range expanded to include specialist cables, USB/LAN isolators and linear power supplies for computers. How would you explain in simple terms why audiophiles might need accessories like these?

They are good products for anyone looking for a way to improve the sound of a system in which a computer serves as the source or transport. Each product reduces noise, resulting in a smoother, more analog sound.

Why is clean, stable power so important to the sound of a computer-based audio system? What sonic benefits come from replacing a typical "computer" switching power supply with a high-quality linear one?

Power is important in every audio component, not just computer-based ones. Most manufacturers use switching power supplies because they are cheap and simple, and they solve the problem of heat generation. A good linear power supply is expensive and makes a product more complicated. It also brings thermal challenges, but in my view, it is essential for achieving high-end sound - the kind that offers good resolution along with body, natural tonality and refinement. Switching power supplies are responsible for the dry, flat and aggressive sound that experienced audiophiles rightly criticized in computer audio.

JCAT has quickly gained customers around the world and earned recognition in the hi-fi industry. It is one of the few Polish audio companies respected even in Japan. I have heard that local reviewers use your cards and power supplies. How did you manage to establish a Polish company in such a demanding market?

It is very difficult to succeed in the Japanese market, which, as far as I can tell, does not pay much attention to international publications. One of my customers, a Japanese man, helped me.

In 2023, you founded XACT. Why? How is XACT different from JCAT? Is it simply an even more advanced, high-end product line, or a completely new direction for the business?

XACT embodies the knowledge, insights, discoveries and experience I have gained during my journey through audio, as well as through creating JPLAY, my audiophile software, and JCAT, my innovative high-end audio hardware. It is the very best I have to offer. It is a new brand focused on fully fledged high-end components.

XACT’s first products include the reference S1 music server, high-end Phantom cables and Immotus anti-vibration feet. What makes a music server like this special compared with more "ordinary" alternatives?

In short, the S1 is not a computer adapted for playing music. It is a device designed from the ground up for that one purpose. That is a fundamental difference. Most music servers and streamers on the market, including some very expensive ones, are based on off-the-shelf PC motherboards designed for office work or home entertainment. Those boards contain switching converters that pollute the power supply and the rest of the system. That is what you hear as a characteristic "digital" edge. Some manufacturers try to fix the problem with a better power supply, but that does not remove the source of the problem, which is on the motherboard itself. That is why I took an approach no one had chosen before with the S1: we developed our own motherboard from scratch, dedicated exclusively to music playback. It is the world’s first motherboard powered entirely by linear regulation. It uses only linear regulators, and its Optimo X1 power supply, with 140,000 µF of filtering capacitance, contains no switching components. It also has an OCXO clock with ±5 ppb stability, resistant to temperature changes, vibration and electromagnetic interference, along with components selected by listening, from Nichicon Muse capacitors to a specially selected SSD with a proprietary XACT Link cable and an industrial SLC card running the operating system. The other half of the equation is the software. After more than a decade of developing JPLAY, I know that software can affect sound just as much as the hardware platform. XACT OS was optimized for this specific hardware. I listened to every change in the system, however small, and logged it. The whole project took six years. Work began in 2016. It took three years to finalize the motherboard, followed by another three devoted entirely to tuning the sound. It is the device I have worked on longer than anything else in my life. The result is that the S1 does not sound "digital." It sounds natural and organic, with an analog sense of ease and presence. At the same time, it is incredibly simple to use, with a single tap in the JPLAY app. That is the difference from "ordinary" alternatives: elsewhere, you are buying a computer in a nice case. Here, you are buying a device built for one purpose, with no compromises.

A few years ago, audiophile network switches began appearing on the market, including the Silent Angel Bonn N8 at $599, the English Electric 8Switch at $800 and the NuPrime Omnia SW-8 at $579. Prices in the $500-to-$800 range came as a bit of a shock, but before long we also saw products such as the Silent Angel N16 LPS at $1,999, the Synergistic Research UEF at $2,295 and the Nordost QNET at $3,199.99. You then decided to outdo them all with the XACT N1 switch, priced at $7,450 in the U.S. market. You have to admit that it sounds like a joke not only to newcomers but even to many audiophiles...

I can see why it sounds that way - until you look inside. If the N1 were like most "audiophile" switches, that price really would be a joke. By the way, I have not actually "outdone them all." There are network switches on the market that cost several times more than the N1. But I do not compete on price at all, because for me, price is a consequence, not a strategy. Almost all products of this type, whether they cost $2,000 or well over $10,000, have one thing in common: inside, they use an off-the-shelf, mass-produced IT-grade switch board, essentially the same kind you would find in a computer store. The manufacturer adds a better power supply, perhaps changes the clock, puts everything in an attractive case and calls the result an audiophile product. That is tuning, not engineering. The source of the problem - noisy switching converters and an architecture designed for the office rather than audio - remains untouched. Much of the price difference comes down to the extras. The N1 was built the same way as our servers, starting from a blank sheet of paper. At its heart is a fully proprietary motherboard designed exclusively for audio transmission over Ethernet. The Optimo N1 power supply is fully linear and has three galvanically isolated rails: one for the main circuitry, one for the clock and one for the isolated output. Timing is controlled by an OCXO clock with its own dedicated linear power supply. The port intended for the player is optically isolated at the data, clock and ground levels. There is also an SFP fiber-optic input, and the firmware was written and tuned specifically for audio. Show me another switch built this way. Developing our own motherboard and software from scratch, using expensive components and manufacturing in small quantities in Poland all cost money. It is also worth adding that the N1 was not created as a box to fill a spot in the price list. While developing the S1 and S1 EVO, we deliberately avoided using any external switches. The servers had to stand on their own. But the network’s impact on sound is undeniable, so eventually I asked myself what a switch would look like if it were designed according to the principles of audio engineering rather than IT. The N1 is the answer. And finally, I am not asking anyone to take my word for it. I invite people to listen, preferably in a system they know well. The N1 is intended for people with reference-level systems, where every component can be heard.

I understand that you hear a clear difference between a regular LAN cable and an audiophile alternative, or between an ordinary switch and the XACT N1. But wouldn’t it make more sense to upgrade your speakers, amplifier, interconnect, fuse or wall outlet - perhaps to a gold-plated, rhodium-plated model with a carbon-fiber faceplate and embedded diamonds? Or are XACT accessories aimed at audiophiles who already have absolutely everything?

I do not entirely agree with the premise of that question, because it puts a music server in the same category as outlets decorated with diamonds. Those are completely different things. The XACT S1 EVO is a fully fledged source, the starting point of the signal chain, not "just some accessory." In a system based on files and streaming, the server plays exactly the role of a turntable in an analog system - and nobody calls a turntable an "accessory." It is much the same with a switch. In a network-based system, it is not a decoration but part of the source infrastructure, through which the signal physically passes before reaching the DAC. In a file-based system, a switch is no small add-on. Of course, a system needs to be put together sensibly, and I would not advise anyone to start with a switch if the rest of the chain is underperforming. But the old rule still applies: nothing that is lost at the source can be recovered by any amplifier or speaker. They can only faithfully reproduce what they are given. That is why upgrading a system only at the downstream end can lead nowhere. As for audiophiles who "already have absolutely everything," I do not know if people like that are even my customers. Most of them simply have good, carefully thought-out systems. What is more, the benefits of the S1 EVO are clearly audible in less expensive setups, including ones with speakers that cost less than the server itself. A good source raises the level of the entire system, whatever its scale. I will end on a slightly ironic note: you will not be paying for gold, titanium or embedded diamonds in my products. With XACT, the money goes into engineering.

Do you stay in direct contact with users of JCAT and XACT products? What are some of the most interesting stories or "cases" you could share? Is there a group of fans who, after years of experimenting, will follow your advice without question?

Yes, and very directly - essentially every day. Since the beginning, I have handled technical support for all my brands almost entirely on my own, so when a customer writes to support, they are effectively writing to the designer. It takes a lot of time, but it is invaluable. There is no better way to learn how products behave in real systems than by talking to the people who use them. Many improvements and features in our software came out of those conversations. As for people following my advice "without question," there is indeed a group of customers who have bought every product I have ever released, from the first versions of JPLAY to JCAT cards and XACT servers. They are not all abroad, either. Some are in Poland, which makes me especially happy. I think that trust comes from the fact that my method has remained the same for years: every change, however small, is tested by listening before it reaches users.

I suspect this conversation will be of great interest to people who hunt down and mock what they call audiophile voodoo. A music app, audiophile LAN cables, power supplies, switches and anti-vibration feet... All that is missing is audiophile air in a spray can, or a high-end chair with a specially shaped backrest and foam developed for space shuttles to eliminate spinal resonance. How would you convince a skeptic that audiophile accessories like these are not a placebo?

I will say something skeptics probably do not expect from me: I have no intention of convincing everyone, and I will honestly admit that these products are not for everyone. Even putting financial considerations aside, people are not equally sensitive to changes in sound, and not everyone is a connoisseur - just as not everyone can tell the vintage of a wine or where its coffee beans came from. There is absolutely nothing wrong with that. I will go further, even though I probably should not say this as a manufacturer: in simpler systems, it is better to invest in the fundamental components rather than accessories. I always say that, and I have never urged anyone to start with a switch or a cable. But there is a fundamental difference between setting priorities sensibly and throwing everything into the "voodoo" bucket. What I do is not magic. It involves concrete, measurable phenomena: power-supply noise, clock stability, electromagnetic interference and galvanic isolation. No serious person disputes that these things exist. The disagreement is about the threshold at which they become audible. That is not something you can settle on an internet forum. You have to listen. That is why my method has remained unchanged for more than a decade: every change, however small, goes through documented listening tests before it reaches users. I do not sell anything I have not heard for myself. I would suggest a simple experiment to a skeptic - one that is free and completely safe for their wallet. Anyone with any kind of streamer and an Apple device can download the JPLAY app and try it free for two weeks. No cables costing thousands, no taking my word for it, just a change to the software controlling playback. If you do not hear a difference, you lose nothing and save a lot of money on audiophile accessories. If you do hear one - and that is how the vast majority of these trials end - you will have to ask yourself a question: if the app alone changes the sound, what else might change it?

What are your plans for the future of your brands? Are there still areas of digital audio that could be improved and that you would like to work on?

I will start with the second part of the question, because the answer sets the direction for everything else: yes, digital audio can still make significant progress. Contrary to the popular idea that "bits are bits," file playback is not abstract mathematics. It is physics: power-supply quality, clock stability, interference, and how the operating system and software behave in real time. Every one of these areas can be improved further, and some of them - especially software - are still greatly underestimated by the industry. That is why my first plan is to keep developing what we already have. My products are not closed systems. XACT OS is constantly evolving, and every change goes through full, documented listening tests before it reaches customers. People who bought the server two years ago are listening to a better device today than they were on the day they bought it, and that will continue. Second, there are new products. I am working on the next generation of streamer, but I am particularly excited about another project: a much more affordable model. I want to prove that XACT’s philosophy - proprietary design, an obsessive focus on power and listening-based tuning of every detail - can be brought to a lower price point without resorting to off-the-shelf, compromised solutions. That would make the brand accessible to people who have only been able to read about it so far. And then there is the JPLAY app, which I am improving pretty much all the time.

You listen to music on a very original, high-end stereo system. Both the components and your listening room are impressive. Could you tell us more about the system? Why did you choose those particular components?

My system serves two purposes. It is a source of pure listening pleasure, but it is also my main working tool - every XACT, JCAT and JPLAY product is developed and tuned on it. That is why the main criteria for choosing the components were maximum resolution, transparency and neutrality. I spent years putting the system together and fine-tuning it, and I think it will stay with me in this form for a long time. The YG Acoustics Hailey 2.2 speakers are essentially the high-end development of a studio monitor in a floorstanding design, and that is exactly what I value about them. I have never heard anything that offers such resolution and transparency while maintaining such good balance. These speakers reveal the shortcomings of most components on the market without mercy. It takes real care to find the right partners for them, but when you do, they reward you handsomely. For a designer, that is invaluable: if a change to a prototype is audible, I will hear it immediately on the Haileys. I have been associated with the brand since 2018, starting with the Carmel 2. I chose SPEC RPA-MG1000 monoblocks partly against conventional wisdom, since they are technically Class D. But their highly sophisticated power supplies and refined, distinctly Japanese tuning give them such a natural sound that they can stand alongside the best tube amplifiers. I mean that quite seriously. The source is an Esoteric K1X SE, a flagship one-box SACD player with a DAC. It complements the SPEC amps beautifully, combining resolution with a natural sound. It also serves a special purpose for me: it is the reference point for my XACT S1 EVO transport. I keep a top-quality physical-disc player in the system partly to make sure file playback always has something to measure up against. It is an honest sparring partner that keeps me from simply taking my own word for it. The rest of the chain consists of my own products - the S1 EVO, N1 switch, Phantom cables and Immotus feet - because everything I make is first developed and listened to in my own home. The room itself is an ordinary living room, and not the easiest one acoustically. Rather than rebuild it, I chose precision: diffusers in several places, speaker and listening-seat positions refined down to the centimeter with a laser, and carefully adjusted toe-in.

After so many years spent refining audio playback, can you still sit down and simply enjoy music without analyzing the equipment? What kind of music do you listen to every day? Are there any favorite albums that move you in particular or serve as references when you test prototypes?

I can, and I do it every day. Otherwise, this work would make no sense. Jazz is the mainstay of my everyday listening. My favorite artist is Bill Evans. I also listen to a lot of more recent jazz, especially from the European scene, as well as classical and electronic music. When I am working on research and development, though, I deliberately avoid getting fixated on a single album - and I would advise every audiophile to do the same.

Looking back on your journey from an enthusiast experimenting with foobar to a designer recognized around the world, what has it taught you? If you could go back in time, what advice would you give your younger self, or young audio enthusiasts who want to follow in your footsteps?

"Be careful what you wish for..." I say that with a smile, but there is some truth to it. I once dreamed that audio would become my whole life - and it did. There are consequences: customers write to me from every time zone, and the boundary between my passion and my work disappeared more than a dozen years ago. I do not regret it, though. More seriously, this journey taught me above all that it is worth having dreams, being passionate and pursuing your goals consistently. It may sound like a cliché, but in my case it is not a slogan. It is a literal description of my life story. There were no shortcuts. I started out like thousands of other enthusiasts, with experiments and late-night listening sessions. What turned a hobby into a profession was hard work and determination. The second lesson is to trust your ears and have the courage to question the consensus. If I had believed that "bits are bits" and nothing more could be done, none of my products would exist.