ne of the most well-known Polish constructors specializing in digital systems asked me recently what we see in the “whole DSD thing”. I frankly said that the answer is “nothing”, because DSD files always come across better in comparison with PCM files. What I said was confirmed by a listening session in which I participated with Dirk Somner and Gerhard Hirt within the Krakow Sonic Society (read more HERE). The session was then repeated in Warsaw during the Audio Show exhibition and over a hundred listeners took part in it (read more HERE). The differences that I am talking about are not subtle at all – on the contrary: when heard once, they will become a prism through which we will look at them from that moment on.

To cut a long story short: PCM files seem to be colored. DSD files are characterized by much smoother frequency response and their sound is more meaningful, and less “mechanical”, whereas the bass is more natural and the treble is close to what we know from analogue master tapes. These observations have been confirmed both by Dirk, who works for the Sommelier Du Son label where he makes recordings and releases them on LP discs, while also experimenting with DSD files, and Damian Lipiński who has remastered a lot of recordings in the analogue domain, which we have been able to buy in Polish shops. Recently, he has remastered two recordings of famous Polish bands: Unu by Perfect and Zemsta nietoperzy by Dżem (although it is still not certain whether they are going to be released). Damian has clearly emphasized a few times that when we listen to a master tape on a well-calibrated Studer A80 cassette player and compare it directly with other formats, a DSD file sounds best and is closest to the master tape.

However, when it comes to measurements, such devices and such type of material get worse results. It is because they are characterized by very high noise, caused by the fact that it is 1-bit signal of a high sampling frequency. The whole noise must be removed from the audio frequency with the aid of aggressive “noise-shaping” techniques. Even if this is done correctly, the noise still exists – outside 20 kHz, but it is still there. Owing to this, mathematical treble resolution is very low in this type of signal, whereas noise is very high. In comparison with DSD, PCM signal doesn’t seem to be affected by any distortions at all. Therefore, a lot of very good engineers think that DSD is a “bluff” (see HERE).

However, based on peoples’ listening experiences it has been demonstrated that a well-prepared DSD file sounds fantastic (read more HERE), similar to SACDs. As regards the latter format, things are more complicated, as SACDs do not always sound as well as CDs with the same material. This may be caused by a complex process of reading and decoding, or by other factors. It seems that audio equipment can be largely blamed for that, as almost all digital-to-analogue converters used today (DAC chips, that is) are optimized for PCM signal and “compatible” with DSD. What does it mean? It is not always known exactly, but DSD signal is often converted to PCM in these devices and only then decoded. So, decoding the original (“native”) DSD signal is out of the question.

However, if this is done like in the Accuphase DP-901 converter, we may obtain excellent results. I assume that Alex Peychev, an engineer who has been working for Sony Electronics’ Broadcast and Professional division, must have come to similar conclusions. He has proposed his own version of a digital-to-analogue converter optimized for DSD signal. What is more, he has also prepared a SACD transport to go with the DAC. The transport has a DSD output via the RJ-45 cable (“Ethernet”). It is what Accuphase also does. Alex runs his own company: APL Hi-Fi, Ltd.

The DSD-S digital-to-analogue converter which I would like to show you has an input through which we can connect SACD transport. It also takes DSD signal (up to DSD128) via a USB port. We can send PCM signal to it via the same port, up to 32 bit and 384 kHz (DXD), whereas PCM 24/192 signal can be sent to the RCA, optical and AES/EBU inputs. However, the DAC is the kingdom of DSD. It is because all PCM signal types are converted into DSD – DSD64 or DSD128 (according to a user’s selection) before being converted to analogue signal. It is because the converter has been optimized for DSD signal, including the output (passive) filters. So, the situation is different than in the case of classic converters in which PCM signal “rules” and DSD just exists “in addition” to it.

ALEX PEYCHEV
APL Hi-Fi, Ltd. | owner, constructor

I made my first stereo cassette tape deck from two mono ones when I was in high school. After that, I have designed at least 50 solid state, tube and hybrid amplifiers, as well as several speakers, EQs, tape decks, phono stages and other devices. In 1989 I went to former West Berlin to work as a support engineer for one of the largest home audio/video equipment stores at that time. When I returned to Bulgaria 3 years later, I was employed at one of the first Sony Electronics authorized service centers in my country.

In 1994 I moved to the USA and worked for Sony Electronics Inc. in San Jose, California, from 1996 to 2003. I was the lead electronics engineer in their Business and Professional Division. I was responsible mostly for professional video systems, such as JumboTron installations, professional digital video recorders, cameras and digital effect systems. Those things make any audio equipment look like, literally, baby toys. Just imagine a $150,000 professional Sony D1 video recorder to understand what I am talking about.

In 2003 I officially opened APL Hi-Fi in Dublin, California. In 2004 already, I received the “Most wanted component SACD Player” award from Stereotimes. There you will also find their review of my version of the Philips SACD1000 player. After the Philips SACD1000, I started re-designing Pioneer and Denon players, but my most well-known project is the uncompromising NWO player based on the Esoteric UX-1 player. A review of the NWO-3.0-GO version was published by the 6moons.com magazine. The latest version of the player is the NWO-M model.

Working on it was very difficult and included using my proprietary clocks, D/A converters, a tube output and so on. Still, people thought it was just a “modification”. This is the reason why I've decided to design my own unique line of audio components. Instead of outsourcing production to a Chinese factory, I assumed I would do everything in my own high-end all-handmade boutique. So, I returned to Bulgaria where I was given help by people I can trust. Today, we offer a complete audio system: from source to speakers. I have been able to exhibit the entire audio system a few times and it was always rated as “the best sound of the show”.

The DSD-S converter was designed with one goal in mind – to obtain sound as close to analogue sound from an all-solid-state D/A converter as possible. It is based on a modular structure which offers a client-friendly upgrade path. Like the rest of our products, the DSD-S is handmade in our manufacture. This includes manual population of the boards inside with components as small as 0.3x0.6 millimeters. Every unit is hand-built from scratch. 

The sound of our digital components has been inspired by the sound of the Kuzma Stabi-M turntable with the Kuzma 4POINT arm and the Dynavector XV-1s cartridge, as well as and our reference PHS-M phono stage. The DSD-S is the product of a long design process. At this point I don’t see any possibility of improving its sound and I don't think there will be any upgrades available for it in the near future. Personally, I am very happy with it.
It is possible to order an optional balanced output and Franc Audio Accessories anti-vibration feet for the device.

Recordings used during the test (a selection)

COMPACT DISC

• Tommy Dorsey, Masterpieces 15, EPM 158342, “Jazz Archives”, CD (1935-1940/1995).
• Miles Davis, The Complete Birth of the Cool, Capitol Jazz/EMI 4945502, CD ([1957] 1998).
• Nat ‘King” Cole, Penthouse Serenade, Capitoll Jazz/EMI 94504, “Super Bit Mapping” CD (1952/1998).
• Krzysztof Komeda, Ballet Etudes/The Music of Komeda, Metronome/Be! Jazz Records, BE! JAZZ 6087 CD, CD (1964/2014).
• Thelonious Monk, Solo Monk, Columbia/Music On Vinyl MOVLP843, “Classic Album”, 180 g LP (1965/2014).
• Grover Washington Jr., Winelight, Elektra/Audio Fidelity AFZ5 203, “Limited Edition No. 0115”, SACD/CD (1980/2015).

BLU-RAY AUDIO

• John Coltrane, A Love Supreme, Impulse/Universal 3735663, High Fidelity Pure Audio Blu-ray, PCM 24/96 | Remaster 2008.
• Queen, A Night At The Opera, Island/Universal 3732771, High Fidelity Pure Audio Blu-ray, PCM 24/96 | Remaster 2011.
• Franz Shubert, String Quintet C major D 956, wyk. Auryn Quartet, Tacet B110, Pure Audio Blu-ray, PCM 24/96 (2014).
• Ellen Sejersted Bodtker, Sonar, 2L 2L51SABD, Pure Audio Blu-ray PCM 24/96 (2008).
• Opeth, Pale Communion, Roadrunner Records RR757375, Blu-ray Audio, PCM 24/96 (2014)
• Depeche Mode, Delta Machine, Blu-ray Audio 24/96 (2014) [w:] Live in Berlin, Sony Music | Columbia 5035642 (2014).

+ DSD, PCM, DXD FILES

Japanese issues available at

While listening to the sound of Mr. Peychev’s converter, I knew right from the start that I had somewhere already heard sound of this type and that I had dealt with similar sensitivity to musical emission before. Having listened to a few recordings, I knew what it was: the DSD-S sounds similar to the Accuphase SACD DP-720 player (read more HERE). These devices were created at two places located far away from one another geographically, but they treat music in a similar way, building up structure as if their creators had been driven by the same thought. Who knows – perhaps they have a shared goal, even though they come from different countries and cultural circles, and speak different languages.

It is mainly about integrating the smoothness and openness of sound. It does not matter what recording we are listening to and what material we are using – these two features will dominate, even when it comes to such old material as Tommy Dorsey’s collection of recordings from the years 1935-1944. From a technical point of view, these are very weak recordings, but have something true in them that can be heard on any reproduction device. The DAC presented the recordings very precisely, without trying to warm them up. There is a lot of upper midrange, but little bass and treble in this music, because of which it is hard to play it without covering the authenticity with tiring hullabaloo.

The tested DAC does not warm anything up, it’s not the case. It precisely shows the treble, which made it possible for me to analyze recordings from different years, with different types of noise, and a different approach to the percussion plates and wind instruments. The sound was clear, but never vociferous. It was explicit, but not in excess. The changes that I am talking about resulted from excellent differentiation and were not an attempt to show everything – these are two different things. In the former case we choose what interests us most in the musical transmission, whereas in the latter case the dominant is chosen for us by the musical transmission.

However, I focused my attention on music, not on elements outside it. Miles Davis’s album Birth of the Cool released five years later opens a completely new, much more complex and modern chapter in jazz history. This early Davis’s recording shares all the problems common for recordings of that time, but its sound is much more selective and distributive than the sound of Dorsey’s recordings. Chronologically, a few or several years passed between the dates of release of these two albums, but, musically and technically, that was a whole era. Davis’s iconic album (my version is a Mark Levinson’s remaster) had a smooth, dense and clear sound with the Bulgarian DAC. The definition of sounds was excellent – it was already clear in the case of the first album, but here it was even better.

However, sound reaches its maturity only with Nat “King” Cole’s recordings from the year 1952 and remains mature for the whole period of the 1950s. Cole is known by everybody, especially thanks to his unique voice, but we cannot forget that he was first and foremost a great pianist. The Penthouse Serenade album that I am talking about is an instrumental album, because of which we are not distracted by the vocal. The leader’s piano was clear and distinct, and it had the right beat and reverberation. Nothing in its sound was rounded or hazy. To be honest, high treble was a little better than in the case of the reference player – the Lektor AIR V-edition. It is because it was better differentiated, there was more information in it, but given in such a way that there was not too much of it.

The percussion was presented even better on Krzysztof Komeda’s album Ballet Etudes, released in 1963 by the Swedish Metronomy company, recently reissued by BE! Records. Eleven years had passed since the release of Penthouse Serenade”, so it was possible to define the sound attack and fill in the double bass better, showing the hidden power of this instrument. However, at the same time, which was nicely shown by the DAC, we are entering a period in music history when the naturalness of sound, perhaps coming from tube devices, got lost somewhere – at least in the case of some recordings. As far as the Solo Monk album (a solo piano record) from the year 1964 is concerned, it is a return to the full glory of tangible, close sound. Thanks to the fantastic resolution of the DAC, the piano had a deep sound and an excellently presented definition – from the bottom to the top end.

So, we are talking about resolution, selectivity, precise top end and bottom. The range of the bottom end is a bit emphasized at about 200 Hz, which results in contoured, selective sound. There is a lot of bass in general. Despite this, the centre of gravity in the midrange is set higher than in my Lektor, but also higher than in the Totaldac d1-twelve D/A converter that I tested at the same time. As a result, the sound of the recordings from the 1970s (Shamek Farrah’s First Impression and Bill Withers’s Just As I Am) is improved a little, although these recordings usually react “nervously” to such treatment.

The vocal of the latter had a brighter timbre than in the case of both devices that I mentioned above. However, it was not distorted. The DSD-S is a high-end product which, by correcting sound in its own “style” (all devices do that), simply proposes a different look at the material which is, at least when it comes to tonality, neither better, nor worse.

Perhaps this is why Whithers’s voice in the only vocal track on Grover Washington Jr.’s album Winelight, which has just been released by Audio Fidelity and was originally issued in 1980, has not lost the warmth and the velvety quality that it is famous for. The 1980s are considered to be a “black hole” in the art of sound engineering, but I don’t agree with that. Well-produced recordings, carefully prepared in an analogue studio, can do miracles. It is the case with the abovementioned Grover’s album – it is very soft, almost “flowing”. The DAC lightened it up a bit, contoured it a little, but has not changed the basic structure, has not destroyed it.

Summary

I would like to finish with this beautiful example of smooth-jazz, because the tested device presented it in a way which well sums up what the DAC is and what it is not. It is a perfectly resolving device. It shows information that one cannot actually hear with other DACs from the same price range. Devices which are “brightened up” (both more expensive and cheaper ones) do not show this, because they only brighten up the sound. The DSD-S does not brighten anything up – its upper treble simply has more energy. It is similar with the bass – there is a lot of it, it is controlled and defined, and the attack is emphasized a little.

A D/A converter which produces such sound seems to be designed for amplifiers which provide a bit warmer sound, in which the midrange is dense and full, but lacks support in the form of good top and bottom ends. Tube amplifiers are mainly coming to my mind, but the Accuphase E-470 integrated amplifier would also be “happy” to play together with the DAC.

As far as playing CDs is concerned, the DAC operates very well and competently. CDs are what I’ve been writing about, but the DAC is probably even better with high-resolution signal from Blu-ray Audio discs and files. Its sound was especially good in the case of DSD files. There was an amazing amount of space – it was much broader than from the reference player – and the dynamics was also excellent. It is important to remember to match the DAC with a balanced and perhaps even a little warm audio system, and then the advantages of the DSD-S will be a blessing.

Mr. Peychev’s DAC has a classic look for this type of devices. There is a thick front aluminum panel, a large blue LCD in the centre and a few buttons. Everything is placed in rigid aluminum casing made of plates and profiles. So, this is a very rigid, vibration-resistant structure which protects electronic circuits from EMI and RFI interference. There is also a detail which tells us a lot about the company’s approach to issues connected with vibration damping: instead of classic feet, the converter is equipped with four flat wooden cones which resemble the ones that I have seen in Japanese SPEC amplifiers. At a client’s request, the converter can be equipped with Franc Audio Accessories feet produced in Poland.

The LCD displays volume (both on a bargraph and alphanumerically), the chosen input and the digital filter applied to it, i.e. conversion to DSD. There are five buttons: standby, a display-dimming button, input change and two volume control buttons.
The sockets used in the DSD-S are of high quality. The RCA sockets were manufactured by WBT (it’s a model from the NextGen series), whereas the AES/EBU (XLR) input is a product of the Neutrik company. There are six digital inputs: a Toslink optical input, USB, DTR, 2 x RCA and AES/EBU. In the tested device, there was only an unbalanced output on the abovementioned RCA sockets. Optionally, one can order XLR outputs. The Schurter IEC socket is also very good. Next to it there are two fuse caps – of the standby and the main fuse. The latter is a high-end expensive AMR fuse with gold-plated conductive elements.

The device has modular construction. As Mr. Peychev says, it future-proofs the product. If a new version of any of the modules is created, it will be possible to exchange it without the necessity of selling the whole converter. The exchange will have to be made either by the producer or by a certified service, because the boards are connected to one another using cables and solder, not pins. It improves the quality of signal transfer, but does not make it possible to exchange the boards quickly.

There are six boards with: a standby chip, an amplifier, display control, digital inputs, signal processing and D/A conversion, as well as with an analogue output circuit. Next to them there is a powerful double-C transformer, described as “P-Core”. It is really enormous – I have seen transformers of exactly the same size in YBA integrated amplifiers (read in Polish HERE ). Four secondary windings come out of it, separate for each module. Rectifier diodes and resistors have enameled housings, so that one cannot identify them (and, as I assume, copy the circuit).

Digital inputs are based on an XMOS (USB) and a Cirrus Logic (RCA and XLR) receiver/converter. They are coupled using matching transformers. There is a separate integrated circuit next to the DTR input. It is here where the ultra-precise clock chips are located – two, for two timing “families” (44.1 and 48 kHz) and DSD. They are a product of the Crystek company. The clock for the XMOS chip (i.e. for the USB input) is also very good.

After selecting the appropriate input, signal is sent further, to the board where it is processed and converted. The first step is to change all signal types into DSD (DSD64 or DSD128), which is done using the DSP Xilinx Spartan chip. Next, signal is transferred to four (two per channel) stereophonic digital-to-analogue Cirrus Logic CS4398 converters. They operate in a circuit developed by Alex Peychev, optimized for DSD signal conversion.

I will remind you that Accuphase also came up with a similar idea. For DSD signal, it uses MDSD (Multiple Double Speed DSD) technology in its SACD players and converters, which allows for most possibly direct conversion of DSD signal into analogue signal. First, signal is up-sampled from 2.8224 MHz/1 bit to 5.6448 MHz/1 bit and then delayed in a special sequence, in programmable ultra-fast FPGA (Field Programmable Gate Array) circuits. Finally, it is decoded in separate D/A converters: the first one receives signal without delay, the second one minimally later, the third one has to wait twice as long, the fourth one – three times longer, etc. As many as eight “DACs” per channel are used in the Accuphase DC-901 converter. On the whole, all the sections comprise a kind of a low-pass filter, but without a filter – it can be found in every SACD player, but it has never been constructed in such a perfect and puristic way.

After decoding, analogue signal is sent to the output board. There are four integrated circuits on its input (the symbols are painted over), as well as one Lundahl transformer per channel. Apparently, the transformers couple the stages together. Next, signal is passively filtered and finally buffered in MOSFET transistors, in a circuit without feedback. According to company materials, it is a unity-gain circuit (i.e. it has a gain of 1). There is an active DC-offset circuit on the output, without capacitors. Internal connections are based on solid-core wires made of OFC copper.

KRZYSZTOF KOMEDA
Ballet Etudes/The Music of Komeda

Metronome/Be! Jazz Records, BE! JAZZ 6087 CD, CD (1964/2014)

For me, the godfather of jazz in Poland is Krzysztof Komeda. The best living musicians did their apprenticeship with Komeda, so I know that through playing with Namysłowski, Stańko or Urbaniak, I partly caught his message.

Leszek Możdżer, Muzyka jest wibracją (Music is vibration), interview by Wojciech Sroczyński, “Newsweek. Kultura” March 2015, p. 4.

It is not only Możdżer’s opinion. For example, in the foreword to a photo album entitled Czas Komedy (Time of Komeda) with photographs by Marek A. Karewicz, we read that “the music of Krzysztof Komeda-Trzciński has grown to become the symbol of a generation, becoming our cultural pride” (Marek A. Karewicz, Dionizy Piątkowski, Czas Komedy, Poznań 2013, p. 9).

On May 3rd 1963, in the Metronome Studio in Copenhagen, Krzysztof Komeda recorded his first full-scale album with his own material, released by the Metronome company in Sweden only. In the album he presented his Ballet Etudes (known in Poland from Jazz Jamboree 1962) and two themes important for his career: Alea and Crazy Girl from the movie Nóż w wodzie (Knife in the Water). He is accompanied by a splendid ensemble of excellent musicians – the irreplaceable Jan Ptaszyn-Wróblewski on tenor saxophone and Allan Botschinsky on the trumpet, whereas the whole composition is driven by Komeda’s favorite rhythmic section: Roman “Gucio” Dyląg on the double bass and Rune Carlsson on the percussion. This is where the album subtitle comes from: A Jazz Message from Poland Presented by an International Quintet. As we read in the note, the recording took place on a rainy, gloomy day…

The sound of the recordings has been very carefully renovated. It is dynamic and very clear. There is no treble cut in it, typical for many recordings from that time. The treble is in an excellent proportion with the rest of the range and the sound of percussion plates is its natural extension and not only the background. The sound can be described as very precise and attack-oriented. The attack is neither fuzzy, nor rounded at all – everything is clear and very selective. Instruments placed at the back are excellently emphasized, with the trumpet placed further away, clear percussion and quite contoured double bass.

What I missed was a little warmth, some natural sound filler. I do not know whether it results from the recordings themselves or from their remastering, but it is something we simply won’t get. I also have the impression that the interference in noise reduction was too deep. I do not mind hearing tape noise at all, especially that there is a lot of micro-information on the instruments in it. In the case of Komeda’s album, effort has clearly been made to make sure there is absolute silence between notes.

However, there are no ideals, so I gladly recommend this album. It is beautifully issued and it sounds great, although it is “only” a mono recording. Its vinyl version is also available, so I naturally bought it (BE! JAZZ 6087, 180 g LP, 2015). It is a limited edition (500 discs), so don’t waste your time!

Quality of sound: 7/10