D/A CONVERTER

CASEA
CEPHEUS

WOJCIECH PACUŁA
Translation: Krzysztof Kalinkowski







The converter of the Czech company Casea fell into our hands due to the mediation of the company's representative from a country, from we received the baptism for the court of prince Mieszko, later named Mieszko I. It attracted my attention from the very beginning due to it's solid construction, worked out circuitry and also due to something that more and more often passes through our mail HFOL: a possibility to improve the sound of a home theater system. A problem of the home theater system, besides the amplituners themselves is the sound of the DVD player – most of the time not very good, diplomatically speaking. The company Casea has a solution to that problem: upgrade of the DVD players (of universal players) – among others Denon and Pioneer, equipping them with a special digital output, with a dedicated DSP, that can output everything – from the classic PCM in the 16/44.1 and 24/96 formats to the – seemingly impossible – 24/192 and the SACD signal (yes, from the stereo SACD layer). “Squeezing” the best possible sound out of the AV devices is one of the main problems, and such a comprehensive service, like the one provided by Casea, is especially welcome.

Cepheus is one of the most technologically advanced devices I saw lately. As I mentioned before, it is able to decode the PCM signal up to 24/192. The problem is, that a consumer standard for transmitting a signal of that density does not exist. In dCS and Chord the signal upsampled to 24/192 was sent to the converter with 2 110Ω cables in the 2xAES2 standard. S/PDIF transmitted by a 75Ω coaxial cable was conceived to transmit signals up to 24/96. At least that was the idea. The bottleneck appeared to be the transmitter and receiver, elements that underwent a significant upgrade lately. And so the company Audiopraise, modifying the transports, equipped the converter with the capability to receive the signal at maximum rates applicable in DVD-Audio discs. However to take advantage from those capabilities a special transport is needed. As mentioned before Audiopraise developed a special module, with the name Vanity, with a DSP processor and a BNC output, that can be mounted inside any type of a DVD-Audio or universal player. The last possibility is especially intriguing, as by the same cable the signal from a SACD disc can be transmitted... How is that possible? I asked this question to the man that conceived all this and implemented for manufacture, Mr Dominik Peklo:

HIGH Fidelity OnLine: What is the responsibility of the company Audiopraise in the Casea-Pioneer combo?

Audiopraise: In Audiopraise we are only responsible for the transport part of the system. The converter Cepheus is an independent product of the company Casea, that is also responsible for selling the Vanity modules with a high resolution digital output.

HFOL: What signals can be transmitted from the modified player?

A: Through a standard S/PDIF link we can send outside PCM signal from the CD (16/44,1) and DVD, including DVD-Audio, so up to 24/192.

HFOL: I thought that the S/PDIF protocol has a limit on the amount of data that can be transmitted and nothing above 24/96 would go out...

A: In reality, this protocol has no set upper limit of the sampling frequency it can “handle”. It's capabilities are determined by the quality of the receiver and the circuit at the converter end, and those days those are adapted to work with the 192kHz sampling frequency. Because in the Cepheus such a modern receiver is implemented, it is capable of decoding two channels with such a sampling rate.

HFL: I am also interested in another thing: You state in your materials, that with the same output the Vanity module sends the SACD signal. Is it DSD?

A: At the moment there is no commercially available standard of DSD signal transmission – Sony does not allow that. In addition – no external converter will “understand” the signal. We would be able to transmit the signal out very easily, but the drive would have to work only with our dedicated converter. And this is against the idea of Vanity, that is to be an as much universal module as possible. This means that before sending the DSD signal out we convert it to PCM. The sampling frequency can be chosen – 88.2kHz or 176.4kHz.

HFOL: Vanity is just a small PCB in the Pioneer, right?

A: Yes, Vanity is a PCB near the BNC output, mounted near the back plate of the Pioneer. This module can be implemented in every universal player and is in no way related to the Pioneer itself. The circuit is built-up around a high class programmable FPGA chip (Field Programmable Gate Array), that works together with a digital filter developed by us and a S/PDIF transmitter. I encourage to visit our web page: Vanity, where much more information can be found.

HFOL: Thank you for the conversation.

A: Thank you too – in case of any questions regarding the Vanity, I am always here to help.

So much from Dominik Peklo from Audiopraise. From this conversation we can draw some interesting conclusions. We can send signals from DVD-Audio and SACD to the Cepheus. The first will be sent in the way they were recorded on the disc, up to 192kHz, the latter will be converted to PCM. The first thing – super, no comment needed, the second needs one. One of the advantages of DSD is a significant simplification of the converter. So converting the signal to PCM we have to deal again with the problems we tried to avoid. Even more – conversion between the standards is a tricky thing and not always worth it's while. Let us take for example SACD discs where the material was converted from PCM master tapes, like the discography of Peter Gabriel or Depeche Mode. It turns out, that the majority of SACD discs on the market were created from PCM masters. The second group are the SACDs created from analog master tapes and only at the far end there are discs natively recorded in DSD. And even here is a problem, as there is no such thing as DSD signal processing, as this format was firstly developed for archiving of data and had nothing to do with audio. So if we want to mix and master the signal in the digital domain, then we have to go to PCM, I mean the DSD mastering tables convert the signal to PCM, process it in that form, and then convert it back. So among SACD disc we have about a percent of discs that were recorded in DSD and mixed and mastered without any processing to two channels.
But what I am coming to: the conversion of the signal is not the best thing around, but it happens commonly. In the hi-end we often see the situation that the PCM signal is converted to DSD and only later to analog. This kind of upsampling was introduced in the players of dCS and Esoteric. But it also happens the other way round, as some of the universal players from Pioneer (eg. DV-696AV), or Cambridge Audio (Azur DVD89) just after reading the signal from the disc converts it to PCM to process it in the same DSP chip. But the same is done by a company that knows it's way around with expensive products, Arcam in the new series of players, the DiVA DV135, DiVA DV137 i FMJ DV139. And, frankly speaking, I think that from the two bad things, it is better to convert from DSD to PCM than the other way round. Testing the dCS I spent much time comparing the plain CD 16/44.1 signal with the upsampled to DSD, and I think, that without hesitation the first one was better. The Vanity does just something like that – converts DSD to PCM and sends it out to the converter. How does it work? We shall hear in a moment.

I approached the test of the Cepheus a few times, for once due to it's broad capabilities - CD, DVD, DVD-Audio and SACD, balanced and unbalanced outputs, but also because I wanted to listen to it paired with an as broad as possible range of products. As a reference point served: used by me for a long time universal player from Pioneer DV656A (identical to the one modified by the Vanity module), universal player Arcam FMJ DV139, DVD-Audio player Arcam FMJ DV29A, SACD player Accuphase DP-800/DC-801, SACD player Denon DCD-CX3, D/A converter AudioNemesis DC-1 and the CD player Ancient Audio Lektor Prime. The Arcams and the Ancient were also used as optional transports for the Cepheus converter. Let us say, that with CDs the Prime was without competition, giving a more stable, more defined sound than the modified Pioneer, but with the DVDs the Arcam was only by a thread more resolving. I used also a few other digital cables like the Oyaide i XLO from the Limited series – with more expensive systems both were better than the one supplied with the system, but the Oyaide not in all aspects, but in systems, where the price of an element does not exceed, let's say 6000-8000zł the differences are so marginal, that there is nothing to get worried about.

SOUND

During the first few discs, when I listened to the Cepheus connected to the Prime and the Prime itself, switching quickly between both sources, I knew they were different products, but I could not pinpoint the difference at first. This is a big praise for the converter – sauté it costs ten time less than the Lektor (of course you have to add the cost of a drive and digital cable, in this case about 10000zł).However A/B testing show only a part of the truth and I usually treat them as something additional. The devices differ, and by large, in many areas crucial for the hi-end, but comparison of converters costing less than the Lektor with this player often showed further reaching differences. The Cepheus held it's stand firmly, and with esteem I went on with the listening tests.

The thing, that makes the converter perform so well, and in many areas surpasses much more expensive players (not all of them, but about that in a few moments) is superb tonal balance. It is the same school as the Prime and the Arcam DV29. The sound is open, quick, dynamic, without euphony and coloring. There is also no trend visible to have it sound “tubey”. Strong treble is a sign of that. It is not overdrawn and very clean. Interestingly, in that area the converter handled CDs best. With the exemplary, in all aspects, disc Blue Monk of Eric Reed Trio (M&I Jazz [Japan], MYCJ-30386, CD; the disc was awarded by an incredibly influential, Japanese magazine Swing Journal) the treble was strong and at the same time it sparkled with timbres, showing differences in hitting different cymbals, reproducing very well the accompanying reverb. The same happens on the other side of the sound spectrum. The bass sounds a bit stronger than from the Lektor, but it has a dynamic, dense character, that translates to enlarged drawing of the instruments. Because in spite of the treble being strong and active, the sound as a whole is slightly darker and situated a bit lower than in the Prime. Actually strong bass and the lower midrange make the drawing a bit larger and thus seemingly more solid.
With DVD-Audio discs – the phenomenal transfers from analog tapes of Muddy Waters – Folk Singer (BMG/Classic Records, HDAD 2008, DVD-A 24/192 + 24/96) and John Coltrane - Blue Train (Blue Note/Classic Records, HDAD 2010, DVD-A 24/192 + 24/96) – compared to the Arcam DV29, which has the tonal balance and other priorities, meaning first of all the resolution, similar to the Lektor, the Cepheus driven by the Vanity module sounded a little slower. A bit warmer, but it is not about warmth here, but without such well defined drawing. The instruments had a large volume, I mean they were true with their size, but from time to time I had the idea, that the bass is a bit too strong. In systems with stand mount speakers this will not be heard, but with larger speakers you have to be careful. But the Arcam had (it is being replaced by the DV139) a three times larger price tag, and was the best DVD-A player up to 10000zł and more. A quick comparison with the best universal players from the low price level (up to 2000zł), I was preparing in parallel for Audio (test – „Audio” 4/07), namely with the Denon DVD-1930 and the Yamaha DVD-S1700 showed, that there is a huge distance between them and the Cepheus, mostly in the area of resolution. .

Very interesting was the listening session of SACD discs. At first I compared the Casea/Pioneer system to the Arcam DV139, but after some time it turned out, that the British player has problems with the firmware, including the decoding of the stereo SACD layer. That is the result of trying to have the test as early as possible – I received the player fresh from the production, two months before the official premiere, and it turned out, that in the last moments some improvements in the programming of the Zoran Vadis 888 chip, improvements my unit did not have. I sent back the unit, and am still waiting for the replacement. So my observations from this comparison are not trustworthy. The standard was set by the Denon and Accuphase players. And it turns out that the conversion to PCM did not ruin anything. The sound was smooth like from SACD and had very well defined edges like from good PCM. The resolution of the treble was not as good as from the Denon DCD-CX3, but the midrange and the bass were better. However while comparing the SACD layer from the very well done sampler of Stockfisch Records - Closer To The Music Vol. 2 (Stockfisch, SFR 357.40006.2, SACD/CD; where, nota-bene, the signal was converted from PCM to SACD) played by the Cepheus and the CD layer from the same disc played by the Prime, a similar effect as with the CD could be heard – slight slowing down of the sound. Again, this is the outcome of enlarging of the virtual sources and bringing them closer to the listener. I did not write about it yet, but this is the main differences between the Cepheus and more expensive sources.

On all discs the Cepheus enlarges and brings closer the virtual sources. It is not heard from the beginning, because by the little deeper and darker sound, the reverb accompanying the instruments and voices are sounding longer than with more expensive devices, and seemingly, the space seems even larger. This is not true – one can have such a feeling, but if we listen closer, change a few discs – not necessarily from the upper sonic shelf – we will hear, that the space is defined worse, that it is covered by the strong direct sound, and the reverb does not define the space, but are only artifacts of the main sound. This is of course a very high level and in cheaper systems, matched in terms of the prize, the differences will not be that fundamental, but everywhere above that, they will surface. We can even say, that in systems based on inexpensive tubes, let's say the tested in this issue amplifier Melody SP9 or a nicely sounding transistor like Primare I21 or I30, Arcam A90 or something from NAD http://nadelectronics.com/index , then it will sound better than the Prime. The latter is quite unforgiving for the discs and accompanying devices and in cheap systems will rather underline their weaknesses than advantages. The Cepheus, on the other hand, plays rather direct sound, so it creates a sound, that can be liked in such a combination.

One of the ways of such presentation is the underlining of the attack of the sound, it's first phase, and this is exactly what the Casea converter does. Everything is quite strong and distinct. It is not sharpening or hardening of the sound – as I have written in the very beginning, the tonal balance is very good. Both in the badly recorded, but my favorite disc of the group Diary of Dreams - One Of 18 Angels (Accession-Records, EFA 03618-2, CD) with prog-electronic-gothic music, as well as with jazz-swing material from the disc of Joe Williams - A Man Ain’t Supposed To Cry (Roulette Jazz/EMI, 71340, CCD), also not very best sounding, the Czech converter sounded without a trace of roughness, maybe with little emphasis on the rough accents in the voice of Williams, but in general without a thorough analysis of the sound tissue, what in this case would result in pointing out the faults. We could even say, that the Diary of Dreams disc sounded better from the Cepheus than from the Denon or Arcam, because the virtual sources were enlarged and did not have a good focus. It was best shown by the Prime with it's outstanding focus: it's sound was thinner, and on more expensive system was perceived as more detailed, telling a story. The converter from Cesea showed the sound with a bit foggy edges of the instruments, what in the same system seemed better at the beginning, but in a few moments ended with lower amount of information. It could be easily heard on the Frank Sinatra disc – My Way (Reprise/Warner Bros., 27049, CD). Interestingly, this is a disc from the early years of the CD with a warning: “Music on this CD was originally recorded on analog equipment, before new techniques of noise reduction were introduced. The CD contains, as closely as possible, the sound of the original recording, but due to it's high resolution it shows the limitations of the master tapes, including noise and distortion”. As it turns out, self-confidence was only one of the symptoms of this medium's “child's disease”... But to the point: the sound of this disc is not rousing, mostly because the voice of Sinatra is recorded worse than the rest of the instruments, and in addition, with completely different acoustics, and thus functions completely separated, not connecting to the rest of the band. Prime showed this momentarily, just like the almost mono character of the recordings of Diary of Dreams earlier. Cepheus, on the other hand, due to bringing everything closer and washing out the edges, “lost” those shortcomings somewhere, put them under the carpet. And? And with the Melody it sounded much better, giving a fuller, nicer sound.

The Cepheus converter is an incredibly versatile (in combination with the Vanity module) successful product. Comparing it with more expensive products it proves itself splendidly, and although it can be heard, that it costs less, it is not a big problem. Comparing it to my reference in the few thousand złoty price range, the brilliant converter DC-1 from AudioNemesis, it is clear, that those products are like hot-cold, yin and yang. DC-1 is very analog, but it has not so defined drawing as the Cepheus. The palpability of the sound of the Italian converter is just incredible, an the other hand, it's dynamics does not match that from the Czech product. So if you have listened to the DC-1, and thought, that it sounds too quiet, that it lacks treble and bass, then the Cepheus will be the response to those accusations. However when the priority is the “analogicality” and continuity of the sound, then AudioNemesis will be the only choice. But it will decode only the signal from CDs, while the Cepheus can handle all types of discs. And using the Vanity module, also the SACDs.

DESCRIPTION

Full name of the device is as follows: Casea Cepheus High Definition Audio DAC, what is related to the fact, that the device decodes the audio signals of high resolution, 24/192. And the converter is not big, as it’s front panel has the same dimensions as the accompanying universal player from Pioneer DV-656A (no longer in production). It is a nicely treated plate of scratched aluminum, with a deeply engraved name CEPHEUS. Under the name a small red LED is placed, indicating that the DAC is connected to the mains. And here a little digression, related to ergonomics of the product: there is no indicator that the signal has been locked. And it would be sufficient to use a two-color LED – green meaning power is on and red that the signal is locked. And if – yes! If – the installation of LEDs indicating the input signal sample rate would be possible, not even mentioning the word length, I would be fully satisfied. I know, that the Cepheus does not cost much, and there is not much room left in the budget, but you can always dream.
If we look at the front panel, we see that the sides are just for decoration, they mask the smaller dimensions of the chassis. I don’t know if it would have been better to make a chassis with the size of the front panel, but there is not much to discuss about. Anyway, the chassis is made from very thick, rigid metal plates, varnished in black. On the back pane a small surprise: the converter has unbalanced outputs (with gold plated RCA sockets) and balanced ones (plain XLRs). There are two digital inputs – TOSLINK and BNC (both in the S/PDIF protocol). Next to them an IEC mains socket was placed wit a small power switch above it.

After unscrewing the top cover, it turns out, that the enclosure could have been even smaller, because one PCB, where all the circuitry was placed, is only marginally larger than the enclosure of the DC-1 converter from Audionemesis. The layout is clean and logical. Some surprise is caused by the writings on the PCB and the converter’s shield, stating that it is a 24/96 converter. So what is the role of the small DIP switches allowing us to set the sampling frequency to 192kHz? I don’t know. Anyway, the PCB is setup in the following sections: power supply 1 (initial part), power supply 2 (rectification), converters and the analog section.

The digital part: the digital S/PDIF signal can be supplied to the BNC connector or the optical input TORX173. this section has been placed in a metal shield, that cannot be dismounted. It is however known, that in the input the digital receiver CS8416 from Cirrus Logic was used, that due to an independent PLL loop oscillator allows for a significant reduction of jitter. Combined with a dual layer PCB, splendid power supply and shielding of the digital section, we should have a guarantee of the lowest distortions (jitter) already in the DAC section. According to the company materials, in the converting section the chip CS43122 from Cirrus Logic is employed. It is a very nice, 24-bit stereo digital to analog converter with a maximum sampling frequency of 192kHz with an integrated digital interpolative filter with a 5-bit delta-sigma modulator, that drives a second generation dynamic-element-matching (DEM) selection logic. The output from the DEM block controls the input to a multi-element switched capacitor D/A converter/low-pass filter with fully differential outputs. This multi-bit architecture features significantly lower out-of-band noise and jitter sensitivity than traditional 1-bit designs. The advanced second-generation DEM guarantees low noise and distortion at all signal levels. The analog part: To those circuits special attention was given. Employed here are high quality opamps, which have erased markings (intellectual property is the most expensive part of this circuit). The symmetric analog signal is being processed symmetrically starting from the converter and ending at the XLR sockets. In the signal path there are no coupling capacitors. DC servos were used to control the DC voltage on the outputs. There is a 3rd order analog low-pass filter applied, and it’s characteristics are matched to the digital signal supplied. If the sampling frequency is set to 48kHz then the frequency response reaches 34kHz (at -3dB). If the S/PDIF signal is sampled at 96kHz then the LPF reaches -3dB at 90kHz. The Cepheus is equipped in a delay circuitry, that switches on the outputs about 10 seconds after switching on the power. After switching of the unit by means of the power switch, the outputs are being switched off immediately to prevent any distortions from the employed amplifier. However it has to be noticed, that in the speakers tree bumps can be heard when the unit synchronizes with the input signal. I would prefer not to hear them.

Power supplies: in the company materials, it is stated, that during the design of the power supply for this D/A converter, special emphasis was laid upon elimination of all high frequency noise, that could pass on from the power supply on to the digital circuits and cause unwanted high jitter, that would cause the effect of granularity, aggressiveness, dryness and speeding-up of the sound. Due to the high level of high frequency noise being present in the AC lines, the input part of the power supply has a discrete LC filter. Next two impregnated toroidal transformers of the known company Talema were used. The larger one, with 15W for the analog section and a smaller one – 7W for the digital section. After the secondary windings PolySwitch fuses were placed, and behind those rectifying elements with ceramic anti-noise capacitors. Just on the inputs of the given DC voltages LC filters with chokes and electrolyte capacitors of large capacity are used, together with parallel placed 1μF polypropylene and 100nF capacitors to eliminate the parasitic high frequency noise not filtered by the input LC filter. In total a quite capable noise filtering circuit, isn’t it? So the high frequency noise is filtered at the very beginning. A revolutionary novelty in connecting the power supplies are the so called “virtual ground”. Unfortunately, most of the competitors used the middle wire of the secondary winding of the transformer as the reference ground, that all circuits used for reference. From the standpoint of elimination of power supply noise this is not a favorable setting, because due to the capacitance of the transformer, the noise from the primary winding has free way to the secondary and from here directly to the circuits. Such a ground, with noise, is the source of many problems and causes faulty function of sensitive digital and analog circuits. In the Cepheus converter the ground is created from a filtered reference voltage, that can be regulated. Such high quality groundings are two in this DAC, completely separated from each other. One is for the digital “ground” and the other for the analog. Two groundings were used to separate the digital and analog domains as much from each other as possible, and reduce the possibility of a feedback loop between them. Most manufacturers use as a regulator and stabilizer one of the circuits like 78xx/79xx or 317/337, but this is not the best solution for sensitive digital circuits and audio amplifiers. Voltage drops and noise on the output have a “hard” character, and their removal is really not easy. Adding a big electrolyte capacitor on the output does not resolve the issue, as it is completely not functional at high frequencies. A RC circuit could be effective, but it would lower the stabilization quotient. In the Cepheus regulators for important voltages are made from discrete elements. Voltages made in relative to the reference voltage are thoroughly filtered and deprived from noise. The actual regulation is provided by cooperation of precise, low-noise opamps and power transistors. Such a technology allows for a perfectly filtered voltage, that is so needed for the proper operation of important, and very sensitive circuits. The Czech converter has six of such circuits. In the less important places the 317 IC were used. So in total there are eight power supply regulators.

The whole circuitry is made in the SMD technique, but in the outputs, in the parallel branches and the power supply of the IC’s we have very nice polypropylene capacitors from WIMA. Regardless from my howling, the device is very nicely crafted, and it can be seen, that it has been thought through in every direction.



CASEA
CEPHEUS

Price: 3500 PLN

Distribution: Casea

Contact:
tel.: +48 609 933 983

e-mail: polski@casea.eu


WWW: CASEA



HOME PAGE



© Copyright HIGH Fidelity 2007, Created by B