Published on: May 1. 2012, No. 96
The present review is primarily focused on the hiFace EVO digital-to-digital converter (digital interface), which converts USB signal from the computer into one of the five types of input signal “understandable” to a classic D/A converter: RCA, BNC, TosLink, I2S, AES/EBU ST. Its operation can be further improved by applying the EVO Supply, an external battery power supply and plugging, through the BNC input, the EVO Clock, an external precision master clock. We will also test review these two devices.
Auditions, tests, reviews, meetings with computer-related products, audio files, especially high-resolution, still resemble walking through a minefield. It would seem that we already mastered some basic skills and that we know quite a lot and learn quickly. It is common knowledge that one of the worst enemies of the digital signal is jitter, i.e. the lack of precision in the signal timing. Less common is awareness of various types of jitter - correlated and uncorrelated, short- and long-term, and their combinations. But let’s assume that the awareness of this problem is quite satisfactory.
However, when it comes to computers, we are usually helpless as babies, because it is the computer geeks, software engineers, IT specialists and other trustworthy, respected people (I say this with absolute conviction) whose voice is most important and significant here. And what they say is that in the computer realm the vast majority of our concerns about digital audio systems is not applicable.
And it is difficult to discuss with them at the meta-level, because according to the theory that is commonly accepted both in academic circles (in terms of pure theory) and engineering (let’s called it applied theory), the computer is an ideal sound source, and USB is the best possible interface. Computing capabilities seem almost endless here, at least in the context of the CD player, and USB transmission bandwidth “precludes” any problems. This, they say, is the sound source of the “pure” form, the perfect source.
Except that in this case theory does not meet practice, at least not in perfectionist terms that we deal with in the audio industry. Because as a result of our sensory experiments we – audiophiles, in other words thinking and practicing scientists, engineers and end users – managed to identify some areas where there seems to be a need to verify the computer world operating theory. I'm not saying here that it needs to be overthrow – absolutely not! I mean something that is normal in every field of science, namely innovation and understanding shifts that do not change basic assumptions, yet significantly affect the results.
What we already know is that the computer is one of the most difficult to master digital signal sources (players). Its computing power is actually the least important component, and what gains importance are methods of internal signal transmission, its treatment, management, power management, etc. The problems are many.
It can be plainly seen in case of the software player JPLAY. The program beats world records of popularity (in audio circles), wins multiple awards, is warmly discussed by the biggest audio magazine, etc. How did it happen that two enthusiasts - Marcin Ostapowicz and Josef Piri – came up with something that easily beats products from well known, recognized software companies, developing this type of software player for many years? The answer is very simple: they questioned everything that appeared to be certain and one by one addressed the sources of problems. And these at first often seemed illogical and without sense. And yet, their elimination has changed sound for the better. It was a simply imagination exercise.
The computer is one thing, and the other is what happens with the signal once it leaves the computer. Now it’s a different game. The computer guys have less to say here; it’s the engineers that have their say. However, as usual, most of them are mistaken, resting on theoretical knowledge and their beliefs. I am almost sure that none of the critics of the solutions adopted in the audio world, repeatedly re-examined, discarded, confirmed, and challenged again, did not carry out such tests himself. And in reality he is either misinformed or - even worse - ignorant. Grrr...
As it has long been known, USB is a very primitive way of audio transmission, very difficult to master. USB cables can dramatically change the sound. The way how DAC receives the USB signal (adaptive or asynchronous) defines the obtained results. And, finally, USB devices abide by exactly the same rules as any other digital audio device. That includes jitter as well as power supply quality.
And so we come to the subject of this review. The Italian company M2TECH have been offering for several years their own solutions dedicated to both USB to S/PDIF signal conversion (the popular hiFace) and its decoding (D/A). This time we will have a look at the hiFace EVO D/D converter, and certain solutions that can improve its sound.
So far we reviewed:
- M2TECH Young converter, review HERE
Recordings used during listening sessions (selection):
- Audiofeels, Sounds of Silence [z:] Audiofeels, Uncovered, Penguin Records, 5865033, rip z CD, WAV.
- David Sylvian, World Citizen (I Won't Be Disappointed) + Angels [z:] David Sylvian, Sleepwalkers, P-Vine Records, PVCP-8790, WAV, rip z CD.
- Kankawa, Dear Myself [z:] Kankawa, Organist, T-TOC Records, UMVD-0001-0004, Ultimate Master Vinyl, 24/192 WAV; review HERE.
- Keith Jarrett, January 24 1975. Part I [z:] Keith Jarrett, Köln Concert, WAV 24/96, HDTracks..
- Nosowska, Kto? [z:] Nosowska, 8, Supersam Music, SM 01, WAV, rip z CD; review HERE.
- Pieter Nooten & Michael Brook, Searching [z:] Pieter Nooten & Michael Brook, Sleeps With The Fishes, 4AD, GAD 710 CD, WAV, rip z CD.
- Sonny Rollins Tenor Madness [z:] Sonny Rollins, Tenor Madness, WAV 24/96, HDTracks..
- Stan Getz & Joao Gilberto, Corcovado (Quiet Nights Of Quiet Stars) [z:] Stan Getz & Joao Gilberto, Getz/Gilberto, WAV 24/96, HDTracks.
- The Alan Parsons Project, Sirius + Eye In The Sky [z:] The Alan Parsons Project, Eye In The Sky, WAV 24/192, rip z DVD-A.
Japanese versions available from CD Japan
As you can read in the section on test review methodology, i.e. the description of the method of its preparation and implementation, I mainly compared the Italian converter against the very successful, in my opinion, asynchronous 24/192 input of the Wyred4Sound DAC2 D/A converter. As you can read in this review, the solutions adopted by the W4S people proved so successful that the said device input has served me a long time as a reference point for other USB-S/PDIF converters and USB “DACs”.
What happened then when I plugged the hiFace converter into DAC2 RCA S/PDIF input? I will come back to this but first I need to return one more time, for a moment, to a more general description.
My impression is that over the years of testing CD players, SACD, audio files, I can see changes introduced by successive improvements, new technologies that form a fairly clear pattern. The biggest changes, system changes, affecting the very basis of sound, its structure, have been effected by jitter limiting procedures. The tone of well clocked device is deeper, warmer, fleshier, better differentiated, and just more natural. It’s amazing how dark such tone can be. Until the first cymbal crash, until we hear an instrument with high treble energy, and also a lot of harmonics. Then we get dense treble, correct cymbals’ body and weight, etc. Devices that are worse clocked, with higher jitter, sounds brighter, seemingly more detailed, but only till we compare it with something better. Then a few bars of a music track is enough to spit out that sound and get on with the new.
A slightly different change in sound can be brought through power supply improvement. One method is to give up the 230 V/50 Hz mains and use battery. The latter is not without its flaws, but in specific applications it can do much good. We can expect most improvements with micro-signals, resulting in better retrieval of space around acoustic instruments including the acoustic character of the recording venue. The sound is usually a little bit further away from us, but the reason for that is a better presentation of proportions between the instruments, acoustics, and their improved relationship with the listening room.
So much for the introduction – so equipped, we will not have to get into details, and the review can be shorter.
Hardball: on its own, hiFace EVO converter sounds better than DAC2’s USB input. The tone is deeper, fuller, with much higher energy; it is more tangible and more complete. It was not bad before, but now, in my opinion, the sound from the computer may be better than the digital output of a good CD player (with 16/44.1 CD rips), or than the digital output of file players (24 bit/ 96-192 kHz files). Maybe not better than the top devices, but outside of the elite and down the price list – it can be very competitive. I’m talking here about something that was once associated only with the very best CD players - coherence, depth of sound. Nowadays, we pay less attention to tonal balance, amount of detail, and other components of sound reproduction, and concentrate more on the music, how it affects us, how it evokes emotions in us. Such focus shift from technology to music is a good thing.
Without thinking twice I can say that hiFace EVO is the best USB-S/PDIF converter (I have not tested its AES/EBU and I2S outputs) that I have ever had in my system. I think that it is one of the best (but that's just my assumption) USB receivers I know of. Actually it doesn’t put a foot wrong; it does not reduce the dynamics (the most common problem), does not sound anemic (equally common), or simply boring and nondescript. Perfectly captures the sound colors, shades of emotional energy.
It won my heart especially with its quality of sound with CD rips. It began with the deep voice of Nosowska in her song Kto. The track was clearly recorded very close, without any anti-pop mic shield, reducing “popping” sounds, i.e. sounds occurring particularly in the pronunciation of the letter ‘p’. This is why we can clearly hear the ‘p’ on this track.
This change has a certain aesthetic and artistic value, and translates into a very intimate, almost “gut” vocal. In terms of soundscape it means deep, soft, quick bass attack.
Few speakers can do it well (and even less amps), usually distorting it or not showing it at all. But it is audio sources that have the biggest problem with this - I do not know why. W4S DAC handles it very well – it has a very powerful and deep bass, and can show this effect on its own. However, the M2TECH converter in full “gear”, i.e. with the external power supply and external clock sounded much better than the CD played on my Ancient Audio CD player, and hence the Philips CD Pro-2LF transport. It was a real experience, because the true meaning of this song is based precisely on the titular, very intimate question. The W4S USB input, although really good, seemed quieter (even with level correction), less involving. Everything was there, but it did not evoke in me the kind of emotions that the Italian converter did.
Hi-res material was similar. What got on my nerves a little was the need to switch between the 44.1 kHz and 48 kHz, because I did not listen to entire albums, but rather groups of a several tracks with different sampling rates. When you listen to the entire album from the beginning to the end it is no problem.
In any case, files with a high sampling rate also showed what I already mentioned discussing the CD and the album 8 as an exemplum. However, in this case I did not instantaneously fall in love; I had to listen a little more for it to happen.
What caught my attention in particular was a more dynamic sound than the USB input on W4S DAC. But there was a transition from ppp to fff; that is not the point. It was something intuitive that came out in every second of the recording, but as something secondary, as a vehicle for the transmission of emotion. While with the CD material it was evident and resulted in my faster heartbeat, hi-res material showed that the former was a little overdone and "for show"; somewhat unpolished. But this is not a problem with the converter – it is a problem of the CD format. Anyone who has heard a good vinyl, or well-played hi-res files or SACD knows what I mean. M2TECH just made this evident. And it did it without “demolishing” the recording, as often happens with expensive "high-end" devices. The hiFace EVO converter saturates CD material with emotions, while showing its disadvantages. It was really something special.
Yes, it is an excellent converter. By itself, with no external clock and no external battery power supply it sounds very well and is easily comparable to the most expensive USB-S/PDIF converters. It may not always win, but will not always lose either.
Adding the battery power supply changes the way it reproduces the acoustics, space, drawing shapes. The sound is a bit further away; it has a more natural perspective, better extracted end reverbs, flavours, etc. It is not better or more "detailed", it may even seem that the mains brings more details; there is just more "wealth". It's like the Harbeth speakers – to the people used to the "chirping" tweeters they may seem dark, less detailed, etc. For all those who are cured of this - of course in my opinion – “artifacts hyper-detailness” they are simply natural, and everything else is warped. So it is with the hiFace EVO, and adding better a power supply simply highlights this difference.
The addition of an external clock leads to the depth of sound that I appreciate so much, to its mature character. This is a very sophisticated change that can elude those who are not yet familiar with high-end audio. Not that they will not hear it, as I think that the changes are very clear, but because at some basic level of familiarity with high-end such changes are not particularly highly valued. But if we know our system well, and have a good comparative scale, this change will polish off what the external power supply and the converter sanded down.
It should be noted that, as always, all the small improvements accumulate into something much larger. And I have a feeling that this accumulation is of greater value than that effected by other components, such as anti-vibration accessories, power conditioners, etc. Something like the whole is greater than the sum of its components. Although it is BS in terms of maths, in the audio realm it is possible.
HiFace EVO converter was connected to the Wyred4Sound DAC2 digital analog converter. The former was compared to the USB port of the latter, as well as to the USB-S/PDIF Haliade Bridge (24/96) and the KingRex UD384 + U Power (32/384) converters. The test had a character of the AB comparison, with the A and B known, with 2 min long samples of music. Listening sessions also included whole albums.
Signal source was a dv7 HP Pavilion laptop with a dual-core processor, 320 HDD, 2GB RAM, Windows Vista operating system and software audio players - foobar2000 and JPLAY. Signal from the computer was sent with the Acoustic Revive usb-5.0pl USB cable.
To connect the converter with the clock and with the "DAC" I used the Oyaide DR-510 digital cable (BNC-BNC and RCA-RCA, respectively).
All three devices share the housings of the same size, made of aluminum, that only differ from each other in their front and back panels. This significantly reduces manufacturing costs. The devices are very small and light - their dimensions are 105 x 46 x 104 mm (WxHxD). And here's a problem - they cannot be placed in such a way that they would not move – any stiffer cable, such as the Oyaide, will move and turn them. I think that it would be useful to think of a some kind of "rack" for all three, holding them together, giving them more weight, and sitting on good anti-vibration feet. Made of wood (like the old Marantz)? Or metal? I do not know; what matters is that it should be functional, should have a system of cable routing. And looking good. All reviewed devices have been made in Italy.
hiFace EVO is an asynchronous, 24/192 digital-to-digital converter (D/D), USB-S/PDIF (ST, I2S, AES / EBU). Its full proper name is 'HiFace EVO 192kHz 24 bit Digital Audio Interface." In short we can say that it converts the digital signal fed from the USB to a signal "understandable" by classic digital-analog converters (D/A).
It features three types of S/PDIF outputs - electrical RCA and BNC and optical TOSLINK. The front panel sports four electrical outputs - I2S, AES/EBU, RCA, and BNC - with optical outputs (TOSLINK and ST) at the back. On the back we also find a USB 2.0 input. The company offers a suitable driver on their website. Next to the input USB socket there is a green LED that shows synchronization status between the converter and the computer.
But there are two other inputs - BNC for external clock, also with a diode, and a socket for an external 7-11 V DC power supply, with another diode.
The whole system is mounted on a single PCB. In its center is a Xilinx Spartan XC3S50A DSP which houses the software implemented converter. The same IC is used in the company’s full D/A Young converter.
All electrical outputs and the clock input feature isolation transformers for galvanic isolation of the PCB. Optical outputs are isolated by definition. Next to the DSP we can see two nice clocks - separate for 44.1 kHz and 48 kHz.
Power supple socket is equipped with an active low-pass filter that improves the quality of power supply.
EVO Supply is a Battery Power supply for the whole EVO family of products from M2TECH, but not only for them. It is based on Li-Ion 2200 mAh batteries (minimum). The power supply provides current to 1 A (peak) and 500 mA (constant) at 9 V (low-noise power supply) or 9.5 V - 11 V (rechargeable).
Its front panel is empty. On the back we have a socket for an external 15 VDC power supply and two voltage outputs. They are really very, very nice. Next to them we find two switches: 'battery/Charge' with a diode and 'On/Off'. The first selects the operating mode – we can either work only on batteries, with the mains turned off, or with an automatic power supply. In the latter case the batteries are automatically recharged if needed. After charging, the power supply is disconnected. Together with the power supply we get one charging cable. Its plug on the power supply side (5.5 X 2.1 mm) is excellent, but the cable itself - rubbish.
Inside the case we have three cells of the said batteries and the accompanying PSU. The outputs for external devices feature some electronics. Unfortunately, I do not know what they are for...
EVO Clock is a precision, external master clock, based on low phase noise and very accurate TCXO oscillators (temperature compensated crystal oscillators), with a special temperature compensating circuit. As we read in the manual, clock stability can be improved even further, using OCXO (oven compensated crystal oscillators), with its own "heater" that keeps the oscillator’s temperature constant, but they are very expensive.
EVO Clock is used primarily to controlling the hiFace EVO converter, but not exclusively - it's very versatile. Using it with the M2TECH converter we use two oscillator frequencies: 22.5792 MHz for files with the sampling rates of 44.1, 88.2 and 176.4 kHz and 24.576 MHz for 48, 96 and 192 kHz. A small switch is used to choose between them. I'd rather have it done automatically, but no problem. EVO clock provides a 3.3 Vpp at 75 Ω signal at the output.
However, you can use this clock for many other devices. Hence the large knob in the centre of the front panel, with which we choose the sampling frequency between 44.1 kHz up to 384 kHz with a "super clock" option of 128x and 256x. The output signal may have the valueof 3.3 Vpp at 75 Ω or Lo-Z (for high-impedance inputs).
At the back of the unit we have two BNC connectors - one for the MCLCK output (EVO devices) and the other one WCLCK (other devices). Unfortunately, I do not know whether the device can send two signals at the same time, allowing a common clock for the converter and the DAC. There is also a socket for a 9 VDC external power supply (with a LED).
The system is assembled on a single PCB. In its center there are two large, nice clocks - one for the 44.1 kHz and its family and the other one for 48 kHz. Next to them we have a large Xilinkx DSP, with software implemented PLL loops to control the desired clock frequency. BNC outputs feature matching transformers, for galvanic system isolation. The power supply socket is equipped with an active low-pass filter circuit, minimizing noise.