£270 soundcard vs £6,000 D/A converter

[ablaze 0]

http://www.avsforum.com/avs-vb/showthread....threadid=236215

 

High-End Analogue Soundcard Shootout: LynxTWO, RME DIGI96/8 PAD and Delta 1010 (LONG) Last weekend Dan (dannydlp) and I have completed the 3rd session of 2-channel analogue audio quality of comparisons. We were critically listening to:

Soundcard 1: RME DIGI96/8 PAD £270

Soundcard 2: M-Audio Delta1010, costs £400

Soundcard 3: LynxTWO model B, costs £900

Control point 1: dCS Delius 24/192 D/A Converter, costs £6,000

Control point 2: Lexicon MC-12B Music & Cinema Processor, costs £9,000

 

* Prices are recommended retail prices in UK inclusive of VAT (1£ approx. 1.5$).

We started this exercise mainly to satisfy our own curiosity in our quest for the ultimate sound quality. We also felt the need to address the question of relative improvement in sound quality that would be afforded by using various (very) high-end sound cards. These cards are frequently mentioned in the forum but there is very little first hand experience of their use in domestic audiophile applications.

Dannydlp decided to invest into LynxTWO mostly based on his expectations after reading various reviews in professional audio arena. For example well respected reviewer Martin Walker in March 2002 issues of Sound-on-Sound Magazine concludesThe Lynx Two proves that a soundcard with internal converters can provide superlative audio quality on a par with many stand-alone converter units if well designed with absolutely no compromises, and I suspect it might give world-renowned product ranges like those of Apogee a run for their money. It's certainly the best-sounding soundcard I've ever reviewedâ€.

 

Many other reports can be found on the Internet that all come to a similar conclusion: LynxTWO is currently the soundcard of ultimate quality both in digital and analogue domain.

The RME DIGI96/8 PAD card has been discussed in this forum more often and recently I performed a test of its SPDIF output quality. The card performed very favorably against dedicated and significantly more expensive CD/DVD transport. At the time the analogue out quality was not considered. However, in February 2002 Martin Walker reviewed RME DIGI96/8 PAD analogue quality stating “Unusually, while the A-D converters are from AKM, and employ the AK5383 chip also found in M Audio's flagship Delta 1010 model, the D-A converters are from a completely different company: the Analog Devices AD1852 chip. Listening tests against my new benchmark Echo Mia proved that these D-A converters were well up to scratch. There were subtle differences between the two, with the Analog Devices AD1852 of the RME card having a marginally sweeter, more natural, and open sound than the AK4528 of the Mia, as well as slightly better stereo imaging. However, on some material I preferred my Mia, so it was a close-run race.â€

 

M-Audio Delta1010 has been much more frequently mentioned in the forum with extensive and authoritative review published by Bill Gaw2 on the analogue out quality of M-Audio Delta1010. Very recently Bill posted following comment:“I have had multiple high end DACS, CD transports, D/A converters, DVD-A and SACD players in my system over the years, have a Walker Proscenium turntable with Kondo IO-J cartridge for analog playback, and, believe it or not, prefer my HTPC with M-Audio 1010 card using Music Center with its DSP and ASIO for playbackâ€

 

Bill Gaw2 credibly establishes that an HTPC configured with good professional sound card can become an equal and valuable component of a high-end audio system.

dCS Delius DAC is high-end of high-end. We are not going to evaluate much about it in this thread but I encourage forum members to check the web for references to this fine unit. Delius was connected to LynxTWO card via SPDIF and had separate connection to card’s master clock so that the jitter was completely eliminated.

For this test we had two PC set-up as follows: Dan’s Windows XP based Audio PC with Delta1010 and LynxTWO cards, both with ASIO drivers. Player was 4am Music console with ASIO plug-in, music files were all uncompressed WAV ripped with Exact Audio Copy. I contributed Carillon Audio PC (Windows XP) with RME DIGI96/8 PAD (driver bypasses kmixer so ASIO was unnecessary). Player was Sonic Foundry Siren (currently reincarnated as Sony Screenblast SIREN), music files were encoded in Perfect Clarity Audio loss-less compression. The rest of audio set up was Spectral DMC12 preamplfier,

Spectral DMA50 power amplifier, Infinity Intermezzo 4.1 speakers, MIT interconnects where possible, and Spectral MIT speaker cables.

Music used was:Antonio Forcione/Live! Track 1

Jane Monheit/Come Dream With Me - Track 1

DiMeola, McLauglin, DeLucia/Friday Night in San Francisco, Track 3

Norah Jones/Come Fly with me - Track 3

O Brother Where Art Thou soundtrack - Track 4

Yo-Yo Ma/Classic Yo-yo - Track 15

 

Several tests have been compiled:LynxTWO-B analogue out vs. RME DIGI96/8 PAD analogue out. Started with four persons present. It took about an hour to get our ears tuned, as the differences were not so obvious to spot. The sound coming from both sources was absolutely trilling and definitely high-end. After a while we have started to notice minute difference and agreed our findings: we all give slight preference to the analogue out of RME DIGI96/8 PAD. The sound had more air and the clarity. The bass was more natural to occasional rumbling-like sound of LynxTWO-B.

 

M-Audio Delta 1010 analogue out vs. RME DIGI96/8 PAD analogue out. This was a much easier task as by then our ears were much sharper. The difference was more obvious and we quickly agreed that M-Audio Delta 1010 is completely outclassed. In comparison with either RME or LynxTWO-B, Delta 1010 sounded as if the speakers were covered with a towel. The precision of overall audio image was noticeably reduced. We all readily agreed that there was not reason to perform additional LynxTWO-B vs. Delta 1010 test and have taken Delta 1010 out of subsequent listening tests.

 

LynxTWO analogue out vs. dCS Delius DAC. This was a difficult test and we think reveals the essence of professional vs. consumer high-end audio. From the 1st test we got accustomed to sound of LynxTWO, its precision, openness and detail. Initial listening to Delius appeared as less airy with possible muting of the higher frequencies, but that was offset with more melodic and overall more pleasing music. In number of tracks when listened via LynxTWO we can quite clearly hear various background noises produced by musicians (their breathing, lips popping, shoe tapping) as well as hiss present in the original recording, when the same track was played through Delius much of this side noise was pushed in deeper background bringing the listening focus squarely on music alone. This would occasionally have some unwanted side effect by de-emphasizing subtle but intended sounds. The explanation for this is probably the additional digital signal filtering and background noise elimination performed by Delius. While LynxTWO would present the digital recording exactly (with all its imperfections), Delius would (attempt to) present it as it was intended to sound. The difference goes to the philosophical meaning of High Fidelity/Audiophile reproduction of music: is it supposed to reproduce, with minimal distortion a recording; or is it supposed to recreate the spirit of live event even if in a process it has to deviate from exact recording.

 

LynxTWO-B analogue out, RME DIGI96/8 PAD analogue out cards and dCS Delius DAC. This was a revealing test that turned very favourable response to the RME audio qualities. RME would sound something in between LynxTWO and Delius. Not a completely open on the lower end of the scale as LynxTWO (to the point that this was overbearing), but not having the smoothness/musicality of Delius for acoustic music. For number of tracks we all preferred RME sound, in many cases vote would be split between RME and Delius, and quite infrequently we would vote to LynxTWO.

 

RME DIGI96/8 PAD analogue out, dCS Delius, and Lexicon MC-12B. In this test, both dCS Delius and Lexicon MC-12B were connected to digital-out coming from the RME card. Analogue out from the RME card and Delius was routed via Lexicon in pure bypass mode. All tests were conducted in pure 2-channel stereo. Speakers in this case were M&K MPS-2510P and MPS-5310 (subwoofer). Between Dan and me we were doing A/B/C test in turn, so that the other person didn’t know what is being selected. This was a particularly difficult test as the difference between the sources was very minute. This was surprising considering the jump in the retail prices of the equipment used: RME approx £270, Delius £6,000, and Lexicon £9,000. After number of tests the pattern started to emerge, where RME and Delius were preferred in approximately 40% of time each, with Lexicon coming as a winner only 20% of time. For the acoustic rock/country music we usually preferred Delius sound, classical music generally went to RME and vocal was split between Lexicon and Delius.

 

The results of these tests left us with some further consolidation work, as the outcome was contrary to what we had expected. We assume that the quality of the other components used is sufficiently high-end and neutral as not to particularly favour one source over other. In retrospect I believe, we should have included a really good CD player in this comparison, but we could not find any available at the time. We leave this for future.

We collated my impressions over several days before writing our final thoughts. We believe LynxTWO-B and Lexicon MC-12B had the most precise reproduction of recorded sound. With ‘precision’, we mean that we were able to analytically explore exactly what was recorded and stored on a CD-ROM (its WAV or PCA equivalent to be more precise) including the recording mistakes, unintended sounds and background noise. This maybe the ultimate goal in the purest sense of high-end High Fidelity, but is it always beneficial to hear the recording in all its (un) intended purity (unless you are duty bound like a studio technician)? We quickly discovered that many CDs in our collection were poorly recorded. Listening them with high quality (transparent) components may result in either off-putting or unpleasant sensation.

With dCS Delius we enter the world of audiophile reproduction of recorded music. In this world the transparency is not the objective as we recognise that the recording is only imperfect imprint of the real event. Delius attempts to correct for the imperfection of recording process and deliver the intended result rather than authentic and uncoloured reproduction on what is actually stored on CD-ROM. This we hear as musicality and smoothness of delivery and in most cases prefer over the rough, but accurate, sound of pure high fidelity. Ultimately, Delius must make certain assumption when it attempts to recreate the actual musical event, as the entire information space is not present in the recording (either due to the limitation of the recording technology or the skill set of the personnel involved). I most cases Delius fills the gaps so that the final outcome is an improvement over the pure high fidelity reproduction, but the success rate falls short of 100%.

Delius is not a pure high fidelity piece of equipment in a sense of LynxTWO or Lexicon MC-12B. The sound as it comes from the CD-ROM is processed (some would even used the word ‘distorted’) in order to achieve the audiophile quality with beneficial results. We expect that as we move towards higher resolution recordings or if analogue sources are used, the benefits of sophisticated digital sound processing would become less important.

Where does than RME DIGI96/8 PAD falls in? Obviously, the physical qualities of on-board DACs are sufficiently high to comfortably stay in this distinguished company. It proved itself as very competent and formidable alternative. We believe that it was designed with purity and transparency in mind, however, obviously using cheaper and less rigorous principles than LynxTWO. This is visible when comparing S/N measurements, RME's are at least 6dB lower than LynxTWO. However, what RME looses in pure high fidelity turns out to majorly benefit its audiophile quality. This was maybe due a cleaver choice of the card’s components at the design stage or even some element of luck, as its the distortion of the original signal adds to overall musicality of RME.

Considering the massive price differential of the components involved, our conclusion is that, for pure 2-channel listening, RME DIGI96/8 PAD presents unbeatable value for money. It is a reference piece of equipment that should be the staring point of any high-end system.

Those who prefer purity and high fidelity may consider upgrading to LynxTWO as the costs is not overly taxing (3 times the cost of RME). On the other hand improving upon audiophile qualities of RME card would require almost 20 times cost multiple.

We hope that for those of us involved in HTPC there would be other alternatives and cheaper routes to audiophile land in a form of various plug-ins that combined with absolute high-fidelity of LynxTWO would introduce digital signal processing that would improve the musicality and make the imperfections of the digital recording medium less obvious. Moves towards the high resolution recording media in a form of SACD and DVD-Audio would offer further improvements without additional signal processing. All of that is likely to make reasonably priced HTPC an extremely competent part of high-end audio systems.

Branxx & Dannydlp

 

[Northern Oak 0]

hmmm... are you very much into high-end soundcard? PC-reliance, right?

[ablaze 0]

yes, I'm PC-centric. was it that obvious?

[Northern Oak 0]

[ablaze 0]

hehe..yeah I'm just one of those people who believe that a good enough sound card will easily rival expensive CDPs

[Northern Oak 0]

hehe..yeah I'm just one of those people who believe that a good enough CDP will easily rival expensive soundcards

[ablaze 0]

haha..well to each his own