Anyone who has done their homework on nice soundcards knows how the sites rate them. Most sites I have seen have the Lynx TWO in a class of its own, and then all the other “Class A” soundcards are in their respective place. As far as I have seen, no one has done a real comparison of these in an audiophile review, with controlled conditions, and in a nice system.
One of our members offered me an audition of his RME HDSP 96/32 with no strings attached. So I accepted, and am pleased to bring you the first real comparison of two powerhouse soundcards. The only catch is, the RME has been modded. The Empirical Audio mods have been performed on this card. I won’t bore you with the details. I am sure Steve Nugent will post the details here, or you can read what he does on his site.
I am comparing this to my Lynx L22. L22 is a two channel version of the Lynx TWO without the video sync and stuff that I don’t need. It’s a lot cheaper than the TWO models, coming in at $749 retail, the same as the unmodded HDSP 96/32. The mods are $650 from Empirical, bringing this card in at a whopping $1,399. My Lynx only has some simple mods which are removable. I cannot disclose what they are because I gave my word to someone who mods these for a living. I have obscured my minor mods from the picture with black spots, but they don’t make a huge difference anyway.
Let’s get on with it, shall we?
First, to describe the review system. I am using a Perreaux Radiance 200i with my Piega C3 Limiteds. My cabling is Audience Au24 balanced, Au24 speaker cable, jumpers, Electraglide Fat Boy 2000 MkII for the amp, and Audience PowerChord for the front end. I have two dedicated circuits with a Wattgate 381 on the analog, and a Hubbell Audio Grade on the digital. My room, as you can see, is fully treated using Eighth Nerve seams, corners, and frames. The PC case is lined with Dynamat and ERS all over. ERS is also around the power supply. To make the review fair, I used the stock breakout cables with each card, although I do have a nice custom breakout cable from Scott Nixon for my Lynx.
A pic of the review setup:
The room treatments:
I used an SPL meter to level match the two cards at 85dB while listening, although this was not necessary b/c the cards both running at -10dBV had the same output, even as measured with the meter. I auditioned the Lynx first, and then the same tracks in reverse on the RME. After letting the RME warm up, I did the test again in reverse and went back to the Lynx.
Lynx settings:
RME settings:
My tracks for this test were as follows:
-Steve Kimock Band – 11/14/2004 – Track 3 - Storm Crow (I was at this show a few weeks ago so it’s a good reference for me)
-Jerry Garcia/David Grisman/Tony Rice – The Pizza Tapes – Track 8 – Summertime
-Leonard Cohen – The Essential Leonard Cohen – Disc 1 - Track 8 – Famous Blue Raincoat
-Pearl Jam – No Code – Track 6 – Off He GoesThe first track up is Steve Kimock. This track sounds pretty close to what I heard live. Carl’s souped up AKSA amp sounds more like exactly what I heard, but this is pretty close, and is a bit softer. I heard more detail on this track with the Lynx card than with the RME. This is the first thing I noticed.
The RME did really exceed my expectations right off the bat, and is a really good sounding card. It has ample bass, although it’s a slight bit lighter than the bass coming from the Lynx card on this track. Another thing that comes to my attention during this track is that the RME seems to lack a certain degree of richness versus the Lynx card, but more on that later in tracks that really display that to a greater degree.
I have always liked The Pizza Tapes, and have listened to it thousands of times. It seems to work well for auditioning, as acoustic guitars are very hard to reproduce faithfully, and it’s very hard for most systems to expose subtle harmonic differences between two different brands or models of acoustics. This is a great recording for testing that, and it was evident that the Lynx more readily exposed the harmonics of these two guitars, and the subtleties of the players’ styles.
By now you’re probably thinking the RME is not a good sounding card. That’s just not the case. It does sound great, and it sounds better than about any stand alone player you can pick up in the $3K range. But the Lynx seems to be making quick work of this heavyweight soundcard in all its modded glory.
An interesting noteworthy fact here is that I never got the HDSP working in ASIO, so I changed to wave out for both cards to make it fair. There are subtle differences between ASIO and wave out for the Lynx, and I assume the HDSP shows these differences as well, but both were in wave out for the test.
Another thing I found out about these cards is that I think the HDSP does the -10dB thing digitally, whereas the Lynx does it mechanically somehow. I guess the Lynx uses transformers or something b/c there is an audible click when you switch to -10. This could have a major impact on this test, but my amplifier does NOT like +4dB with balanced cables. Too much gain, and lots of crazy things happening at +4dB. But I guess you can’t blame RME for this. They aren’t making soundcards for US. They are making them for the pros.
Moving right along, on Famous Blue Raincoat, there are these wind chimes at the beginning in the background. With the RME, these are quite obscured, and it seemed as if there were actually LESS individual chimes as well. This was strange because with the Lynx, it was easy to hear the individual chimes and the wind that moved them.
The natural decay on Cohen’s guitar is not really there as much with the RME card. There is a subtle buzz on Cohen’s E string that is also very easy to notice with the Lynx, and quite obscured by the RME. I could barely hear the backup vocals at the beginning on the RME, but the Lynx was excellent at this. The RME was a bit boring compared to the Lynx on this track. It just wasn’t exciting to listen to as the Lynx was. There was a bit less noise on this track with the RME, which I think is a function of the overall veil that seems to cover the subtle nuances. This is an old track with lots of tape noise, and it’s evident on the Lynx. What’s funny though is that there is a bit MORE noise on the Pizza Tapes track. It’s very subtle, but the noise level is raised on this track, and lessened on tracks where there really is noise. Interesting.
The RME has a wide soundstage. It’s probably as wide as with the Lynx, but it’s not nearly as deep.
On this Pearl Jam track, which is really a super track that Carl turned me onto, there are a couple of acoustic guitars, and an electric also. It’s a very good recording, and a great song to boot. The background in this song is much blacker with the Lynx. It’s a very quiet recording at parts, and it’s easy to hear the noise differences between the two cards.
There is a guitar on this track which is on the right, and it’s always crystal clear, and tonally it jumps out of the mix a bit. With the RME, it didn’t. It was still there, but it just wasn’t what I would call crystalline, and the harmonics were missing. This was the biggest difference I could hear so far between the two cards. Again the RME’s excellent bass came through, but it didn’t match up with the Lynx. The whole song sounded muffled and unexciting on the RME.
I am sure that the RME would probably perform better in a different configuration. Perhaps if you were running +4dB straight to an amplifier, or a passive, it would be closer to the Lynx. But I just don’t see how it could make up 150% difference, which it would have to for the price.
There seems to be a wool blanket in front of the listener when the RME is playing. It’s not a very thick blanket, but it’s dark grey wool, and it’s probably itchy too.
Nothing you’d want between you and the artists, especially when this card retails for the same money as the Lynx with no blanket.
The Lynx L22 truly is in a class by itself with it’s Lynx TWO brethren. It is an exciting card to listen to, and I believe it has few peers among the best digital I have ever heard at any price. I can highly recommend the Lynx to anyone, and I would be willing to throw mine up against any front end you’d want to compare it with. The RME would be a great value at $350. $750 is a hard pill to swallow, and $1,399. . Well. . That’s just. . Ohh nevermind.
Talk amongst yourselves. . .
