Hi everyone,

I'm new to head-fi although I've been lurking for a while.
My quest is to build the ultimate computer-based transport solution.
I own a seriously modded Cambridge player hooked up to an external
battery-powered PCM1704 DAC with a built in low-jitter master clock.
A few months ago I purchased a Squeezebox 3 along with
a Hagtech USB/SPDIF converter just being able to compare what works
better. I must say that sound-wise my CD player as transports
smokes both the SB3 and the USB converter solution, although not
by much. I love the convenience of computer audio (creating playlists
and playing with future room correction plugins...) but I am almost ready
to go back to CD playback due to higher fidelity. Yes, I'm quite a demanding
audiophile. What I'm looking for is a computer audio solution that
equals or bests some of the most expensive cd transports available
(I'm talking about transports in the $5.000-15.000 range)
As a last resort I'm willing to give an internal PC soundcard with I2S
and wordclock inputs a try. Has anyone ever tried this? Shouldn't this
be considered the holy grail from a technical standpoint?
Any other ideas or suggestions wheter it is at all possible to build the
ultimate computer playback system?

Thanks,
Peter

well i can hardly comment on the topic but...perhaps if you want to take out a small mortage get a lynx two + wadia w9series decoding computer thingamabob

Quote:



Contradictory. You can't smoke something by a small margin.

You are not the only one in that quest Peter, you can count me in and probably many many others, I know philodox for example that is also in that quest.

There seems to be a little something that high end CD players do as transport that PC audio is not able to reproduce. I've come to the same conclusion as you with my Arcam CDP and Philodox also with his Eastsound.

There are probably more and more solutions that will come up in the next few months/years. Here is a list of things that are announced or already on the market for you to check out:

- Naim has announced a series of products based on a music server.
- The Olive music servers are supposed to have an incredible sound.
- The upcoming BelCanto DAC3 that will have a USB input.
- Empirical Audio has some outrageously expensive USB to spdif converters that are supposed to be incredible sounding.
- You could have your SB3 modded by Bolder or RWA, apparently it improves the digital out dramatically.

Hope this helps. If you try one of these out, please post your results, I'm very interested. I almost bought a Squeezebox a few weeks ago but can't live with the fact that my CDP as transport will have better sound quality.

Quote:



Is the hagtech bit-perfect.
The squeezebox should be pretty much near perfect.

Does your DAC have any dejitter circuit? In that case the sound will depend on the amount of jitter the digital source emits.

If your transport isn't garbling the bits that can be the only dfference in the digital domain.

Do you have your receiver isolated? Otherwise that could be another factor.

Cheers

Thomas

Yes, both HagUSB and SB3 are bit perfect-sources. That's apparently not the problem. My DAC uses a Crystal ASRC and therefore should have quite effective jitter attenuation circuitry in place. However, my sources don't sound equal. My cd player usually wins but the SB3 is quite close. I'd say the USB/SPDIF converter is the weakest link and makes my music sound a little bit muddy and lifeless. I notice the difference when listening to vocal jazz recordings. Midrange is crystal-clear from the cd player but already less good with the SB3. I'd personally stay away from other USB converters as they probably utilize the same TI PCM27xx/PCM29xx chips. After studying its internals and transmission modes, concept of time-optimal PLL..., I am less convinced about USB as an ideal source for high-end audio. Modding the SB3 might be an option although some fellow audiophiles have reported that even after mods, it still doesn't equal the best cd transports out there. Knowing the stock SB3 sports very low jitter numbers - what else could be going on here? So this suggests there might be other unknown variables in the digital game . If jitter and/or lousy implementations of the interfaces cause such differences, I would think the ultimate transport must be a master/slave I2S connected system. I would love to know if anyone has ever tried a similar setup or even got reference quality out of their computers. I'd be extremely curious to see what happens if I slave both my cd transport and computer or SB3 to the DACs MCLK and use I2S. Shouldn't even the cheapest source (computer or SB3) provide identical results? That's a key question.
Sounds like an academic challenge but I really want to know what the cd player has that a computer cannot provide...

Regards,
Peter

This is interesting, since it all boils down to the best way to get a S/PDIF signal out of your computer. This is interesting to me because I use the relatively cheap M-Audio Transit to give me bit perfect audio, but your post has me wondering if the USB to S/PDIF conversion is a bottle kneck in my system.

Why not opt for a solution like the E-MU 1212M, which takes the USB conversion factor completely out of the picture. You go from the PCI-Bus right to high quality conversion chips and then straight to S/PDIF.

Is there a way to objectively measure the qualify of S/PDIF outputs?

There is one more effect we did not talk about yet.

Depending on how you drive your ASRC chip you get a lot of resampling going on just just by the frequency difference between the source and the local clock even if the input clock has 0 jitter. This problem will be true if you use a fixed clock on the local side of the ASRC chip.

The difference in sound could simply be different output frequencies on the player and the SB3.

If you can get your hand on a Lavry DAC it would be interesting to evaluate whether you get the same differences with a synchronous dejitter circuit.

I2S really brings not much benefit. You just get the clock on a separate cable but you are still following that clock.

The other way to do this is via a word clock link back to the source. Unless the source really screws up this will give you sound that has nothing to do with the source. Linn, Meitner, and DCS build players and DACs that work this way. Many sound cards have word clock inputs. Most combined A/D D/A converters also ouput a master clock and can be used in this mode.

I agree, I would stay away form USB for high quality for the time being.

Cheers

Thomas

Hhmph my reply was lost.

If you have a sound card with good analog input you could try RMAAing the three sources to see if the result is objectively different and if so which is better.

Apart from jitter, errors and interpolation in the CD player, and upsampling if you have that, may be affecting things.

You can do wordclocks with the right E-MU cards. It should allow good connection to be made over long distances.

i hate to sound like a this, but has anyone ever tried putting a WAV file and seeing how that compares to the FLAC.....

Quote:



They're the same...

You could try the Empirical Audio usb stuff, that might get better results. Or get your squeezebox modded by one of a number of people. If all else fails you could get the dac from lessloss.com and have it synced with a lynx card as it's detailed on their site.

I did a very careful comparison between the coax s/pdif output of the E-Mu 1212M and the M-Audio Audiophile USB (which has coax s/pdif output). I had been using the 1212M for quite awhile, and had become very comfortable using the patchmix software, and had a bad experience with the Audiophile USB (due to bad drivers) at one point, so I was somewhat biased in favor of the 1212M, but I decided to compare them again because new drivers became available for the Audiophile USB which were said to fix some of the problems I had with it in the past.

To my surprise, the Audiophile USB sounded better. I called my wife in to verify which sounded better (I just called them A and B...she really doesn't pay attention to my complicated audio setup) and she picked the Audiophile USB also. There are more details about my comparison in my responses to the thread about the new drivers here at head-fi.

So, in my experience, it is possible to get better sound (from a coax s/pdif output, at least) from a USB solution than from a reasonably high-end, carefully engineered PCI solution like the E-Mu 1212M.

Scrith.

Your DAC1 is somewhat of a special case. Since it drives the local side of the AD1986 you get various amounts of resamling going on simply by differences to the input clock speed.

The sound is jitter free but the bits still might be heavily munged due to the resampling to accomodate for the clock difference.

If your sample of an Audiophile USB happens by accident to be very close to the clock in your DAC1 it might sond much better.

A synchronous DAC won't show this behavior and might give the opposite result. The technology of the DAC matter for this.

Cheers

Thomas

Thomaspf,

I'm not sure I understand your explanation of why I appear to have better sound using USB output to an Audiophile USB rather than s/pdif from my 1212M (I am lucky that the clocks are in sync? Maybe this has something to do with them both being plugged into the same PS-Audio P-500?), but I'll take your word for it. By the way, I found that I could further improve the sound by resampling in Foobar2000 to 96K (using SRC). I attribute this to SRC probably doing better resampling than the DAC1 (which resamples all inputs to 110K, I've heard? If so, maybe it does a better job with a 96K than the 44K I was originally giving it?). The odd thing is I don't remember hearing any improvement by resampling to 96K when using the 1212M.