Does this dac exist?

Good question. I don't know the answer.

I2S is not an external standard primarily because most manufacturers that have added it (about 10) implement it differently, and usually poorly. The designers simply dont understand transmission-line effects and termination techniques. S/PDIF standard was created by a standards committee, mostly staffed by Sony and Philips.

It's a lot like Word-Clocks on transports and DAC's. Very few of them are implemented well and it can be hit and miss whether they work well or at all. You dont know what signal levels are expected and the termination is questionable or not there at all. There was never any committee to my knowledge to standardize word-clocks.

The reason for the variability in I2S, besides the fact that no manufacturer seems to want to be a follower, is that most of the signals in I2S are high-frequency with fast edge-rates. Almost 20 times higher frequency than S/PDIF in some cases. Guys that have been primarily designing analog and maybe copying a little digital are lacking in experience to deal with this. This is the reason that some designers denigrate I2S on the forums. They dont want anyone to know that they are not competent enough to pull it off, so they call it a bad thing.

For someone with a digital background and experience with high-speed signalling, this is like falling off a log.

Steve N.
Empirical Audio

But how come I2S is still not the digital standard for a CD transport and DAC interface

Mostly because it does not provide copy right protection, and other user programmable features.

As I mentioned, I2S is just one of the standards a DAC chip can take, such as MSB and LSB, but let's focus on I2S as you know it.

I2S contains 4 signal lines; data and 3 clocks. S/PDIF contains exactly the same unaltered signals, less the 4th line, the Master Clock. The real fault of S/PDIF comes from the PLL used to "recover" the Master Clock inside the S/PDIF receiver. This usually results in relatively very high jitter level and poor sound quality.

Providing a high-quality external Master Clock, or re-clocking solves all “issues” with S/PDIF. In my experience, such solution is not even a bit worse than I2S.

Best,
Alex Peychev

Bryon,
Back to your original post - If you want to use computer software, then you either have to
a) install a sound card with i2s output
b) use a USB to i2s converter - there are several of these, then feed the i2s signal to the DAC of your choice.

In my opinion it is better and more practical to place your computer next to the USB DAC (pref with async protocol like Empirical Audio's) and control the computer remotely.

I use a Mac mini, controlled via screen sharing from another computer, feeding a USB DAC, both are placed on my hifi rack.

Steve/Alex - Thanks for your observations. It always amazes me the extent to which technologies are determined by institutional momentum, short term market forces, lack of cooperation and planning, etc.. Computer audio hasn't quite emerged from the era of kludges. Having said that, I love the ease of use in my computer audio setup, and would never go back to discs. I just wish technologies would progress a little faster in this area.

03-17-11: Audioengr
You should definitely go for an adaptive USB interface with low-jitter clocks however.

Steve - Did you mean to say "asynchronous" USB here, or are you recommending adaptive USB?

Sirotseta - Thanks for your advice. Mac mini + USB dac close to the computer + screen sharing is how I'm leaning at the moment.

Do you happen to know if the Mac mini can be controlled by an iPad?

Bryon