Will computer to DAC replace transports and cdp's?

Alex P. wrote:
"The answer is NO, if superior audio quality is desired of course. Otherwise we might as well get I-PODs."

I cannot understand this statement. The fact is that computer-driven audio is intrinsically superior to reading of optical disks. The only way CD's can compete is if they are read into a FIFO buffer and then read out with a precision clock, which is essentially just equivalent to computer-driven anyway. I wrote this white paper on the subject for Positive-feedback:
http://www.positive-feedback.com/Issue22/nugent.htm

Even my own best modded transports playing rewritten low-jitter CD-R disks (similar to Reality Check) are no match for well-designed computer driven audio. The jitter can be extremely low using the computer, particularly with I2S interface or with the right clock topology. It is superior audio quality and it is available today, not in the future...

Steve N.

Alex - I agree that if the entire track is read into RAM and then played back from RAM, then it has the same flow-control and clocking advantages of a computer, but to my knowledge, none of the available CD players do this, except for maybe the Meridian, which is evidently just a CD-ROM drive and a computer anyway.

If you go to all this trouble, then why not just put the tracks on hard disk, where you have complete control over them in software, unlike the tracks on an optical disk?

Steve N.

Mburnstein - You dont want the computer to have the I2S interface. This would introduce computer clock, power supply and ground noise into the signals. The best way is to use an external WiFi or USB to I2S converter with it's own power supply, preferably battery. My company makes these. The result in unlike any digital audio you have heard. Eliminating the S/PDIF interface makes a huge improvement by reducing jitter to infinitesimal levels. Here is a link to a recent professional review of an I2S-based system:
http://www.americanwired.com/audio/empirical/offramp.html

Steve N.
Empirical Audio

Pardales - no this is not a good thing. The power is coming from the computer.

Power the Converter from battery or at least it's own power supply and you will realize a big improvement.

"Thanks, Audioengr. The U24 has no other power supply options."

If you open it up and cut the trace with +5V on it and inject your own 5V, then this will do the trick.

"It would seem to me that the future lies in USB DAC's, so that extra step of conversion can be skipped altogether. What do you think?"

Absolutely, although separates like USB to I2S and then I2S DAC input work just as well. USB DAC's, at least the good ones, use this technique. Skips the S/PDIF conversion and clock recovery. The USB DAC that I am designing is simply a USB to I2S converter and a DAC board in the same case. The DAC chips all need I2S or some variant anyway. You cannot skip that part. The best sound that I can offer is using I2S interfaces. 4 DAC's currently support it.

Steve N.
Empirical Audio
Manufacturer/Modder

"Steve, Can you make a I2Se interface with the Sonic Frontiers method of I2S connection as I have one of their Platinum 3.1 Sig DACS with I2Se interface. I know it is not compatable with the Muse Audio I2S interface"

Probably. My converter board is designed to go into virtually all DAC's. It depends on the DAC chip used in there. Do you know which one?

Steve N.

Alex - you bet. I can configure it for 128fs or 256fs.

Steve N.

The Burr Brown 1704-K chip uses left-Justified data which can be supported by the Off-Ramp I2S with some mods, however, the timing is unclear on how the data is multiplexed between the left and right channel chips. This could be a problem.

Steve N.

"there are problems transmitting separated Data and Clocks via single cable, especially in the case master clock is included."

What exactly are the problems? Seems to work perfectly and sounds significatnly better than any S/PDIF I have heard.

"The only "problem" with S/PDIF is the master clock recovery which is done from the Bitclock by multiplying it and PLL-ing it. This of course results in high jitter in the master clock. But since 44.1 (or multiples) sampling is of primary use with your application,"

Actually, it's not. I only use 24/96 and I'm doing systems with the Northstar transport that are 24/192. Both of these sound superior to 44.1.

"you can easily use one of the latest Crystal or AKM S/PDIF receivers and configure it in Master Mode with own master clock generator. In this case the master clock is on-board with the DIR and DAC and if precision external XO or TCXO is used you can have as low as 5-10pS (RMS) jitter. This will be free of the problems with the I2S cable."

This is still using a PLL to derive the clock. Seems like it will have more jitter than the simpler I2S interface.

Steve N.