Jplay

[GuillaumeB]

The reclocking in the TotalDAC is necessary for converting USB to S/PDIF so it can output it over coax or AES/EBU.

Music server PC's that usually output USB are optimized to lower noise and are regularly outfitted with purpose built USB interface cards (like the JCAT I own, see specs here: http://jplay.eu/jcat/) with high levels of noise filtering and a high quality clock. This component can in fact also be seen as a 'reclocker' (but one operating fully in the digital domain), the data received by the card is reconstructed and outputted by the card using it's local clock.

Note that there's no S/PDIF involved in which the data stream has an embedded clock signal that has to be reconstructed by the receiver, this is where usually a 'reclocker' as known by audiophiles (Digital Lens, Wyred4Sound's Remedy, Apogee Big Ben, etc.) is used.

An interesting Q+A with Brient, the developer of TotalDAC. In this context especially Q+A 3 is relevant.

source: http://singaporehifi.blogspot.nl/2014/09...ilter.html

Mr. Brient kindly agreed to answer some questions I had about his cable and some other general questions. The questions and answers are reproduced in verbatim below :-

Q1. The totaldac usb cable/filter is described as containing "high performance filters to suppress digital pollutions coming from the computer or the music server.". Are both the data and power lines filtered ? Does the box provide galvanic isolation from the computer ?

A1. It filters both data and power. There is no galvanic isolation as this doesn't exist for high speed USB. It exists only for old non-asynchronous DACs with a 12Mbit/s USB bit rate, unable to do 192KHz/24.

Q2. The totaldac cable uses a single cable instead of the increasingly popular route of using separate cables for the data and power lines. Do you have any thoughts on this ?

A2. The filter box is placed very close to the DAC connector, so the cable separation is no critical. The long part of cable is before the filter.

Q3. Some audiophiles feel that USB (even if asynchronous) does not sound good as other connection methods like AES or coaxial. Many are using separate USB/SPDIF convertors, claiming better sound compared to direct hookup to their DAC's USB input. Do you think think this approach has any merit ?

A3. All computer approach need a computer bus to output the digital audio signal. USB, PCI and so on, all have the same difficulty due to the clock and pollution of the computer. There is no computer with true native spdif output, then can only fill and asynchronous fifo, just like the USB does, even if the fifo can be in the processor itself.

USB, even asynchronous, is a sensitive link, this is why the filter helps. For the best sound my customer use the d1-server using the USB cable/filter for the internal loop, then the signal is rebuilt again in AES-EBU in the asynchronous reclocker, then goes to the DAC via AES-EBU.

A computer straight to a USB DAC is not the best way.

All in all the difficulty is not especially the USB, it is the computer for audio, but optimisation are possible, I think that now the d1-server with its USB filter is better sounding than a CD drive.

Wow thanks for sharing Antoine! Very interesting indeed.

Guillaume