I2S vs SPDIF/USB

Dedicated to the silver disk spinner
Post Reply
iansr
Old Hand
Posts: 401
Joined: Thu Nov 08, 2007 10:44 pm
Location: West Riding

#1 I2S vs SPDIF/USB

Post by iansr »

I've often read that I2S is inherently superior to SPDIF or USB for transmission of digital audio data but I've never had any glimmer of understanding of why that is. If you are in the same boat here is an explanation aimed at the layman, see post 329:

http://boards.psaudio.com/showthread.php?t=3412&page=33

BTW, this thread is about a new transport (more of a digital portal device actually) that PS are developing that sounds really interesting; if it does half of what they say it will be a great bit of kit. if you want to know more about it but don't want to waste time reading the long thread, go to the PS Audio site and have a look at the newsletters.

For the avoidance of doubt I have no affiliation with PSA (unfortunately).
"Its good enough for Government work."
User avatar
Nick
Site Admin
Posts: 15711
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

#2

Post by Nick »

I assume you mean the post that contains

"The problem with ANY serial interface is jitter. Jitter comes from timing problems occuring (in this case) from the process of taking the serial data and separating it back into its three components of clocks and musical data. The reason I2S is the perfect way to send data is it never gets embedded - it never is in a serial format and therefore it never has to be taken apart - because it was never put together."

This is a bit of a miss-truth, i2s is definatly serial in any sensible definition of the term, ie, its doesn't present data in a parallel form. However the key point is that it contains a distinct clock line, so the reciever doesn't have to extract the clock from the bi-phase signal that is at the center of sp/diff.

i2s should be a robust format to transfer data, its at TTL levels, but using the correct buffers should be fine via cat-5 type cables. Delay isn't a issue, as long as all three signals are delayed by the same amount (data, bit clock, word clock). It has space for 24bits of two chan data, and its upper speed is limited by the method used to transfer it. 44k CD data requires a 2.1 MHz bit clock which is well within the limit of standard logic, 192k requires a 60MHz which is more of a challange, but in this worl of 1Gb ethernet should be do-able.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
iansr
Old Hand
Posts: 401
Joined: Thu Nov 08, 2007 10:44 pm
Location: West Riding

#3

Post by iansr »

Nick
yes that was the post. Thanks for the further illumination.
Ian
"Its good enough for Government work."
Post Reply