|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
211.26.193.92
In Reply to: Always an issue posted by audioengr on October 30, 2006 at 10:04:25:
It was this interesting article that prompted the question. There are about two pages on the 8 contributions to jitter, but the two lines:“Computer-driven audio usually eliminates jitter contributors 1, 2 and 3. In some cases it also eliminates or minimizes jitter contributors 4, 5, 6, 7, and even 8”
I'm a natural sceptic (is it too good to be true?)
I’m not suggesting the statements are at all not valid, I am simply hoping someone could explain or ‘verify’ those statements.
Follow Ups:
Well, 1, 2 and 3 are obvious. These situations simply dont exist in a computer audio system. Items 4,5,6,7 and 8 were not specified as being intrinsic to computer audio, only that they could also be improved in computer audio implementations as well as Transports.As for 4,5,6,7 and 8:
4: If the S/PDIF interfaces are well-executed, then the jitter will be significantly reduced for both Transports and Computer Audio converters. This includes faster buffers, power current noise control and impedance matching. This is not unique to computer audio, but certain implementations can be better than that possible with Transports. It's in the design details.
5: Digital cables are necessarily usually 1-1.5m long to extend from a Transport to a DAC. With Computer Audio converters, it is possible to conveniently have a 9" long cable. This is a definite advantage for lower jitter. Digital cables experience both losses and dispersion, the magnitude of these being dependent on the materials used and geometry of the cable. Dispersion and losses exhibit themselves as jitter in the system because they tend to be pattern-sensitive.
6: Optical interfaces can be eliminated for both Transports and Computer audio converters, so there is no real advantage of Computer over Transport here. Avoiding Toslink is just a good idea - the jitter is significantly higher than a well-implemented S/PDIF interface due to the two coptical-electrical conversions.
7: The edge-rates of the interface and the resultant jitter in the recovered clock are closely related. This is another obvious conclusion and applies to both transports and Computer audio. If you understand digital electronics, transmission-lines and interfaces then you would know this.
8: Noisy power planes and crosstalk can cause jitter in ANY digital signal. Again, this is common knowledge for digital engineers.
I have a done a little digital design myself. About 30 years worth, including design team lead on the Pentium 2 and architect of the Intel massively parallel supercomputers.
Steve N.
I was also thinking the one more jitter source specifically for PC audio. S/PDIF output signal interference by other computer softwares or operation system itself. Since most of Microsoft windows system are not a real time system, will not well response S/PDIF signal request at time.
This has more to do with the protocol used for USB or Ethernet. USB can have Isochronous or Asynchronous protocols. Asynchronous allows bursting of large packets to the device so that latency is not much of an issue. Most devices do not use asynchronous protocol...yet anyway.Linux will have the same problems I believe.
Steve
Thanks for your comprehensive response; there’s no doubting your knowledge, and a good job of outlining digital electronics to non engineers.
One aspect> Digital cables experience both losses and dispersion, the magnitude of these being dependent on the materials used and geometry of the cable.
I’ll look for a good moderate cost USB cable
Im talking about S/PDIF cables, not USB cables. USB cables are less of an issue because of the reclocking in the converter. I use a good Belkin 16-footer myself.
This post is made possible by the generous support of people like you and our sponsors: