00940
Headphoneus Supremus
- Joined
- Nov 6, 2002
- Posts
- 4,493
- Likes
- 47
Quote:
It doesn't work that way. The signal timing is recovered from the edges of the packets. If the edge of the packet isn't square enough (something a cable could potentially influence), you introduce incertainty in the recovered clock, but it's not like you don't have a clock, you just have a jittery one. The protocol is sufficiently robust to work within such incertainty and to reduce it to a certain extent (that's the job of a PLL). However, those variations in the time domain will translate as distortions at the analog output (not as crackle or pop). The math for that is in some papers I've posted in the other threads (about spdif but the reasonning is the same). That's why some usb cables with big ferrites are fine for data transfers but might be problematic for usb audio transfer, as they "round" the edges of the packets.
Still, my point of view is that interface jitter introduced by a short, decent cable is no big deal. I myself use a cable grabbed from an old HP printer.
If a cable (are there any on sale?) distorts the waveform such that it cannot recover the timing signal, then surely it would stop working rather than affect sound quality?
When I say stop working I mean even if it just for a part of a second and it introduces a crackle or other sound that is not part of the music. Not just if the signal comes to a complete halt.
It doesn't work that way. The signal timing is recovered from the edges of the packets. If the edge of the packet isn't square enough (something a cable could potentially influence), you introduce incertainty in the recovered clock, but it's not like you don't have a clock, you just have a jittery one. The protocol is sufficiently robust to work within such incertainty and to reduce it to a certain extent (that's the job of a PLL). However, those variations in the time domain will translate as distortions at the analog output (not as crackle or pop). The math for that is in some papers I've posted in the other threads (about spdif but the reasonning is the same). That's why some usb cables with big ferrites are fine for data transfers but might be problematic for usb audio transfer, as they "round" the edges of the packets.
Still, my point of view is that interface jitter introduced by a short, decent cable is no big deal. I myself use a cable grabbed from an old HP printer.