ciuchino
New member
Di questo discorso che ne pensate :
http://www.avsforum.com/avs-vb/showthread.php?threadid=436801&highlight=spdif
There has been much debate on this issue, but packet based transmission does not solve the jitter issue at all. While collecting a number of bytes over SP/DIF to receive a complete packet (the packet delay), the clock of the receiver is always synced to the clock of the sender (your player).
Which basically means that the start and end timepoints of the packet will be sensitive to jitter as well : they are determined by the source.
If SP/DIF had flow control the digital=digital reasoning would be true. We can then just buffer everything at the receiving side and inform the sender when the buffer is full (or when empty but it would be more interesting to send the data a little to fast so the buffer is always full). As there would be always data in the buffer, we could read-out the buffer using a very good clock to drive DSP's and DAC's.
Unfortunately SP/DIF doesn't work this way and the receiver just has to take whatever comes in, which means trash in = trash out. However, newer protocols like i-link are bi-directional and do not suffer from these shortcomings. There was even a post on the Tag Mc Laren forum once that if Tag would start using bi-directional protocols between players and procesors (such as firewire), there wouln't be any difference between a cheap and expensive transport.
There have been expensive reclocking and anti-jitter solutions (up to 2500$) for SP/DIF but none of them work perfect as they still reveal differences in the sources, which shouldn't happen.
http://www.avsforum.com/avs-vb/showthread.php?threadid=436801&highlight=spdif
There has been much debate on this issue, but packet based transmission does not solve the jitter issue at all. While collecting a number of bytes over SP/DIF to receive a complete packet (the packet delay), the clock of the receiver is always synced to the clock of the sender (your player).
Which basically means that the start and end timepoints of the packet will be sensitive to jitter as well : they are determined by the source.
If SP/DIF had flow control the digital=digital reasoning would be true. We can then just buffer everything at the receiving side and inform the sender when the buffer is full (or when empty but it would be more interesting to send the data a little to fast so the buffer is always full). As there would be always data in the buffer, we could read-out the buffer using a very good clock to drive DSP's and DAC's.
Unfortunately SP/DIF doesn't work this way and the receiver just has to take whatever comes in, which means trash in = trash out. However, newer protocols like i-link are bi-directional and do not suffer from these shortcomings. There was even a post on the Tag Mc Laren forum once that if Tag would start using bi-directional protocols between players and procesors (such as firewire), there wouln't be any difference between a cheap and expensive transport.
There have been expensive reclocking and anti-jitter solutions (up to 2500$) for SP/DIF but none of them work perfect as they still reveal differences in the sources, which shouldn't happen.
)
.
