I currently use audio over IP/Ethernet (AoIP or AoE) and thus the quality of the Ethernet network card (NIC) is quite relevant.
- My NIC is a PCie Intel Pro/1000 (see picture below). I would like to experiment and replace the clock with a more precise one.
- In the same fashion, I would like to replace the clocks in the media converter (copper<->filter) pictured below.
Any suggestions please? 🙂
Thanks a lot for any help
-Roberto
P.S. Ideally I would like also to power the NIC directly with DC, but this will be the next step 🙂
Intel NIC:
Media converter:
- My NIC is a PCie Intel Pro/1000 (see picture below). I would like to experiment and replace the clock with a more precise one.
- In the same fashion, I would like to replace the clocks in the media converter (copper<->filter) pictured below.
Any suggestions please? 🙂
Thanks a lot for any help
-Roberto
P.S. Ideally I would like also to power the NIC directly with DC, but this will be the next step 🙂
Intel NIC:
Media converter:
The NIC is simply an interface to a communications network.
I don't see how data packets containing only audio information are affected and not other types of data packets.
Modifying the clocks won't make a difference to audio quality, the clock has to be stable in the first place considering the 1Gbit per second data transfer rate.
All the data control mechnisms are performed by the Data Link Layer and sub layers.
I don't see how data packets containing only audio information are affected and not other types of data packets.
Modifying the clocks won't make a difference to audio quality, the clock has to be stable in the first place considering the 1Gbit per second data transfer rate.
All the data control mechnisms are performed by the Data Link Layer and sub layers.
Have you please any suggestion on what I asked? I know exactly what I am doing 🙂
Please check the PTPv2 protocol for AoIP and the Delta clock drift parameter that is adaptively chosen according to the physical layer of IP, so the better the smaller the Delta.
Cheers
Roberto
Please check the PTPv2 protocol for AoIP and the Delta clock drift parameter that is adaptively chosen according to the physical layer of IP, so the better the smaller the Delta.
Cheers
Roberto
Check your connection information for data loss
If there's no data loss, there's no data loss
99% chance there's no data loss
If in the rare case there was an issue with the card, 15-20$ for a new one
FWIW, I use shairport-sync and an older laptop with a terrible wireless connection that dips to about 2% gigabit speeds.
If there's no data loss, there's no data loss
99% chance there's no data loss
If in the rare case there was an issue with the card, 15-20$ for a new one
FWIW, I use shairport-sync and an older laptop with a terrible wireless connection that dips to about 2% gigabit speeds.
Unless this is for a professional broadcast application I dont see the point in a home hifi situation since the tcp/ip stack provides a buffer for out of order delivered data packets.
I think you need to post this in a professional audio forum.
PTPv2 protocol is a timing mechanism, still wont affect audio quality, the data packets all arrive at their endpoints at the same time.
I think you need to post this in a professional audio forum.
PTPv2 protocol is a timing mechanism, still wont affect audio quality, the data packets all arrive at their endpoints at the same time.
Last edited:
P2Pv2 AOIP is interesting... I guess this NIC becomes the master clock and is either point to point or broadcast through a switch? If through a switch will that need to be dealt with too or some algorithm applied to subtract that jitter?
I assume there are FPGA Dev boards that could be better tuned as an alternative to a nic with internal DLLs and unknown ASIC design practices?
I assume there are FPGA Dev boards that could be better tuned as an alternative to a nic with internal DLLs and unknown ASIC design practices?
This could be interesting in diyAudio too, whole house audio, synchronization wireless speakers, etc.
Its really intended for professional applications. If you wanted to send audio around your home you could just use multicast. No need for femtosecond data packet synchronisation.
Probably so but still can trickle down just like Ethernet itself. As you might not care that much with jitter in whole house audio, you probably will in a home theater wireless setup.
Perhaps moving to 10Gb would be even easier.
Before attempting to power via a separate DC source, I would focus more on either replacing, adding if not present, capacitors where needed and more vital. Often, the traces to components are bypassed minimally and minimal traces that could make trying to power it separately useless. If you can find the capacitors, you can even solder another identical on top of the other. If you're changing out the clocks, this would be easy.
Before attempting to power via a separate DC source, I would focus more on either replacing, adding if not present, capacitors where needed and more vital. Often, the traces to components are bypassed minimally and minimal traces that could make trying to power it separately useless. If you can find the capacitors, you can even solder another identical on top of the other. If you're changing out the clocks, this would be easy.
Are you talking about the slave clock offset? If so that would be affected way more by your switches and cabling than the NIC clock drift. Besides, by default PTP syncs every second so the clock must be pretty bad if it has any meaningful drift inside a second.
- Status
- Not open for further replies.