Ethernet cables

[acatala]

2 hours ago, Cebolla said:

 

Remember to follow the thread of certain posts to avoid quoting utter twaddle (ie, take a look at @mansr's replies to such, above)!

 

Which reply from @mansr talks about MAC layer? Am I missing posts?

 

[mansr]

37 minutes ago, acatala said:

Which reply from @mansr talks about MAC layer? Am I missing posts?

I can talk about it now. The Ethernet link provides error detection per frame using a 32-bit CRC. Retransmission and flow control, if required, are handled by higher level protocols such as TCP. UDP based application protocols can of course also include such features. This can be useful if packets vary in importance, in which case retransmission might be requestion only for important packets.

[davide256]

51 minutes ago, mansr said:

I can talk about it now. The Ethernet link provides error detection per frame using a 32-bit CRC. Retransmission and flow control, if required, are handled by higher level protocols such as TCP. UDP based application protocols can of course also include such features. This can be useful if packets vary in importance, in which case retransmission might be requestion only for important packets.

All of which is unaware of clocking done at DTE endpoints. Clocking and clocking recovery for audio purposes happens at a layer above/outside IP packet. COS in IP helps prioritize audio/video packets over  delay insensitive data packets where near real time behavior is desirable... near because speed of light limitations and packet switching delay. Whats interesting is that this works, bridging Ethernet to router and renderer from music server... you would think using a simple switch should sound the same...

 

 

[acatala]

5 hours ago, mansr said:

The Ethernet link provides error detection per frame using a 32-bit CRC. 

 

Correct. So in a non-defective network devices scenario, is it reasonable to say that the bit error rate at layers over Ethernet will be negligible?

[mansr]

28 minutes ago, acatala said:

Correct. So in a non-defective network devices scenario, is it reasonable to say that the bit error rate at layers over Ethernet will be negligible?

Assuming normal 1500-byte frames, it is extremely unlikely that an error will be undetected by a 32-bit CRC. In typical conditions, even detected errors are very rare. It's not something one need worry about.

[acatala]

2 hours ago, mansr said:

Assuming normal 1500-byte frames, it is extremely unlikely that an error will be undetected by a 32-bit CRC. In typical conditions, even detected errors are very rare. It's not something one need worry about.

 

100% agree!!

 

In a scenario where the LAN hardware (electronics and cables) is not defective and the network load is not over the saturation threshold the rate of lost IP packets and UDP/TCP segments  will tend to zero. Therefore, from the point of view of data loss, any UTP cable that just does its job will serve to our purposes and there is no real benefit in using an expensive one. 

 

In regard of presumed phase problems, any buffer technique  would solve or, at least, minimize them. Is this right?

 

 

 

[mansr]

17 minutes ago, acatala said:

In regard of presumed phase problems, any buffer technique  would solve or, at least, minimize them. Is this right?

I'm not sure what problems you're referring to. An Ethernet stream is bursty by nature, each frame arriving at precisely the link rate, typically 1 Gbps, so for audio use, the receiver must have a buffer in order to deliver a continuous data stream to the DAC.

[jabbr]

55 minutes ago, acatala said:

In regard of presumed phase problems, any buffer technique  would solve or, at least, minimize them. Is this right?

By "presumed phase problems" I assume you mean that the input Ethernet clock is not in phase (nor frequency) with the DAC clock. A FIFO buffer is used to cross the async clock domains and present data to the DAC in sync.

[iiDexii]

I heard clear difference in favour of Supra Cat8, versus heavy folded freebe Huawei Cat5 connected to bluray/streamer.

 

It may be a strong Placebo thing going around the world.

 

But if it is the Strong P, it has definitely changed my listening habits.

 

Before I only listened to my setup- Spotify and youtube music on home party's or as background music mixed to 5.1, stereo 2.1 sounded poor with no details and timing/bass.

I rather played Cd or Sacds.

 

Now good quality music on Spotify sounds detailed an with precised and timed full bass.

 

I shock now listening to Spotify critical in 2.1 it's fun. Netflix also sounds more like a dvd or bluray...

 

Its more understanding, when Spotify from phone to cheap 27 bucks Sony headphones and blutooth speaker Audio Pro Addon T3 sounds great on Spotify...

 

My 4700 bucks setup, should sound nice on Spotify with the free cable if it's not like this.