Digital to coaxial

[Dwaindibly]

I'm currently using an audioquest pearl optical cable from my wiim to my marantz amp streaming from tidal at 24bit 192khz I could not find the specs of the cable so I contacted audioquest and have just had a reply stating that the cable supports upto 24bit 94khz but yet it's sounds fine

 

[voodooless]

it's sounds fine

Then it is fine

 

[somebodyelse]

There is no such requirement as far as I know. You can also get by with special line drivers that offer transformer-live floating outputs, like, for instance, from THAT (1606), but also the classic TI DRV134 or SSM2142.

Maybe I'm misremembering from when I went digging into the specs, or paraphrased incorrectly. I'll have to go digging again to check.

 

[boxerfan88]

Or something like this USB-to-SPDIF converter…

DA&T 谷津音響 - Product Information

谷津音響認為音響是一種商品，也可以是一種藝品，但不應該是一種神話。

www.da-t.com

 

[Berwhale]

This is very far removed from a fiber optic network connection that uses glass fibers to transport multi-10-gbps

Indeed, for the OPs benefit... Here's an example of 10G Active Optical Cable (AOC) that I happen to have lying around at home...

The electronics (SFPs) on the ends of the cable contain the transducers (LEDs in this case) to covert and electrical signal into light to send down the fibre. The fibre is only OM2 standard for this particular AOC because it's short (2m). You can also use separate SFPs and fibre, but integrating the electronics with fibre gives the manufacturer more control and allows them to use lower standard (and cheaper) components without compromising overall performance and functionality of the AOC (there are downsides to AOCs as they can be a PITA to install and vendor supplied ones may not be fully compatible with your all your equipment)

There's more about the types of fibre used in networks and their capabilities here: https://community.fs.com/article/advantages-and-disadvantages-of-multimode-fiber.html

 

[antcollinet]

Folklore had it that jitter was worse on toslink. The sample rate limitation may have come into it too.

Coaxial may have galvanic isolation too. Transformer isolation of the transmitter is required in the AES/EBU spec, so pro gear will have it. Domestic gear might, but you will probably have to test it to find out. If they can put one in a <€20 Fun Generation USB interface the cost really shouldn't be an excuse.

I've never seen a device with transformer isolated outputs, and finding out before purchase is not that easy, since it isn't generally specified.

 

[MCH]

I've never seen a device with transformer isolated outputs, and finding out before purchase is not that easy, since it isn't generally specified.

Old datasheets used to recommend them:

For some reason, SPDIF output hats for raspberry pi often have them:

As do all minidsp products of which I have seen a teardown (flex, shd...), okto DAC pro....

 

[antcollinet]

Old datasheets used to recommend them:
View attachment 410385
For some reason, SPDIF output hats for raspberry pi often have them:
View attachment 410386

I'm not saying they don't exist. Just that they are not that common - and knowing what you are getting is difficult.

 

[somebodyelse]

There is no such requirement as far as I know. You can also get by with special line drivers that offer transformer-live floating outputs, like, for instance, from THAT (1606), but also the classic TI DRV134 or SSM2142.

Following up, it appears to have been mandatory in the first version of the EBU spec, but not the AES version. From the EBU Engineering Guidelines section 1.3:

This specification, AES3-1985, was put forward to ANSI, the American national standards authority, for
ratification and also submitted to both the EBU in Europe and the EIAJ in Japan for their approval. Both
bodies ratified the standard under their own nomenclature, although small modifications were made to both the
text and the implementation. The most significant being the mandatory use of a transformer in the transmitter
and receiver in the EBU specification. Despite these small discrepancies the interface is now commonly
referred to as the "AES/EBU" interface.

I guess it was dropped in later EBU versions - it's not in the third edition.

 

[Dumdum]

If you listen without knowing the connection type and without peeking, you will never be able to tell the difference.

Unless you have a ground loop that the optical solves. Or compare the two between different DACs not level matched, or different filters, or one of the DACs is junk.

Some people believe they hear differences between optical and coax with everything else matched. They actually can’t and fail to identify the differences in a proper test where they are not allowed to see which connection is being used.

It is after all data. And audio is on the extremely easy side of data I/O.

We did a blind abx test with coax vs optical and it was indeed indistinguishable…

I was curious and we used a topping d10s to feed both into the same Dsp and switch randomly… zero difference

We tested a couple of dsps that people said had different sound characteristics… yup… zero difference too…

The one thing we did find was one rca we could pick out when randomly swapped with three others, we could pick out the one regardless of what combination we used of the four… I made three of them and they measured identically in a loopback setup… and yet we could pick out the one repeatedly

Van Damme Silver series lo cap 55 was just a little brighter, but that’s it… a tiny difference, I’d picked it up in my car, but didn’t know if it was real… so a blind test was needed to confirm if it was in my head or not

 

[Randyman...]

We did a blind abx test with coax vs optical and it was indeed indistinguishable…


Van Damme Silver series lo cap 55 was just a little brighter, but that’s it… a tiny difference, I’d picked it up in my car, but didn’t know if it was real… so a blind test was needed to confirm if it was in my head or not

A digital cable somehow only affected the bits that represented high frequencies, and changed them to become louder (brighter)? And was significant enough to be detected in a car audio system by listening? I wonder how the cable knew which bits were which, and how it changed the bits?

What did it measure when tested?

 

[somebodyelse]

A digital cable somehow only affected the bits that represented high frequencies, and changed them to become louder (brighter)? And was significant enough to be detected in a car audio system by listening? I wonder how the cable knew which bits were which, and how it changed the bits?

What did it measure when tested?

More likely the coax connection wasn't transformer isolated, and the cable's shield resistance was sufficiently different to alter the ground noise. Assuming the tests were properly controlled, including that nobody had found a 'tell' - the paper from one test included one participant having noticed a slight difference in the sound of one of the switching relays, so the test was no longer blind.

 

[antcollinet]

More likely the coax connection wasn't transformer isolated, and the cable's shield resistance was sufficiently different to alter the ground noise. Assuming the tests were properly controlled, including that nobody had found a 'tell' - the paper from one test included one participant having noticed a slight difference in the sound of one of the switching relays, so the test was no longer blind.

But noise (including ground loop noise) sounds like noise, not brightness - which is mainly a frequency response issue.

 

[Dumdum]

A digital cable somehow only affected the bits that represented high frequencies, and changed them to become louder (brighter)? And was significant enough to be detected in a car audio system by listening? I wonder how the cable knew which bits were which, and how it changed the bits?

What did it measure when tested?

The rca was analogue not spdif coax, sorry that wasn’t clear… I would hope you’d realise that… as a loopback test with different coax digital cables would show no differences

 

[Dumdum]

More likely the coax connection wasn't transformer isolated, and the cable's shield resistance was sufficiently different to alter the ground noise. Assuming the tests were properly controlled, including that nobody had found a 'tell' - the paper from one test included one participant having noticed a slight difference in the sound of one of the switching relays, so the test was no longer blind.

I wasn’t clear, it was analogue rca we could distinguish one cable from all others

 

[Randyman...]

Was not initially clear - it is now.

 

[Vigovsky]

I create my original VR-1 AudioPCs based on INTEL DN2800MT for several years, specialized for output with upsampling
24/ 96, 192 for balanced XLR AES/EBU, BNC coaxial outputs and USB. AES/EBU and BNC outputs have transformer galvanic isolation.
For the system and storage of different audio formats, PCI Express mini card SSD is used.
SATA DVD/CD drives are also used, programmatically configured for silent operation хх10.
Power is supplied by external pulse and linear power supplies 12V 10A.
The case was created specifically for this Audio PC, the panels and tray were CNC cut and powder coated.
These PCs have been working reliably and stably for many years. The sound is excellent. Galvanic isolation is important.