Coaxial S/PDIF on Gigabyte X58A-UD3R out of spec?

[sapperlott]

Hi!

To test my new oscilloscope, I decided to have a look at the signals coming from the coaxial S/PDIF output of my Gigabyte GA-X58A-UD3R Rev 2.0 mainboard (yes - with the correct 75 Ohm termination of course )

I was quite irritated when I measured > 1 Vpp since according to the S/PDIF spec (at least according to this document) this should be 0.6 Vpp max.

So I contacted Gigabyte and attached a screen capture of my scope. Gigabyte then forwarded my mail to Realtek which replied that they verified the coaxial S/PDIF output with their Audio Precision analyzer and that there's nothing wrong with it.

I've asked them if I could see those AP eye diagrams and the attached diagrams are what I just received in response. According to Gigabyte, Realtek confirmed those as compliant.

Seriously? I can spot a number of WTFs right off the bat. For starters it can be clearly seen that the signal exceeds 0.6 Vpp by far. Secondly they seem to have used the AES3 settings which won't match the physical interface of S/PDIF no matter what. To top it all off it looks like they didn't even read the nice little remark at the bottom of the diagram stating that they have to adapt some files for sample rates other than 44.1 kHz (since the x-axis for the 48 kHz test has exactly the same range as the one for the 44.1 kHz test).

So it looks like the people at both Gigabyte and Realtek have no clue when it comes to the standards their products claim to implement. If they can't even get something as simple as a coaxial S/PDIF output right, I wonder what one might discover when plugging one of their Gigabit Ethernet cards into a network analyzer *shudder*

It would be interesting to see if other Mainboards with Realtek AC97 / HDA chipsets show the same problem. So if you got one of these and a oscilloscope handy, could you check and post your results here?

By the way - does anyone have a softcopy of IEC 60958-1 / -3 lying around and is willing to share? The prices at the IEC webshop are a bit steep for just a quick peek.

[fmak]

Originally Posted by sapperlott

I was quite irritated when I measured > 1 Vpp since according to the S/PDIF spec (at least according to this document) this should be 0.6 Vpp max.

I wouldn't get so irritated by the signal level; the spdif out from the Realtek stuff is suspect anyway. I like the way they stretched the horizontal axis to hide variations in the vertical one!

What is disturbing is the way Realtek/Gigabyte configure the output, where you cannot be sure if resampling is taking place.

At least in XP, KS is available in the output and you can unmap the device. With Win 7, even Wasapi offers different settable output sample rates, with no support for 32 bit float files nor 88.2/176.4k.

It's really best to avoid MB spdif and use a sound card, until these computer guys have put in a bit of effort to understand digital audio and implement it properly.