Wednesday 1 May 2013

MEASUREMENTS: TosLink digital optical audio cables.

Let us now finish off testing the digital audio cables at my disposal. Already we're seen that USB and coaxial SPDIF cables did not have any measurable differences even with poor construction and length. Remember that the benefits of TosLink ("Toshiba Link") is that they are non-electrical so have the benefits of being resistant to electrical interference. Furthermore, it has been said that they can be used in long runs without loss (specifications allow up to 33 feet).

I have a few old TosLink cables hanging around of varying qualities and build. Let's have a peek in the closet:

Cable A:
This is the cheapest cable I have. I believe it was a freebie from one of my old DVD players. The only marking is along the side of the cable it says "VITOnet". Length 6' and relatively thin.

Cable B:
A small step up from the freebie. Branded "Thunder Cable" TosLink I got as a package with some other home theater cables I bought maybe 10 years ago from Costco locally. Again 6' and thickness of the cable about the same as the VITOnet, so it feels quite fragile in the hand.

Cable C:
Another step up in construction and price. A 6' Acoustic Research Pro TosLink cable. Nice gold plated connector now with good strain relief and quite a thick caliber. Maybe about $20 when I bought it. Has served me well for years between the DVD player and a Denon receiver.

Cable D:
Up to this point, all the cables have been plastic optical fiber. This next one is of glass cable construction. Can be purchased off Amazon here for ~$20. Construction is also very good with nice metal connectors, gold plated terminals, and good cable thickness. According to the "specs", the cable supposedly has "280 individual stands of glass". Again, the length is the same at 6'. Note that it's not necessarily a given that glass is superior to plastic; for LED light sources, it has been said that plastic cables might even be better.

Cable E:
The "worst" Dr. Frankenstein cable I could construct from what I had was with the use of a $3 TosLink coupler off eBay. This adds an extra transmission interface and of course extends the length of the cable which could worsen signal integrity. I coupled the 2 cheapest TosLink cables I described above - the VITOnet (Cable A) and Thunder Cable (Cable B) - for a total length of 12'. I tested the cable looking just like this with it coiled up.

Analogue Tests (RightMark, 24/96):

Standard methodology:
Win8 laptop --> 6' shielded USB cable --> CM6631A asynchronous USB-to-SPDIF --> *test TosLink cable* --> AUNE X1 DAC TosLink input --> shielded RCA --> E-MU 0404USB --> Win8 laptop

Okay folks, here's the summary...

I think as you can see from this table, the test results are essentially exactly the same. 'Cable E' was measured about 2 weeks after all the others (I was waiting for the coupler to arrive from Asia); notice how reproducible the test results are.

Frequency Response:

Noise Level:

THD:

Stereo Crosstalk:

No difference.

Jitter Analysis (Dunn J-Test Signal 16-bit & 24-bit):

Cable A (6' VITOnet):


Cable B (6' Thunder Cable):


Cable C (6' Acoustic Research):


Cable D (6' Glass):


Cable E (12' Coupled VITOnet + Thunder Cable):


For the 12' Cable E, I even started waving the cable around at the coupler joint to see what would happen to the jitter spectrum... Nothing of significance!

Notice that the CM6631A USB-to-SPDIF converter has quite low jitter overall despite this being TosLink - few nasty sidebands and little spectral spreading of the primary signal (usually in my other measurements TosLink not as good as the coaxial interface).

Conclusion:

TosLink cables measure the same based on these results. Inexpensive 6' plastic, better quality plastic, glass, or a longer stretch of 2x6' cable with a coupler did not show any measurable difference in the analogue output from the AUNE X1 DAC used to test. Likewise, jitter analysis with the Dunn J-Test in both 16 and 24-bit versions did not demonstrate any appreciable difference between the cables. As far as I can tell, jitter does NOT appear to be affected by the TosLink cable quality based on these tests.

Subjective listening to the 12' coupled Cable E likewise did not show any musical anomaly - was able to enjoy Jorma Kaukonen's Blue Country Heart thoroughly through the AKG Q701's :-).

I think this concludes my survey of digital cables - at least for now...

13 comments:

  1. Thank you for these instructive tests. I guess the results would be the same with an active toslink switcher like the viewhd model? Jean

    ReplyDelete
  2. Thanks, reading. Awesome blog.

    ReplyDelete
  3. Arch,
    "Already we're seen that USB and coaxial SPDIF cables did not have any measurable differences even with poor construction and length."
    I think your also excellent test here http://archimago.blogspot.co.uk/2013/02/measurements-logitech-touch-as-transport.html shows that cheap coax S/PDIF does measure differently ...

    ReplyDelete
  4. Thank you for the tests with statistics, and settling the debate.

    ReplyDelete
  5. This comment has been removed by the author.

    ReplyDelete
  6. Great and helpful. The only post that made real sense. Kudos

    ReplyDelete
  7. Reading this interesting test, I wondered if it is possible to conduct a comparable test with a three-way passive switchbox. Using such a box I can not hear the difference compared with a situation without switch box.

    ReplyDelete
  8. What if your measuring chain is noisier than the toslink cables to be measured? In that case you could not distinguish one from another.

    Thanks
    -R

    ReplyDelete
  9. Interesting test, Thank you.
    I was wondering if you are able to test which cables are able to pass 24/192?

    ReplyDelete
  10. I wonder if a better DAC would reveal more? relatively speaking Im not sure? Good write up.

    ReplyDelete
  11. I notice the crosstalk measurement is remarkably consistent (0.2db range) except for the one cable that's minimum 0.5db clear of the rest. The fact this is the sole glass cable seems too much of a coincidence to ignore?

    ReplyDelete