does TOSLINK / ADAT cable length matter?
If you’re here, you probably decided to get an optical cable and are not sure if TOSLINK / ADAT cable length or brand will have large impact on the audio data it will be transferring? That is exactly what I was trying to figure out several months ago and I did not find any solid answer on the net. And that is exactly why I wrote this post and did a little test – so me and you can have some closure concerning ADAT cable length (here’s a link to test files for you to start downloading while you read).
I try to only keep gear that I really need. Because of that I sometimes get into situations where minimal creativity is needed for fulfillment of unusual tasks. 99% of time I only record one voice actor at a time so RME Babyface with a Sennheiser MKH-416 is enough. But there are exceptions of 1% where a client wants to sit in the studio and direct the talent or even two talents acting at a time. Because RME Babyface only has two mic inputs there’s only one way to connect a talkback mic to it – using ADAT input of the interface.
Few months back I bumped into situation where I needed third mic for the first time in many years. I decided to use my Sony PCM-D100 as a fancy talkback mic. The only problem was I needed an optical cable. This is where I got confused.
There’s ongoing debate about quality and length of ADAT cables and how these factors affects data being transferred. Obviously there are huge price differences too. For 9EUR (~10$) you can get a cheap no-name 10 meter long cable. On the other hand for 20EUR (~21USD) you can get an expensive branded 2 meter long one.
Occasionally I like to record lots of people using just a stereo mic so I thought 10 meter would be a good call for such future projects in case I need to hide myself further away to judge the mix objectively. But at the same time who wants to lose quality running data that was recorded on a 700$ recorder trough a cheap no-name optical cable?
TESTING CONDITIONS AND STEPS
Long story short – I got myself a 10 meter long cable but after all the research I still wasn’t sure what effect on audio data ADAT cable length does have. So I got myself another 1m long cord and decided to do some testing.
- Both 10m and 1m long cheap no-name ADAT cables were used for testing.
- I recorded my voice using internal mics of Sony PCM-D100 (48kHz/24bit).
- Then I transferred the file into my PC and Reaper.
- Played it back directly from RME Babyfaces optical output and recorded it using its optical input.
- Played it back from D100’s “Optical out” and recorded into RME’s optical input. Additional adapter was used to connect standard ADAT cable to recorders optical output.
- For sake of variety I used both “Internal” and “Optical In” word clock sync.
- All files were normalized to -3 dB to compensate for pan-law inconsistencies introduced by software settings.
- All files were sample-accurately aligned in the timeline.
- Then I used “phase-reverse” switch on one of the two random channels to check if they cancel each other out.
“PHASE REVERSE” METHOD
As far as I know “phase-reverse” method is the only objective way to check if two digital audio files are identical. Because if judged by ear and eye humans usually tend to love prettier GUIs, newer and more expensive things. Not to mention the esoteric warmth, fullness and body some of us hear in identical signals while testing out things we adore more vs. things we adore less. Ah, Placebo…
So does ADAT cable length really matter? Conclusions:
- For carrying stereo data the length and brand of an optical ADAT cable doesn’t make any difference (if cables are not defected). Because all the files re-recorded using the original file were canceling out each other completely. All except two…
- Two files – the ones re-recorded using RMEs optical output routing it directly into optical input – did cancel each other out, but there was noise peaking at -130dBFS present trying to cancel them out with each of other files. Although the difference might be only because of some kind of dithering Babyface applied while sending the data out of optical outs. Not sure what that was, but after rendering out the sum with -130dBFS noise and normalizing it to 0dBFS it was only white-ish noise that was audible – no sign of any kind of useful data related to original file.
- As those tests were performed with only two channels (stereo it is) of audio, thus I can only state that you won’t lose any quality using cheap no-name ADAT cables transferring TWO channels.
Let us keep in mind that TOSLINK / ADAT is intended for 8 channel transfers of audio data however and to check if these could be lossless using same cheap mediums we will require corresponding tests. If you’ve performed such – would be great if you shared your insights! If not – thanks for reading, hope you enjoyed it and found it useful. Questions / suggestions – feel free to drop a comment little lower. Cheers!
P.S.: once gain here are all the files I recorded and re-recorded. Check ’em for yourself. Maybe you’ll find something different.