I recently ordered a couple of DACs and I've only ever used the internal ones in various devices.
If I use one on my computer, I'll obviously use the USB input. But what about other devices that have both an optical and coax output. Is there any reason to choose one over the other?
I'm sure there are technical differences, which I'm interested in, but I'm mainly concerned with any audible differences.
Does it just vary depending on the particular DAC, in which case you just try both and listen? Or is there some consensus 🙂rofl🙂 that applies to DACs in general?
If I use one on my computer, I'll obviously use the USB input. But what about other devices that have both an optical and coax output. Is there any reason to choose one over the other?
I'm sure there are technical differences, which I'm interested in, but I'm mainly concerned with any audible differences.
Does it just vary depending on the particular DAC, in which case you just try both and listen? Or is there some consensus 🙂rofl🙂 that applies to DACs in general?
I've read opinion in the Hi-Fi press that an optical connection is superior provided you use real glass fibres as in this example from QED:
QED Reference Optical Quartz
QED Reference Optical Quartz
If you have a DAC which don't rely or is influenced on the clocked conveyed on the optical link (i.e. a good internal clock generation solution) and don't need higher Fs than 192ksps - I would choose opto thanks to its superb isolation.
//
//
Coax wins in many if not most cases (from standard cheap Toslink cables) but it is worthwhile to check if either source or DAC have isolation/pulse transformers.
+1 with J-P with the true ohmic matching of the plugs and jacks with the cable.
Not tried ECDESIGN new way with led though nore a true high end glass fiber spidf (but cheaps ones : pouya, not for me !)
Not tried ECDESIGN new way with led though nore a true high end glass fiber spidf (but cheaps ones : pouya, not for me !)
I keep meaning to try coax, I presently use Atlas Mavros Glass Optical Toslink to Chord Qutest - perfect galvanic isolation. It sounds much better than my high end USB cable and source. Some of this must be down to the coding on the FPGA chip by Rob Watts as well.
Coax seems more popular, in terms of available high end sources.
I have a spare coax cable, I'll knock one up and compare.
Coax seems more popular, in terms of available high end sources.
I have a spare coax cable, I'll knock one up and compare.
I tried ECDesigns, great in the technicalities, but for me it was just to neutral (the DAC that is). It's my belief that toslink works just as well. The ECDesigns DAC was still sensitive to source (although lesser than some DACs).
True - I use a matrix audio x spdif 2 which is a darn fine DDC, I will try a ultra clocked pink faun card soon.
This, plus a chord DAC, ‘should’ mean that toslink is superior, but proof is in the listening.
This, plus a chord DAC, ‘should’ mean that toslink is superior, but proof is in the listening.
More snake-oil... That statement is utter nonsense.
The quality of the emitter and photodiode are _way_ more important than the fibre, so long as the losses aren't too great in the fibre you use. These audio signals are _extremely_ low frequencies by the standards of most fibre links which are in the 0.5GHz to 500GHz range (per optical carrier), but you don't tend to see the properly fast optical transducers for audio links as they cost _very_ much more, so jitter is typically dominated by the cheap transducers used. 75ohm coax is a great alternative, and low jitter interfacing is typically easier/cheaper with copper, but there is the downside of ground-loop formation.
Its very telling that the "reference optical quartz"link actually mentions boro-silicate glass (which is _not_ silica, despite the marketing use of the word "quartz"). Pure amorphous silica is used for long-haul ultra-high speed single-mode optical links as its the best there is for this. For short haul and very low frequencies the issue of optical dispersion is irrelevant and any material sufficiently transparent is completely fine as an optical fibre, single mode, multimode, or just free-space light beam.
Light travels 100 metres in the time one bit is transmitted in TOSLINK, so the optical medium is a tiny fraction of a bit time in its effect, and an optical signal can handle rise-times of tiny fractions of a picosecond, so the 10's of nanosecond rise/fall-times for cheap optical transducers are completely swamping any time-domain effects of the optical medium, be it plastic, silica, boro-silicate, diamond or dish-water....
Sorry that you have a jitter sensitive DAC - try another receiver circuit and enjoy the benefit of opto isolation!?
//
What do you think of andrea's solution for the I2S input on upcoming FIFO (input I2S isolator > LVDS over HDMI > output I2s isolator) in terms of isolation, good as opto?
The CMR of LVDS would make it more resilient to external RFI pickup I guess, maybe not quite opto level. Not sure about the isolation level of opto compared to some of the other digital isolator types.
Recalling digital theory the optical fibre will introduce more jitter but any DAC worth it's salt will have re-clocker on the input so try both. Out of interest I tried the coax and optical feeds from my disc player into my Dac/amp. The optical sounded marginally cleaner but neither came close the analogue feed ;-)
Me readings you results in a nervous breakdown and I'm bringing back my Rega Planar 3 before my pills ! 😉
I think you are confused with noise. Toslink has 2 extra conversions going on which induces noise. Distance wise it reigns supreme yes.
If not i²s than BNC would be my choice. And I think many DAC manufacturers would agree. I know of a couple that prefer BNC over Toslink.
If you buffer and retransmit the signal (usually inside the DAC, if not with a little external hardware), then the source noise means nothing as long as you don't drop a bit due to extreme jitter/poor interconnects.
Reclocking in or at the DAC IC achieves this and so do SPDIF receivers like the WM8804 - FIFO buffers so the input and output clocks are as isolated as they can be.
Reclocking in or at the DAC IC achieves this and so do SPDIF receivers like the WM8804 - FIFO buffers so the input and output clocks are as isolated as they can be.