What is the best connection to use if you have a choice between Toslink, SPDIF, USB and AES?
Also does anybody have any opinions on Toslink cables, is it worth buying more expensive Toslink cables? is there any difference in SQ?
SPDIF is the best choice.
Toslink is limited, itâs the Toslink ends that are the limitation and itâs not worth spending money on more expensive âcablesâ.
1 Like
Latest high end dacs seem to favour more the USB. With spdif, aes/ebu and toslink being a legacy.
2 Likes
You really donât need to spent a lot for a (very) good Toslink cable: Lifatec, one on the very few to use REAL glass.
3 Likes
There is no simple answer, it depends. If you need to support high sampling rates or native DSD, then either USB or I2S are your only options. If DoP is acceptable, and it should be, then either AES or coaxial are good for up to 8X Frecuency sampling (384 kHz). If you can avoid USB do it, since those are generally the noisier interfaces, specially the ubiquitous Amareno.
Toslink and coaxial are both SPDIF. Coaxial (8X) supports higher sampling rates than Toslink (4X), but Toslink gives you galvanic isolation. Lifatec and Sys. Concept are very good Toslink options if you choose to go that way.
4 Likes
I used a TosLink cable between my CD Transport and my NDX2 to begin with, but ordered a Nordost SPDIF Phono(CDT) to BNC(NDX2)
The SPDIF cable came today and I changed it this afternoon. It definitely sounds better. Much more detailed and a fuller sound.
1 Like
AES/EBU if you have balanced components connecting. Just like DIN it has the ground covered, the impedance of the cables is conform specs and good cables are very affordable, since it is used so much in Pro Audio. It suffers from none of the drawbacks of one of the other connections. You will hear it instantly.
If one of the components connecting is not balanced, then SPDIF (COAX) is the next best.
USB is very much depending on how it is implemented in the components used. It can be good, it can be horrible.
TOSLink: Van Den Hul the OptoCoupler is one of the few TOSLink cables with a glass head you can still find with good ends, they are not very expensive.
1 Like
AES3 does not transmit line level audio signals, so it doesnât really matter if the connecting components are balanced or single ended.
1 Like
I would have thought that the benefit of balanced cabling applies not only to audio analogue signals, whether line level or lower, but equally to picking up other unwanted noise, notably RF, which can cause adverse effects in the DAC. It therefore makes every sense to me.
It may also be that using AES with one connected end not balanced might modify impedance and thus depart from the benefit of impedance matching, but I am unsure whether that would be the case, nor whether it would affect the signal at all unless so bad as to corrupt data.
2 Likes
With USB, clocking is at the DAC side, done by the DACâs built-in clock. With S/PDIF, clocking is at the source and DAC needs to reconstruct the clock from the data stream using PLL.
USB inputs are not inherently noisy! USB is an excellent choice for a DAC input, as long as the USB SOURCE is designed to be low noise (meaning a purpose built streamer/renderer designed for high end audio use). In fact, wired Ethernet sources have a lot of network noise on them. Just as is the case with USB, an Ethernet input on a DAC needs to be very carefully implemented to not produce noise which gets to the DAC clock circuit and analog circuitry.
A well implemented USB input can perform as well as, or better than, any other DAC input.
So technically speaking, USB should sound better. However, your experience is the other way round, but as always, your ears tell you what you like best.
1 Like
No amount of fiddling, e.g with linear power supplies can stop a computer producing RF because of the very nature of computer processing, using RF clocking, therefore as even the best streamer/renderer is a computer inside I think it will still have significant RF superimposed on the digital signal unless heavily filtered at the output from the computer stage⌠which of course can be done with any renderer. But in practice the key to good USB audio data transmission is rejection within the DAC, then the noise quality of the renderer becomes less significant.
1 Like
I think thatâs a thing of the past.
Iâm pretty sure the ndac doesnât rely on timing info from the transport.
1 Like
The USB interface on the DAC will generate RF noise too, thatâs why Chord DACs for example work best with the dual coax inputs.
1 Like
I meant exactly what I said, the clock master is the at the source, and the nDAC has to reclock it using PLL.
1 Like
It depends on the boxes and their interface implementation. I prefer AES with my DAC and a single cable but then most of my content is ripped CDs so the 24/192 limit doesnt bother me.
.
1 Like
If I understand it correctly, the nDAC has a data input buffer and data gets clocked out of this buffer using the nDACâs master clock, so there is no need to reconstruct the clock from the input signal. This methodology is featured in all Naim DACs since the nDAC.
2 Likes
I see, so the nDAC is an exception, considering that it was designed in 2007! thanks.
4 Likes
After comparing USB ASIO (normal) with USB Bulk Pet, I canât really recommend USB as implemented by most things. The difference between ASIO and Bulk Pet was large. It significantly reduces processing noise at the receiving end by reducing the number of packets by 3/4. Which also reduces some ground plane noise at the sending end that is piggy backed on the cable.
When compared, Bulk Pet came out the clear winner, then coax, then regular USB.
As mentioned Toslink provides galvanic isolation which is a good reason to use optical from cheap ânoisyâ sources instead of the electrical methods. But itâs audio standard is throughput limited so high resolution surround sound wonât work.
Toslink can go up to 192/24, which is DAC dependent, so it is more than adequate unless you want to upsample the music stream to higher resolution.
2 Likes