Wanted to chime in a bit in regards to the digital connection discussion. Apologies in advance for the lack of formatting, will get this tidied up a bit later!
The first thing is that there are three key aspects to consider when it comes to a digital audio connection.
- Noise
- Data Integrity
- Jitter
Noise:
The first thing to consider is simply how noisy your source is, and what effect this may have on the receiving device. Using a higher quality source like an SMS200 Ultra will give you a much lower noise level on the output than a raspberry pi, and a raspberry pi will give you way less noise than a beefy gaming PC or something.
But how much this affects your DAC will depend on the level of noise and your DAC. Some DACs are incredibly susceptible to source noise, and some have full galvanic isolation on data and power and are effectively immune to even incredibly bad sources.
However you don’t need to spend tons of money. In fact you can get devices like the Intona 7055-C which provide full galvanic isolation and an incredibly low noise output for about $300.
There are of course also options like using an optical connection which inherently will not pass any noise as there is no electrical connection, it’s just light.
Grounding is also an important consideration, and some DACs WILL be affected by an ungrounded or galvanically isolated source (stuff like the ADI-2 for example is not grounded by the PSU and relies on USB ground, so can encounter significant issues when used behind a galvanic isolator), but many/most will not.
Data Integrity:
The bare minimum for a digital audio connection to be considered ‘good’ is that the data itself gets to the other end without being altered or misread. And in the vast majority of cases, this is not a problem.
Unless you are using extremely long USB connections, it’s almost certain that you’re not going to have any data integrity issues.
But even if you do, it’ll be obvious, and will usually manifest as either clicks/pops or even a static/robotic effect.
There won’t be any sort of typical audiophile descriptions like improved resolution, blackground etc. It will either work or it will be obviously problematic. So don’t fuss about this one unless you have very clear issues.
Jitter:
This is the big one, and the one that many streamers, DDCs etc will seek to improve.
The digital input method you use with your DAC and the source feeding it, can and will have a measurable impact on the performance of the DAC.
Whether it’s enough to be audible is situational, but the fact remains that it IS a genuine factor to consider.
BUT…only with sources other than USB.
With sources using I2S, AES, SPDIF (and some proprietary connections), the audio data is sent in a constant stream, with a clock signal either sent alongside it in the case of I2S, or embedded in the stream in the case of AES/SPDIF.
The DAC must convert the incoming samples/data with the timing instructed by the clock signal. And therefore, the level of jitter on the incoming clock signal has a direct impact on the timing accuracy of the DAC itself.
There are solutions such as a PLL to help attenuate jitter from an incoming signal, and these operate to varying degrees of effectiveness, but the quality of the incoming clock signal still has a measurable effect, just less so than if operating without a PLL.
This behaviour means that if you were to purchase a high performance DDC with incredibly low jitter, such as an Audio GD DI20HE or Singxer SU-6, you could conceivably gain a performance benefit if the clock signal provided by the DDC is better than the one provided by your DAC’s internal clocks. Though conversely if you picked a bad DDC, you might end up with a perfomance penalty.
But, given as the conversation above was in relation to USB, it’s important to note that Jitter from the source does not affect USB.
USB audio does not operate in the same manner as I2S/AES/SPDIF. It is NOT sent in a constant stream and a clock signal is not provided to the DAC.
With UAC2.0 audio (the standard for modern DACs), audio is sent in chunks. A chunk of audio is sent to the DAC, and this is then put into a buffer.
The DAC then takes samples from this buffer and converts them, using it’s OWN clock to determine when to do so.
More chunks will be provided to the DAC as needed to ensure that the buffer does not end up either empty (a buffer underrun) or full (a buffer overrun). The timing precision of this does not matter at all, so long as the next sample is ready by the time the DAC needs it.
So why do a lot of devices advertise ‘USB Reclocking’?
Well, to be frank, because people will buy them. Reclocking SPDIF/AES/I2S is genuine, as you are replacing the clock signal with a (hopefully) better one.
But a USB reclocker is actually just a repeater or transparent hub. And anything from a USB hub, to an innuous phoenix, to an intona 7055-C is technically a ‘reclocker’.
This CAN have the benefit of providing better signal integrity for longer distances, and that’s the main reason you’d want a USB repeater anyway. But it will not improve jitter, because as mentioned above, there is no mclk or word clock signal with USB audio and the DAC relies entirely on it’s own clocks for timing accuracy.
USB reclocking is nothing more than retransmission of the USB signal, acting as a middleman between the host device and DAC. The audio is still sent in chunks at semi-random intervals and as long as the audio is handed to the DAC before the buffer runs empty performance will be unchanged.
:
)
