Jitter, CDPs, Transports and Streaming

I got to thinking that DACs have buffers that they read from and realized that the upstream source shouldn’t matter, but they apparently do. Why doesn’t the buffer completely eliminate the relevance of the quality of the source? Are there types of DACs, like asynchronous DACs, that make the CD transport or computer source quality irrelevant?

It depends on the way the buffer works. The buffer in the Benchmark DAC3 for instance is not a resampler or a PLL receiver. I have not personally heard it, but customers have told me that putting my Synchro-Mesh reclocker ahead of it has no effect on it, indicating that it may actually reject most jitter. I have no idea how good these customer systems are however, so I may still hear a difference in my own system.

Making the source jitter irrelevant is a good idea, but only if the internal DAC clocks and circuits used deliver really low jitter also. This is in no way guaranteed. Those circuits can easily take a 1psec jitter spec oscillator and convert it to 300psec and they usually do.

This is exactly why I prefer DAC’s that have no reclocking on the S/PDIF inputs. This way you can drive those inputs with lower and lower jitter sources and reap the benefits.

Steve N.

Empirical Audio

There is always a small buffer of a few words, but it is clocked synchronously with the clock recovered from the incoming stream by the "Receiver" chip. The clock is contained in the S/PDIF data stream and it is recovered by this chip.  Some of the newer Receiver chips can reduce jitter even more if they are implemented on the circuit board optimally.

All DAC's work basically in "real-time", even if they have asynchronous buffers.  The buffer delay is much smaller than anything audible.

Steve N.

Empirical Audio

Every source, whether it's a Transport or a computer server will sound different because the clock circuit designs are different.  Some are better than others, so they will have lower jitter.  A DAC that reclocks asynchronously will usually be affected by this jitter. I have only evidence of one DAC, the Benchmark DAC3, that is generally immune to incoming jitter, but this conclusion is only anecdotal based on customer feedbacks.  I have not tested this myself.

Steve N.

Empirical Audio

Reclockers or reclocking designs, chips, circuits and said implementations of such... can make for worse jitter going in (to the dacs proper), than was originally received from the given source point, so one has to be careful. Re-clocking is not a wonder ointment cure-all. It is generally considered to be a good idea but it does, or can... create a brick wall jitter reduction effect, that the reclocker is then crowned the king of jitter issues for that system.

There are certainly poor reclockers and good ones. Reclocking is a great cure-all actually, if it is done right. Measuring the jitter directly at the end of a 4 foot cable is not what most vendors do, because it requires specialized, expensive equipment. This is what I do, so my measurements speak for themselves. Many, many customer feedbacks also speak volumes. You can read all of these yourself.

The other benefits of reclocking are usually glossed-over, including input and output galvanic isolation and upsampling to 24/96 so the DAC uses a better digital filter. These are all benefits of the Synchro-Mesh.

Among my own products, which all deliver extremely low jitter, less than 10psec at the end of a 4 foot cable, my Synchro-Mesh reclocker actually delivers the lowest jitter. The artifacts created by upsampling to 24/96 by the Synchro-Mesh are minimal too, really insignificant compared to eliminating the jitter. Most don’t hear anything due to the upsampling, just better SQ.

Steve N.

Empirical Audio