Why do digital cables sound different?

I am asking your experience with digital cables since you seem to know quite a bit about them as well as peoples perceptions of them. Surely your opinions must be based on something. Very simple Palerider.

Wire properties can certainly affect digital pulse characteristics. But, up to the point where a data "one" can be misinterpreted as a "zero" pulse characteristics don't affect the information which goes into the D/A converter.
So I don't find it "mysterious" at all that various cables sound the same. The mystery is why some folk think they sound different.

The only possibile explantion is increased jitter that you can get from passing signals between devices. I agree about the "ones and zeros" - these should be recognized properly and completely (so the bits are the same).

In this case, it may not be the fault of the cable but the way in which the components clock timing interacts.

Perhaps a device with superior PLL & timing/receiver ciruitry might sound the same with any cable when connected to a digital source (even a relatively jittery one).

Perhaps a device with poor PLL & timing/receiver ciruitry might sound different (more or less jitter depending on characteristics of signal received)...

Shadorne...If a cable causes a pulse risetime anomaly (such as a spike) it will happen for every pulse. All will be delayed or trigered early by the same amount, so jitter is not the result.

Shadorne...If a cable causes a pulse risetime anomaly (such as a spike) it will happen for every pulse. All will be delayed or trigered early by the same amount, so jitter is not the result.

What if the shape of the leading edge changes according to the signal pattern.

"Most digital audio links have a limited bandwidth. This attenuates the higher frequency components of the digital audio interface signal, and slows rise and fall times. A transition delay is introduced, and this varies depending on the data pattern in the previous time slots."

See this paper for more details.

This is probably why Ed Metiner uses "preamble jitter" for clocking, as preamble jitter is usually less. At the end of the day, since interface jitter is a fact of life then the best approach is to design a PLL clock that does not "jump around". So in the end, it is the quality of the PLL loop in your receiver that will determine how well it maintains sychronization whilst making a steady and unjittered clock. (The simplest approach being asynchronous - the data is passsed to the receiver digitally in perfect form and then receiver does D to A with its own superbly accurate clock)

The paper concludes "Digital audio equipment has to be designed to cope with jitter from its synchronisation source, as this is a characteristic of the interface. This is particularly important for the recovery of sampling clocks from interface signals."

Therefore either

One has equipment that is well designed and which can cope with jitter and all decent quality cables will all sound the same (jitter below audibility).

or

One has equipment that is not that great at rejecting jitter and cables and anything might make a subtle difference...you can blame the cables but in the end, since jitter is a fact of life - it makes more sense to blame the equipment.

Much of the jitter problems come from power supplies - so if you can clean up the power then it may be enough.

Once again - you'll notice a common theme with my posts - I tend to blame equipment or equipment interface issues for observed differences in quality rather than the interconnects and cables. (In a sense one cable may work well with one setup but not at all in another...a lot depending on the gear quality.)

Since jitter has been well known since the late 80's early 90's - my suggestion (for those who are worried that they are getting digital hash from jitter)is to

1) stick to gear made after 1995 or later (once manufacturers got robust PLL loops working well that could handle all this jitter).
2) use asynchronous approaches (like a USB DAC)
3) consider a reclocker - particularly with older classic digital gear that might be expected to be rather jittery (you know older stuff that does not upsample etc. before the engineers go their heads round all this)

Eldartford - assuming ideal square pulses spikes/distortions will happen to every pulse, but in reality signal has jiiter all the way thru with electrical noise etc. Imagine signal with slightly different rise time from bit to bit and cable with mismatched characteristic impedance. This will result in creation of staircases (transmission line efect) different at different bits since slew rate is different. Different shapes of different bits will be converted to jitter. Cable with perfectly matched characteristic impedance won't produce jitter in this scenario.

Second possibility is noise. Imagine shielded digital cable in the presence of electrical noise. Amount of jitter (caused by limited slew rate + noise) will depend on quality of shielding and will be different for different cables.

Third possibility - limited bandwidth. There is inherent noise on the top of digital signal (power supplies etc). Cable with limited bandwidth will reduce rise time and make this noise "visible" to receiver (gate will recognize level at different times on the slope of rising signal). Perfect cable with very high bandwitdh won't reduce slew rate and noise won't be visible.

I suspect that cable experts might bring many more explanations - I'm not an expert. It is, in my opinion, much more difficult to explain why power cables help with imaging but I don't question that they do. Audio is very subjective thing. How one knows that other person only "thinks" he hears difference? Trying to find scientific reasoning is a noble thing but we shouldn't question experience of others.