Why do digital cables sound different?

Gmkowal just a follow on my previous response. If you don't believe me, prove it yourself. Attach a scope to the output of your DAC. Play some test tones on your transport, ( Use frequencies above 8k that's where the effect starts to be more noticable). Get some different cables and measure the signals that come out of the DAC. I have measured some pretty significant differences. I've always believed if you can measure the difference, it exists.

Blues_man - Given the inadequacies of the CD protocol and your experience in the area of digital audio data transfer interfaces, what are the benefits brought to the environment by using a high-end digital interconnect over a basic, well-built digital interconnect? That was really the question at the beginning of this post, and I think we're all still curious. -Kirk

The problem is really in the transmitter and receiver. Even though there is a "standard", there are always difference in implementation of the hardware. I started out believing that there was no difference in digital cables. That was before I was familiar with the CD standard which isn't very robust. After a while I started measuring lots of cables and interfaces just out of curiosity. Its not easy to determine which cable sounds best with a particular transport / DAC combo. You have to rely on what other people have tried. First definitely goe with AES/EBU interface over RCA cable. It's not that much more. The one I liked best was the MIT reference. Its very expensive, $800, I believe its not worth the money but if cost is no object. I'll be putting the one I tested for sale here soon (at least half off). As I posted in another thread, the best thing is to get a Transport / DAC with a custom interface. I got the Spectral because I thought it sounded best. If you have access to a scope, use the method I suggested above with a high frequency tone. Whichever cable gives the most accurate signal is probably the best.

I understand that the transmission is real time. I just do not believe a cable that is in good working order will cause a bit or two to be dropped. I agree with Blues_Man that the problem is probably in the transmitter or reciever if bits are being lost. I ran a quick test on my 3 digital cables at work today using a logic analyzer ( a scope is not the right tool for a real time test) with a pattern generator and deep memory (16 meg). I simply output a 16 bit burst every .5 seconds. The rep rate within the burst was set to 10KHz. I simply tied the pattern generator's clock to the logic analyzer's clock so that every time the pat gen's clock went high I captured the data until I filled up the memory and saved the data. I tried this with alternating 0's and 1's, all 0's , all 1's, a pattern of 0110111011001111 and it's complement. Once I had captured the data I saved them as ASCII files and I wrote a small visual basic program to look for missing bits in the patterns and found none. I also fed a repeating pattern of 0's and 1's into the cables and terminated the cable with what I approximated was the impedance of my D/A. I looked at the waveforms with a scope and looked for any droop, overshoot and undershoot. The risetime of the pulses appeared to be close on all 3 cables but I did notice some some other differences. I noticed one cable inparticular did cause more overshoot than the rest but when I varied the impedance used to terminate the cable I could minimize the overshoot (probably more capacitance in this cable causing an impedance mismatch). I marked this cable and gave all 3 cables to a another engineer who has a a separate DAC and transport to take home to see if the cables sound any different from one another. I am sorry but I did not hear very much of a difference between the cables to begin, with but I thought this would be a more subjective test. As for real time and loosing bits, the logic analyzer does not lie. I will let you know what he thinks. I can not think of a better way to test the real time data transmission characteristics of a cable. I burned up today's lunch time trying this, tommorrow I think I will have something to eat :-)Thanks for the post

One more thing, the logic analyzer used actually had 1 meg memory, not 16 meg. Someone borrowed my 16 meg aquisition module.