2. digital
  3. 1279591648

Does the quality of a digital signal matter?

I recently heard a demonstration where a CD player was played with and without being supported with three Nordost Sort Kones. The difference was audible to me, but did not blow me away.

I was discussing the Sort Kones with a friend of mine that is an electrical engineer and also a musical audio guy. It was his opinion that these items could certain make an improvement in an analogue signal, but shouldn't do anything for a digital signal. He said that as long as the component receiving the digital signal can recognize a 1 or 0 then the signal is successful. It's a pass/fail situation and doesn't rely on levels of quality.

An example that he gave me was that we think nothing of using a cheap CDRW drive to duplicate a CD with no worry about the quality being reduced. If the signal isn't read in full an error is reported so we know that the entire signal has been sent.

I believe he said that it's possible to show that a more expensive digital cable is better than another, but the end product doesn't change.

There was a test done with HDMI cables that tested cables of different prices. The only difference in picture quality was noted when a cable was defective and there was an obvious problem on the display.

I realize that the most use analogue signals, but for those of us that use a receiver for our D/A, does the CD players quality matter? Any thoughts?
mceljo
stanwal's avatar
stanwal
3,147 posts
 
If Shadorne's argument is correct then why should the quality of the DAC or CD player matter? Even cheap one seem to measure very well. "Signal quality much less of a factor than in analog"? No wonder my LPs sound so much better. Why do transports make such a difference? A friend of mine didn't believe they would until he heard different ones on his system. Why do the best CD playback systems cost so much unless they are susceptible to degradations just as analog is?
shadorne's avatar
Ag insider logo xs@2x
shadorne
7,477 posts
 
CD's are not 0s and 1s. They are pits burnt into the metal layer and are measured for lenght by the lazer and then converted into a digital format.

These pits (or the transition) represent bits: a 1 or a 0. All digital information must be stored in analog form including what is on your computer hard drive. However, the digital approach allows the use of a threshold level or clear demarcation between a 1 and a 0 that does not exist in analog approaches.

Example of a digital scheme (not from a CD)

Signal level between -.5 and +0.5 Volts = 0. Signal level between +0.51 to 1.5 volts = 1.

This means you can have a lot of analog error or noise in the media and still get a perfect translation of the data as exactly what it should be - a 1 or 0.

If you add parity bits or polynomial redundancy check bits to the data you can also improve the robustness further (allows detection of data errors or even allowing for recovery of completely missing data)

Using the same example, compare this to an entirely analog approach where the difference between 0.O and 0.4 volts may be significant.
shadorne's avatar
Ag insider logo xs@2x
shadorne
7,477 posts
 
If Shadorne's argument is correct then why should the quality of the DAC or CD player matter?

I think I covered this:

In fact the biggest source of quality differences with digital audio is the conversion to analog - this is where differences are audible - in the quality of the D to A converter.

Clock & converter accuracy as well as the quality of the analog circuitry in the output stage can still make a difference.

However, digital eliminates the problems of media storage degradation and analog read errors from media. (dust, feedback, surface noise, pressing imperfections, pre-echo, poor chanel separation, lack of dynamic range of analog storage methods etc.)
jea48
4,451 posts
 
Just 0s and 1s?

Then there is this......
paulsax
275 posts
 
keep in mind that while digital info is just 0's and 1's there also has to be a timing element that insures that the bits are in the correct place. Easy to get them in order but if your disk is spinning too fast then your read is too fast and such. Mind I'm not advocating anything as I am suspicious of so many things here but I could see timing perhaps being affected by cheap stuff. One problem with isolation is that there is a delay between setups and the listener often knows which is which. Any blind tests out there with multiple listeners and under statistical or sampling control?
More to discover