Does the quality of a digital signal matter?

Error checking and correction is very loose in CDPs since it has to read CD in real time (dust, scratches). There are programs like MAX for Mac that read music CD as data going multiple time to the same sector until finds right checksum. My CDP plays and Itunes rips CDs that MAX refuses to read or reads extremely long time.

Digital data from CDP is jittery (contains jitter - noise in time domain). Jitter creates sidebands at very low level (in order of <-60dB) but audible since not harmonically related to root frequency. With music (many frequencies) it means noise. This noise is difficult to detect because it is present only when signal is present thus manifest itself as a lack of clarity. Jitter can be suppressed by asynchronous upsampling DACs (like Benchmark DAC1) or reclocking devices. Jitter depends on quality of CDP transport and power supply. Typical digital transition of CDP is in order of 25ns making it susceptible to noise (slow crossing of threshold). High quality transports can transition many times faster reducing noise coupling but creating a lot of problems with reflections on cable characteristic impedance boundaries (therefore require better digital cable).

Jitter in D/A playback can be suppressed but recorded jitter in A/D process stays forever. For some early A/D conversions the only option is to convert it again if analog tapes still exist.

Paulsax - Jitter is a function of CD pressing quality, transport quality, digital cable quality, jitter suppressing scheme, electrical noise etc. It is a function of whole system. Even if we assume that amount of jitter is constant at given moment effects of jitter after D/A conversion are proportional to magnitude of the analog signal. Second page of Stereophile article (thank you Jea48) describes audible effects of jitter. They describe loss of detail and change in sound of instruments (harsh sounding violins) that might be effect of burying lower level harmonics in noise. Effects that they describe are often called "digititis".

Some people believe that as along as exact digital data gets to DAC timing doesn't matter. Try drawing sinewave on moving paper by marking predefined points (horizontal lines on paper to make it easier) in exact time intervals and then joining them. If intervals are not exact resulting sinewave won't be smooth - it will be jagged. Horizontal/time error got converted to vertical/value error.

Bob - Yes, error correction scheme will take care of most of the problems but used scheme (Cross Interleaved Reed-Solomon) can only correct 4000 bits of data (about 0.1"). If you have tiny scratch along the disk longer than 0.1" correction fails (only for this error). CDP won't try same sector again resulting in loss of sound quality. On the top of it transport might have poor tracking (skip track) because of CD vibrating, poor light reflection etc.

I like the fact that CDs surface can be re-polished (that's what our library did to all CDs). It tried to re-polish LP once but for some reason it didn't work.

Paulsax - IMHO everything plays role. In addition to sound of jitter that was described in mentioned Stereophile article there is digital signal processing (oversampling, non-oversampling, upsampling) and filtering algorithm, type of DAC (traditional or sigma-delta, voltage or current output, single or dual differential etc.), particular DAC selected (sound differently) and chosen update rate, type of current to voltage conversion (transformer, tube, op-amp), analog electronics (tubes, op-amps, discrete), type of op-amp or tubes, quality of components and PCB, quality of power supply. It is endless list. Even something trivial like "mute" circuit can be responsible for loss of sound quality.

At the end sound in your system is the only thing that matters.

Shadorne - CD copy should be bit perfect only if CD is copied as data and not as a music CD. For example - with I tunes I can make copy of CD that is not readable as data using MAX (with "do not allow to skip" option). You can make 10 copies of scratched CD with Itunes or similar program and every copy will be different (because many sectors might be on the edge of readability).

I use MAX for ripping but have few CDs that required use of Itunes to rip them because MAX with multiple attempts was failing.

The worse part is that Cross Interleaved Reed Solomon error correction code is approximating/interpolating uncorrectable data. Read what I found i Wikipedia under "ripping"

"CD audio has two major design constraints that make it difficult to obtain accurate copies in the form of a standard digital file. First, the system is designed to provide audio in real time in order to ensure continuous playback without gaps. For this reason, it does not provide a reliable stream of data from the disc to the computer.

Secondly, the designers felt that it would be preferable for major scratches in the disc to be covered up rather than resulting in total failure. Normally, an error correction system such as Reed Solomon would provide either a perfect copy of the original error-free data, or no result at all. However, CD audio's Cross-interleaved Reed-Solomon coding includes an extra facility that interpolates across uncorrectable errors. This means that the data read from an audio CD may not in fact be a faithful reproduction of the original.

Another practical factor in obtaining faithful copies of the music data is that different CD drives have widely varying quality for reading audio. Some drives are thought to deliver extremely accurate copies while others may do little or no error correction and even misreport error correction information."

Shadorne - I'm not sure how this interpolation works. Is it happening also when I use program that rips CD as data (like MAX) - I hope not. Do you know?