Frequently Asked Questions About Compact Discs

Please contact Media Sciences if your questions are not answered on these pages.

CD jitter test equipment seems to be expensive. Correlation problems also seem to exist. Is jitter really important? Is there an inexpensive way to measure jitter?

CD-ROM discs contain physical pits and lands. Red Book parametric measurements (HF, asymmetry, push-pull, radial noise) and error rates (BLER and Exy) accurately evaluate pit geometry. If these results are good, jitter is usually low and may not require measurement.

CD-R discs only imitate pits. Various dyes produce different simulation patterns. Experience shows that all parametric measurements and error rates may be satisfactory for a CD-R disc yet jitter may be unacceptable. Jitter therefore is an important test for CD-R discs.

Correlation can be difficult for many tests including jitter. Three problems can result in poor correlation. First, differences in players and especially in the laser beam and optics can give different results. For example, results from a player having an ideal circular spot where the beam is focussed onto the track may not agree with a player where the spot is elliptical with the long axis in the direction of the track. An elliptical spot with the long axis perpendicular to the track can give results that differ from the previous two examples. Second, test electronics can affect test results. Third, test software can have an impact.

A player is always required for jitter measurements, and Philips Test Sample 5B.2 can provide jitter and length deviation standards. An analog oscilloscope is useful for jitter estimates, and can help to identify electronics or software correlation problems. It also may be a less expensive instrument, although a test player is still required. Note that the following procedure only provides estimates for jitter, and cannot replace precise jitter testers. It only provides one value for jitter, not nine values for pit jitter and another nine for lands, and does not give values for length deviation.

Begin with a 100 MHz oscilloscope or better if available. Although 100 MHz high frequency response is not required, its higher intensity and better phosphor are beneficial to the measurement. Remember that we are measuring random variations, not a stable signal. Ground the scope input, and adjust focus and astigmatism for a small, circular spot. Connect the vertical input of the scope to the HF output of the player used for testing. Trigger from the vertical channel, and adjust triggering to observe the classic eye pattern of the EFM signal. Set the sweep rate to about 200 nsec/div for a 1X player, proportionatly less for higher speeds, and adjust intensity to obtain a normal, bright display. Although the eye pattern will seem fuzzy, do not touch astigmatism or focus. You are now observing jitter!

Examine the pattern carefully near the decision level. This is near the vertical center of the eye pattern where diamonds, or cats eyes, are visible. Now concentrate on the "X" crossing of two traces, one decreasing (a land-to-pit transition) and the other increasing (a pit-to-land transition) while sweeping to the right. Sweep magnification may be helpful. With a normal, bright beam, estimate the horizontal width of one trace near the "X" crossing. Designating this width as W, one standard deviation jitter can be estimated by dividing W by 3.6. This method is very simple and surprisingly accurate.

The value 3.6 is not magical but is derived from two established principles. First, visually observable extremes on an analog oscilloscope having a normal persistence phosphor and typical brightness will include about 99% of a population (remember that white noise theoretically can have rare variations of infinite amplitude). CD jitter is usually Gaussian, therefore 99% of the population occurs 2.57 standard deviations either side of the mean. The width, W, that we measured is thus twice 2.57 or 5.14 standard deviations. Therefore the standard deviation of a pit/land transition is W divided by 5.14. This relationship can be used to estimate DVD bit jitter.

Second, time interval analysis (TIA) is conventionally used to determine CD jitter. This measures the variations of a time interval between two successive pit/land transitions. Result W represents only one pit/land transition on the oscilloscope. TIA jitter and W are easily related since the variation between two events is the square root of the sum of the squares of the variations of each event. Since both pit/land transitions have the same variations, TIA jitter is greater than the pit/land variations by the square root of two, or 1.414. The final result is that one standard deviation TIA jitter for a CD is 1.414 times W divided by 5.14, or just W divided by 3.6.

Remember that this method only estimates jitter, and is not a precise measurement. It is a proven method that can help to resolve correlation issues, and provides realistic estimates of jitter when sophisticated equipment is unavailable.