Handbook for Sound Engineers

810 Chapter 24

side sounds to prevent these from being added to the
delayed sound.

24.3.3 Tape Delay

A more practical early implementation method for con-
tinuous delay was to use a magnetic tape loop. An
example of such a device is shown in Fig. 24-12. The
sound is recorded onto the tape at the record head and
then is read back by one or more playback heads. The
tape then passes an erase head and loops back to the
start. The delay time is given as

(24-1)

Only the length of the tape limits the maximum delay
time. The performance of this system depends on the
quality of the recording system. Dynamic range, fre-
quency response, and SNR are affected by the tape speed
and track width. These parameters may be improved by
using the usual tape recording tricks such as compres-
sion and various forms of preemphasis/deemphasis noise
reduction. Such systems need regular maintenance
including head cleaning and replacement of the tape to
maintain optimum performance.

24.3.4 Analog Shift Register Delays

Analog shift registers as illustrated in Fig. 24-13 appear
in two forms, the bucket brigade and the analog charge
coupled device (CCD). They both operate in a very sim-
ilar manner and differ at the silicon level in the type of
switches: a metal oxide-semiconductor capacitor
(MOSC) structure for a CCD and a metaloxide–semi-
conductor junction FET capacitor (MOSJC) structure

for the bucket brigade.^2 They are classed as shift regis-

ters because they move single samples of a signal in the

form of an electrical charge from one stage to the next

in response to timing signals. The delay (T ) of a shift

register is proportional to the number of register ele-

ments (N) and inversely proportional to the frequency

(fs) of the timing signal.

(24-2)

The term charge transfer device (CTD) has been

applied to both the bucket brigade and CCD-based

structures of an analog delay. The term CCD has

become colloquially associated with a type of light-sen-

sitive array used in cameras but actually refers to the

method used to read the information off these devices.

The idea of a CTD is that it stores a sample of analog

information as a packet of charge on a capacitor and,

under control of a timing signal, transfers it to the next

storage site. All the requirements of sampling theory for

audio-frequency band limiting should be met. The per-

formance parameters for a CTD include transfer effi-

ciency (H), the fraction of charge left behind in each

transfer; the leakage of charge from a cell during the

holding period; and the leakage of charge into a cell due

to semiconductor thermal effects. Taken together, these

effects degrade the SNR of the signal as it passes

through the CTD and also lead to distortion due to the

nonlinear nature of the leakages. The practical use of

CTD is limited to applications requiring less than

100 ms of delay or longer where SNR and distortion

may be tolerated. CTDs have largely fallen out of use as

delay lines because digital systems have become

cheaper.

24.3.5 Digital Delays

Digital delay operating principles have undergone a

number of important changes since they were first used

in sound systems. Probably the most significant is in the

type of storage or memory. Shift registers were almost

universally used in the first commercially produced

Figure 24-12. Tape loop delay system.

Time distance between heads

tape speed

= --------------------------------------------------------

Figure 24-13. Bucket brigade: switched alternately open

and close to hand charges along the line.

T N

fs

--- -=