f various kinds of 
ssing performance. 
or to recognise the 
t that lenses may 
mally specified in 
. The sine-wave is 
me IMIN. 
  
X.- IMIN. 
(.+ IMIN. 
ation. 
orms, of whatever 
n the same series. 
r phase, when put 
aning any system 
. in the wave-form 
anges sign, corre- 
cal targets cannot 
electro-optical case 
becomes negative. 
escribe simply the 
specified independ- 
object and image. 
ion (high contrast 
1ere is no modula- 
1itrast transmission 
The chart in Fig. 3 
ce ratio. It will be 
ximately equal up 
ame: we require a 
-scan frequency is 
'k along each line. 
en the intensity is 
; for synchronising 
he bandwidth is 2 
issing 2 megacycle 
ss, then since each 
mm” in the image 
or smaller frames 
THE PHOTOGRAPHIC IMAGE, BROCK 
9 
the line-frequency for no attenuation would increase in proportion. Since the visual 
resolving-power of lenses is measured in hundreds of lines per mm, it might be thought 
that the lens introduces negligible loss into the overall television performance. As in the 
  
05 r 
LOG (LUMINANCE RATIO) 
5 
  
1 I 1 L 
  
  
a2 0.4 06 08 
MODULATION 
64 
32 
LUMINANCE RATIO 
Fig. 3. Chart relating modulation to luminance ratio and 
to log luminance ratio. 
analogous photographic 
case, however, the lim- 
iting visual resolution 
is no guide to the per- 
formance on a different 
receptor, and lenses 
considered to be good 
by visual standards are 
found to attenuate ap- 
preciable at frequen- 
cies well above their 
nominal resolving- 
power, and even at 8 
lines/mm. 
The case just discus- 
sed was over-simplified, 
because in practice one 
does not photograph or 
televise pictures of 
sinusoidal variations in 
luminance, and perfor- 
mance is usually meas- 
ured on sharp-edged 
lines. In electronic ter- 
minology the luminance distribution across the latter is a square wave (Fig. 3). However, 
there is a connection between the two. 
It is shown mathematically with the 
Fourier theorem or can be demonstrated 
graphically, that a square wave of basic fre- 
quency ‘n° can be duplicated by a sine wave 
fundamental of frequency ‘n’ plus harmonics 
at frequencies 35, 5n,.... etc; i.e. all the odd 
harmonies up to infinity. All these harmon- 
ics must start in phase and have amplitudes 
inversely proportional to their frequency. 
Although the infinite series is theoretically 
needed to give a perfect square wave (anal- 
ogous to a perfectly sharp-edge) the first 10 
or so give a close approximation and even 
the first two give a recognisable resemblance 
(Fig. 4). It is perhaps a philosophical ques- 
tion whether sharp-edged lines are actually 
made up of sinusoidal series, but the im- 
portant practical point is that they behave 
INTENSITY 
1.0 
DISTANCE 
Fig.4. High contrast line test-object 
shown as a square wave. 
as if they were. Square wave signals generated by purely electronic means can be rounded 
off and eventually turned into sine waves of the same fundamental frequency by putting 
them through appropriate low-pass filters. Photographic reproduction of test charts 
always results in a rounding-off of the edges well above the resolution limit, due to the 
loss of the higher harmonics, and the record resembles the fundamental sine wave. 
Archives 4