Sony 900 사용자 설명서

9

Chapter 9 Glossary – Terms and Definitions

174

HDC-900/950/930 Series Product Information Manual

compared in numbers, it does not appear in the 
specifications of the brochure. However, it is very 
important for evaluating picture quality intuitively. 
Usually, it is only called “DTL (detail)”. 

In an IT type CCD, beside the pixels there are columns 
of cells dedicated for vertical transfer. The electronic 
signals are transferred from the pixels to the columns 
of cells during vertical blanking. Finally, the signals 
reach a row of cells driven by a horizontal block for 
horizontal transfer, from where the signals are output 
as video signals. Although there is a cover to protect 
the cells from light leakage, sometimes the leakage still 
happens under a strong light, which causes smear. 

To avoid this phenomenon, in an FIT type CCD, 

there is  additional storage added to the CCD which 
enables the vertical transfer to perform much more 
quickly so to minimize the chance of light leakage even 
under a strong light. However, this additional storage 
increases the cost. 

Recently, there has been great advances in IT CCD 

technology and the gap between IT and FIT has been 
dramatically reduced. 

Contrasts always become worse when compressing in 
the direction of the bright level during knee correction. 
Knee aperture is used to improve this point to 
emphasize the contour to which the knee correction 
applies. 

The dynamic range of a natural scene is very large, but 
that of a video standard is limited. When we use the iris 
to adjust for the subject photographically, the bright 
portion of the picture will become almost completely 
white. However, when we focus on the bright portion, 
the subject for photography will become very dark. 
Knee correction is a function to solve this problem by 
compressing the level of the bright area to meet that in 
the video standard so as to make the dynamic range 
seem larger. 

Knee point is the starting point of the knee correction 
process, and the knee slope is its compression rate. 
Usually, knee slope is the peak value under an 
environment of 200% incident light. 

Because knee correction performs independently in R/
G/B channels in a conventional knee circuit, the R/G/B 
percentage of the portion to which the knee correction 
applys will change, and thus the color phase and 
saturation change. This is called “white clip” or “wash 
out.” Knee saturation is used to balance the coloring to 
avoid this phenomenon.

Just like Crisp reduces noise, level depend is a circuit 
that decreases the DTL value in dark areas. 

A limiter is a circuit used to constrain the DTL peak 
value that may ruin the picture quality when the 
luminance difference is too big that it will generate too 
large of a DTL.

The linear matrix is actually a circuit used to make the 
color reproduction look better, although some 
textbooks mention that it is used to correct the 
difference between camera chromaticity and that of 

the signal standard. It seems that the linear matrix 
should be more important to emphasize the visual 
appeal of the picture apparently rather than just be 
theoretically determined by some coefficients.

Recently, most of the cameras create V DTL from all 
three channels of R/G/B. Depending on how to apply 
these three values of V DTL to the add operation, there 
are two types of mix: the Mix DTL – a simple sum of the 
values and the NAM DTL – a non-additive mix of the 
values. Both of them have pros and cons. The Mix DTL 
is not always better than “out of green”, while the NAM 
DTL helps make a picture sharper but sometimes it 
makes the S/N become worse or adds the DTL 
opposite to the luminance. The NAM DTL does not 
apply to H DTL, because it will increase aliasing. 

Multi matrix is used to easily implement color 
reproduction that complies with color matching and 
visual appeal between different cameras by adjusting 
the matrix coefficient to match the color phase. For 
instance, it is possible to change the color phase and 
saturation of red color only. In most of Sony’s cameras, 
it works with 16 axes for color phase on a color plane 
like a vector, each of which has its own matrix 
coefficient to perform the linearity corrections between 
each axis. Normally, each axis needs six matrix 
coefficients (that means 96 matrix coefficients are 
needed). However, the number of coefficients can be 
reduced to two for each axis in most of Sony’s cameras 
using some dedicated software - color phase and 
saturation. 

To obtain a better quality picture, -the more pixels the 
better for a CCD. However, increasing the number of 
pixels means decreasing the pixel size. When the pixel 
size becomes smaller, less incident light can be 
gathered, which degrades the sensitivity. The on-chip-
lens is a micro lens put on each pixel to help increase 
light gathering so to improve the sensitivity. 

There are two types of optical filters, Neutral Density 
(ND) filter and Color Temperature Conversion (CC) 
filter. The former is like a pair of sunglasses for the 
camera to avoid the iris to be stopped down too much 
under a very strong light, while the latter works like a 
pair of colored glasses to compensate for the color 
temperature beyond the white balance compensation 
range. Both of them are integrated on a built-in filter 
disk, can be manually or electronically controlled and 
can be remotely controlled using the remote controller 
for the electronic filter. 

Shooting with a CCD camera means sampling a 
picture via the use of the pixels on a CCD. An optical 
low pass filter is made of optical crystal to cut off the 
unnecessary high frequency that might cause beat 
interference caused by the relation between the 
number of pixels and the pattern of the shooting 
subject during the sampling. 

The older version cameras create V DTL from the G 
channel and add it to each channel to carry out image 
enhancement. This method helps use a less number of 
delay lines, but does not perform well in some cases 
such as a deep red flower picture. Some cameras 
create V DTL from R/G channels.