Digital camera technical indicators and selection knowledge

In addition to using subtitle machines or computer animation equipment to produce television images, the first step in TV program production is to use a camera to capture images. The camera is the most important device for generating television material signals. If the quality of the material signal is not high, then it is simply impossible to produce high-quality works. Generally speaking, after processing, recording, and transmission, the signal quality will always deteriorate to varying degrees, and it is difficult to improve significantly. Therefore, choosing an excellent camera is of great significance. In recent years, camera technology has made great progress. CCD camera devices have gradually replaced vacuum tube camera devices, and later digital processing cameras appeared.
Digital processing cameras have only been around for 10 years. At the beginning, digital processing cameras still had some shortcomings. Therefore, for a period of time, digital processing cameras and traditional analog processing cameras coexisted and were evenly matched. With the rapid development of technology, the functions and technical indicators of digital processing cameras have now fully surpassed those of analog processing cameras. Therefore, analog processing cameras are inevitably destined to be eliminated. This article explains the main indicators of CCD cameras, some characteristics of digital processing cameras, and the future development direction, for reference when choosing cameras.

1. Brief History of Camera Development
The earliest television cameras used electric vacuum camera tubes as camera devices. Camera tubes must be preheated before working. After a long time, the working state is prone to drift. Camera tubes themselves have multiple adjustments, such as coincidence adjustment, mechanical focus, electric focus adjustment, white-black balance, black spot correction adjustment, etc. Often after a period of use, the working state changes and must be readjusted. Camera tubes have poor vibration resistance and require a relatively complex power supply system during operation. The circuit of this part is very complex, so the failure rate is relatively high. Later, cameras with CCD
as camera devices appeared. CCD is a product of large-scale integrated circuits (VLSI). When CCD first appeared, some technical indicators, such as signal-to-noise ratio, clarity, and grayscale characteristics, did not reach the level of camera tubes, and there were some inherent defects of CCD. However, with the advancement of VLSI technology, the technical indicators of CCD devices have made great progress in recent years. Now it can be said that the technical indicators of CCD devices have comprehensively exceeded those of camera tubes. In this way, camera tube cameras have completely withdrawn from the camera market.
Digital processing cameras are developed on the basis of CCD devices. In 1989, when Panasonic of Japan launched the world's first digital processing camera AQ-20, it caused a lot of discussion. Some people believed that under the conditions at that time, the functions of analog processing cameras had reached a very perfect level, and it seemed unnecessary to adopt digital processing methods. However, after nearly ten years of practice, it has now been proved that digital processing has unique advantages that analog processing cannot match. Now, all newly launched cameras are digital processing cameras without exception.

2. Three major indicators of TV cameras
Sensitivity, resolution, and signal-to-noise ratio are collectively referred to as the three major indicators of TV cameras. They are the most important technical indicators of cameras.

1. Camera sensitivity
It is the value of the camera aperture under standard camera conditions. Standard camera conditions refer to the value of the aperture when the sensitivity switch is set to 0DB, the reflectivity is 89.9% white paper, the illumination is 2000 lux, and the image signal reaches the standard output amplitude under the lighting conditions of standard white light (iodine tungsten lamp). Usually the sensitivity can reach F8.0, and the sensitivity of new and excellent cameras can reach F11. It is equivalent to the sensitivity level of high-sensitivity ISO-400 film. In the
technical indicators of cameras, the data of the minimum illumination is often provided. When choosing, this data is more intuitive, so it has a certain value. The minimum illumination is closely related to sensitivity, and it is also related to the signal-to-noise ratio. The minimum illumination indicator of the latest camera is that the aperture is F1.4, the gain switch is set to +30DB, and the minimum illumination can reach 0.5 lux. When used under ENG conditions, you can choose a low-light camera. In this way, when shooting outside, you can reduce the requirements for lighting, and even in the evening when the naked eye cannot see clearly, you can shoot acceptable images without lighting. For studio applications, using a high-sensitivity camera can reduce the requirements for studio lighting. Lower the temperature in the studio, improve the working conditions of the cast and crew, reduce energy consumption, and save production funds.

2. Horizontal resolution
Resolution is also called clarity. It means how many vertical black and white lines can be distinguished within the range of the horizontal width equal to the screen height of the image. For example, if the horizontal resolution is 850 lines, it means that in the horizontal direction, in the center area of ​​the image, the highest ability to distinguish is the vertical black and white lines whose adjacent distance is 1/850 of the screen height.
At present, the horizontal resolution that high-end business-level cameras can achieve is 800 lines. Some cameras use pixel dislocation technology and claim to have a clarity of 850 lines. In fact, it is meaningless to pursue a very high resolution unilaterally. Due to the limited frequency band range of the signal processing system in the TV station and the signal processing circuit in the TV receiver, especially the bandwidth range of the video recorder, even if the camera has a high resolution, it will suffer losses during the signal processing process, and the final image cannot show such a high clarity.
Even if two cameras have the same resolution, if the modulation of the image signal is different, the visual effect of the image obtained will be very different. Therefore, when comparing the quality of cameras, the comparison should be made under the same modulation conditions. The higher the resolution, the better the quality. Generally, camera products do not provide modulation data. When purchasing, you should compare them to determine the quality.
The vertical clarity of the camera mainly depends on the scanning format, that is, the number of scanned lines. Therefore, the vertical clarity of the camera does not need to be considered.

3. Signal-to-noise ratio
An indicator that indicates the noise component contained in the image signal. In the displayed image, it appears as irregular flickering dots. The smaller the noise particles, the better. The value of the signal-to-noise ratio is expressed in decibels (DB). Recently, the weighted signal-to-noise ratio of the camera can reach 65DB. Observing with the naked eye, the influence of the existence of noise particles will no longer be felt.
The noise of the camera is related to the choice of gain. Generally, the gain selection switch of the camera should be set to the 0DB position for observation or measurement. In the gain increase position, the noise naturally increases. Conversely, in order to clearly see the effect of noise, you can observe in the state of gain increase. Under the same state, compare different cameras to determine the quality.
Noise is also related to contour correction. While enhancing the image detail contour, contour correction also enhances the noise contour and increases the noise particles. When performing noise testing, the contour correction switch should usually be turned off.
The so-called contour correction is to enhance the detail components in the image. Make the image appear clearer and more transparent. If the contour correction is removed, the image will appear hazy and blurred. Early contour correction only performed contour correction in the horizontal direction. Now digital contour correction is used, which performs correction in both horizontal and vertical directions, so its effect is more perfect. However, contour correction can only achieve an appropriate degree. If the contour correction amount is too large, the image will appear stiff. In addition, the result of contour correction will make the facial marks of the characters more prominent. Therefore, the new digital camera is equipped with the function of reducing contour correction in the skin color area, which is an intelligent contour correction. In this way, while improving the overall contour of the image, the face of the character is kept relatively smooth, which improves the image effect of the actor.

3. Other main indicators of the camera

1. CCD types and specifications
CCD is a photoelectric conversion device manufactured by large-scale integrated circuits. It is literally called a charge-coupled device. According to the different manufacturing processes and charge transfer methods, it can be divided into three types: FIT type - frame interline transfer, IT type - interline transfer and FT type - interframe transfer. The first two types are commonly used. The FIT type has a more complex structure, higher cost, and better performance. It is mostly used by high-end cameras. The IT type is relatively cheap, but it may produce vertical tailing. In recent years, due to technological advances, the tailing phenomenon has improved. Because of its low price, it is mostly used by business-level cameras.
According to the length of the diagonal of the CCD device, there can be different specifications such as 1/3 inch, 1/2 inch and 2/3 inch. CCD is a semiconductor device, and each unit is a pixel. The clarity of the camera mainly depends on the number of CCD pixels. Generally speaking, the larger the size, the more pixels it contains, the higher the clarity and the better the performance. Under the condition of the same number of pixels, the larger the size, the richer the displayed image level. Under possible conditions, a camera with a large CCD size should be selected, and of course the price will be more expensive. The number of pixels of advanced cameras may reach more than 600,000. The number of pixels of 2/3-inch CCD devices used in high-definition cameras can even reach 2 million.
According to the number of CCDs used in the camera, it is divided into two types: single-chip CCD and three-chip CCD. The cameras used by TV stations are generally cameras with three CCDs. RGB are imaged by independent CCDs respectively. Lower-end cameras may also use single-chip CCDs. Single-chip cameras only use one CCD device to process the three-way RGB signal. Its price is relatively low, corresponding to the relatively poor color reproduction ability. In addition, advanced cameras can also use four-chip CCDs. In addition to RGB, a CCD is also used to generate the Y signal to improve processing accuracy.

2. Grayscale characteristics
The natural scenery has a very rich grayscale level. Whether it is photos, movies, paintings or television, it is impossible to reproduce the grayscale level of nature absolutely and truly. Therefore, the number of grayscale levels is only a relative concept. Since the luminous characteristics of the cathode ray tube are nonlinear, the growth rate of the luminous amount is slow in the input low voltage area, and the luminous efficiency gradually increases as the input voltage increases. However, the photoelectric conversion characteristics of the camera device are linear (this is true for both the vacuum camera tube and the CCD device), so gamma correction must be performed in the circuit. In fact, the correction amount that the gamma correction circuit should have is inferred from the electro-optical conversion characteristics of the cathode ray tube. In order to obtain a good image grayscale characteristic effect, the main measure is to accurately adjust the gamma characteristics of the camera.
In indoor observation, the ratio of the lowest brightness to the highest brightness in the image is appropriate within the range of 1:20. If this ratio is too large, visual fatigue is likely to occur as a result of long-term viewing. Within this range, the grayscale level is around 11 levels, and a satisfactory viewing effect can be obtained.

3. Dynamic range and inflection point characteristics
Sometimes, television shooting is carried out under strong lighting conditions or in sunlight. Some reflectors reflect particularly bright light spots, and the camera will generate particularly strong signals. If it is not restricted, then the signal may be limited during the circuit processing, that is, it will be subject to white clipping. In the displayed image, there will be a pale, layerless part, which affects the visual effect of the image. In the circuit processing, the super bright part is gradually compressed so that white clipping will not appear in the subsequent processing. The super bright part in the image retains a certain degree of layering, which can greatly improve the visual effect of the image. In the amplitude relationship curve of the uncompressed input signal and the compressed output signal, it is shown that the inflection point of the curve appears at the high amplitude position. This is the inflection point processing. The proportion of the input light flux that the camera can handle exceeds the normal maximum light flux is the dynamic range of the camera. Now, the dynamic range of excellent cameras can reach 600%.

4. Number of quantization bits
The sampling of modern digital cameras generally conforms to the 4:2:2 sampling specification of ITU-R 601 (i.e. CCIR 601). That is, the sampling frequency of the Y signal is 13.5 MHz, and the sampling frequencies of the RY and BY signals are 6.75 MHz respectively. The quantization level can be 8, 10, or 12 bits. The larger the bit level, the smaller the quantization noise generated. Quantization noise is an important inherent noise source of digital cameras. For cameras used in studios, cameras with high quantization bit levels should be selected as much as possible. In addition to low quantization noise, higher processing and adjustment accuracy can be obtained in calculation and processing, and better results can be obtained. Some cameras use a sampling format of 4:4:4, or even a sampling format of 4:4:4:4. In addition to the brightness and two color difference signals, a dedicated key signal is added for use in the signal processing process. The sampling frequency of each signal is 13.5 MHz. Because cameras generally use uncompressed (bit-transparent) digital processing, the signal bit rate is very high and the signal processing speed requirement is very stringent.

5. Neutral filter
New cameras are sometimes equipped with multiple neutral filters. The function of the filter is to reduce the light flux. This is because, in strong light conditions, due to the effect of the automatic aperture, the aperture will become very small, the resulting image will appear relatively stiff, and the lens will not work in the best state. Use an appropriate neutral filter to make the automatic aperture open a little wider, the image will appear softer, and the overall effect of the TV image will be improved.

6. Lens selection
Modern cameras all use zoom lenses. Lenses with different zoom ranges should be selected according to the actual use occasion. If used to capture conference images, a short focal length zoom lens must usually be selected to facilitate wide-angle images. If used to capture outdoor images and long-distance photography, for example, to capture wildlife, or to take long-distance candid photos, a long focal length zoom lens should be selected. You can choose an appropriate telephoto magnifier or wide-angle magnifier. For lenses used in ENG, the focal length of wide-angle lenses can now be less than 4.8 mm (lens model A10×4.8BEVM-28). For specific cameras, the image size depends on the size of the CCD, and the image distance is also determined. Therefore, the camera angle can be calculated based on the shortest focal length. When the shortest focal length is 4.8 mm, the camera angle is about 85 to 90 degrees (related to the size of the CCD). The focal length of the telephoto lens can be greater than 700 mm (lens model A36×14.5BERD-R28). Sometimes, a camera can have two or more lenses as spares. According to the actual needs, the corresponding lenses can be replaced at the appropriate time. There is also a lens device called a fisheye adapter, which is very suitable for use in certain specific occasions (such as candid photography). Another important parameter of the lens is the maximum relative aperture. The larger the relative aperture, the smaller the distortion and the higher the efficiency of light utilization. Excellent large-aperture lenses have high resolution not only in the central area, but also in the edge area, with small image and color distortion.
Color reproduction ability is also an important feature of the camera. However, it is difficult to explain it with test indicators, and general camera manufacturers do not provide indicator data on this performance. It can usually be judged by comparison based on the actual observation effect.
For electric vacuum type camera devices, there is also an indicator of RGB coincidence accuracy. For CCD cameras, this indicator is no longer a problem.

4. Advantages of digital processing cameras
The following briefly describes the characteristics of digital processing cameras.

1. Suitable for computer processing
In modern digital processing cameras, microprocessors (MCPs) are generally used as central processing components to achieve control, adjustment, and calculation functions, and a variety of dedicated large-scale integrated circuits are used to greatly enhance the processing capabilities and automation functions of the camera.

2. Simplify the adjustment mechanism and adjustment method
Most analog cameras use adjustment components (potentiometers, adjustable capacitors, coils, etc.) for adjustment. The adjustment components of many cameras are located on the circuit board. Therefore, the shell must be opened to perform the adjustment operation, which is very inconvenient. Once an analog processing camera is adjusted incorrectly, it is very difficult to restore it to its original state. Therefore, the adjustment of analog processing cameras is generally performed by experienced technicians. Even so, they must act cautiously. If the adjustment is messed up, it will be very troublesome to restore it to its original state. Digital
processing cameras use menu display and increase and decrease adjustments by buttons. In this way, from the user's point of view, the work that originally had to be carefully adjusted by skilled technicians can now be adjusted by general technicians or even cameramen. The adjusted data is saved in the memory in the form of a file. If you are not satisfied with the data you have adjusted, you can call out the data when the machine left the factory (standard) or before this adjustment for comparison, so there is no need to worry about messing up the data due to lack of experience. The operator can make repeated adjustments with confidence to obtain satisfactory results. From the manufacturer's perspective, it is easy to automate the manufacturing and debugging processes, improve production efficiency and reduce costs.

3. Accurate and detailed adjustment can be achieved.
In essence, digital processing cameras transform the signal, converting the original analog signal into symbols represented by 0 and 1 codes. In this way, digital processing cameras process symbols rather than the analog signal itself. Symbols have the following main characteristics:
(1) Analog signals will inevitably produce distortion and noise during processing and transmission. Once distortion and noise occur, the signal quality is reduced and it is difficult to recover. Digital signals have excellent stability and are not prone to distortion during storage and transmission. In the digital field, the problem that may arise is bit errors. Certain measures, such as channel coding and forward error correction, can effectively overcome bit errors. Of course, when designing circuits, caution should be exercised and the system should not be allowed to approach the collapse zone. Signal regeneration should be performed at appropriate times. Otherwise, the bit error will be too large to be corrected and will be out of control. However, with proper design, this situation can be completely avoided.
The essence of analog processing is signal replication, which is accompanied by signal distortion. The essence of digital processing is signal regeneration, so it can accurately reproduce noise-free and distortion-free signals.

(2) It is difficult to adjust a specific amplitude or frequency characteristic of an analog signal. To adjust a certain characteristic, other characteristics may change. On the other hand, if several parameters need to be adjusted in coordination, it will be very difficult. In digital processing, a certain part of the signal characteristics can be adjusted separately, for example, gamma correction adjustment and inflection point adjustment can be made separately; a certain frequency or a certain amplitude level can be adjusted separately, while the characteristics of other parts of the signal remain unchanged. In addition, several adjustments can be combined and correlated to make coordinated adjustments. For example, in the application process, it is often necessary to make related coordinated adjustments to chromaticity, contrast, brightness, etc. Under analog conditions, this is almost impossible, but under digital conditions, this type of function can be realized relatively easily.
Some people say that in the digital field, there are only unimaginable functions, but no unrealizable functions. The actual situation should be that as long as it is a reasonable idea, digital technology has the ability to realize it. Even though the computing power at that time was still difficult, it was eventually realized with the development of algorithms, storage, and computing speed.

(3) High processing accuracy. For 10-bit quantization, it can be divided into 1024 levels for adjustment. High-end cameras can achieve 12 or 14 bits, which is equivalent to 4096 or 16384 levels, or even higher levels of processing. To make such fine adjustments, analog processing methods are powerless and can only sigh in despair.

5. Disadvantages of Digital Processing Cameras

1. Conversion loss
Starting from the camera device (whether it is a camera tube or CCD), the signals generated are all analog signals. After a series of processing and transmission, analog signals must still be used when displaying images on the picture tube. It can be seen that the so-called digital processing first converts analog signals into digital signals (ADC), but in the intermediate processing and transmission process, the form of digital signals is used. Finally, the digital signals still need to be converted into analog signals (DAC). Analog-to-digital conversion and digital-to-analog conversion will produce quantization error distortion. In the process of signal processing, various operations must be performed, and rounding and carry errors will be generated during the operation process. If the code stream conversion is also required, new conversion errors will be generated. The results of these errors are superimposed one after another, forming an overall loss. Therefore, any digital system should avoid various types of signal conversion as much as possible.
Increasing the number of quantization bits and the number of bits in signal processing can reduce these errors. The earliest digital processing camera AQ-20 uses an 8-bit quantization level, compared with a 256-level quantization level. The resulting quantization loss cannot be ignored. If the number of quantization bits is increased to 10 bits, the quantization level can reach 1024 levels, and the corresponding quantization error can reach 60dB, and its noise is actually close to being negligible. However, considering the accumulation of rounding and carry errors in calculations, excellent cameras usually use 12, 14 or even higher bit signal processors.
Cameras usually use bit transparent processing, that is, non-compressed, full-bit processing. Therefore, there is no quality loss caused by compression and decompression. If the signal is stored, processed, and transmitted using a compressed method, the quality loss caused by compression and code stream conversion should also be considered.

2. High bit rate brings difficulties to signal processing
According to the sampling parameters recommended by ITU-R 601 (CCIR 601), that is, the 4:2:2 sampling method, the sampling frequency of the Y signal is 13.5 MHz, and the sampling frequencies of the RY and BY signals are 6.75 MHz respectively. If 8-bit quantization is used, the data bit rate of the video signal can be calculated:
(13.5+2×6.75)×8=216 Mbps
Deducting the invalid information in the blanking area:
216×(64-12)/64×(312.5-25)/312.5=161.46 Mbps/s.
Because the amount of synchronization information is small, it is ignored. The information of the sound signal is not considered here.
It can be seen that when the number of quantization bits is 8, the bit rate of the video signal is still as high as 161 Mbps. In other words, the camera must process 160 million bits of data per second. If quantization levels of 10, 12, or 14 bits are used, the amount of data will increase dramatically in proportion. As can be seen, the amount of data processed by digital processing cameras is extremely large. Moreover, all processing in cameras must be done in real time. Therefore, very stringent requirements are placed on the computing power and speed of the built-in IC and microprocessor.

3. High power consumption
As mentioned above, digital processing cameras must process extremely large amounts of data at high speed, and each step of the operation is accompanied by energy consumption. Therefore, the temperature rise of early digital processing cameras was very high. In particular, the VLSI on the main processing circuit board was so hot that it became hot. Obviously, power consumption causes temperature rise, which is the root cause of instability and failure. High power consumption increases the burden on the power supply. In many cases, the camera is powered by a rechargeable battery. Due to the high power consumption, the same battery can only work for a short time, which brings many inconveniences to the actual work. As the functions of the camera become more and more powerful, the workload of signal data processing becomes larger and larger, and the energy consumption becomes larger and larger. On the other hand, due to the improvement of algorithms and the advancement of integrated circuit technology, power consumption continues to decline, which is a dialectical development process.
It can be predicted that the shortcomings of existing digital cameras will be gradually overcome with the development of technology.