- Joined
- May 1, 2013
- Messages
- 10,668
- Reaction score
- 6,561
- Location
- Dublin
- Country
- Ireland
- Dash Cam
- Street Guardian UK / EU / EEA Distributor
About CMOS technology this Russian article below will help you to understand more. There is a sample OV vs Aptina, but explanation rules about pixel size and all other aspects apply to all CMOS dash cams.
http://translate.google.ie/translate?sl ... -sensor%2F
CMOS-sensor OmniVision OV2710 and Aptina AR0330
On this page of our website we will try to explain to customers what the usual "matrix" of how it works, features and differences of the different matrices. Yes, let us forgive professionals if our presentation will seem to them too primitive, and the terms are incorrect. We do not aim to write a scientific article, but we want to just communicate clearly convey the essence of the theme handily.
Now there are new modern matrix is significantly more sensitive and "low-noise". They are able to generate enough high-quality video at night compared to the old matrix. OmniVision OV2710 and Aptina AR0330 U.S. production. This matrix has been successfully used in professional video surveillance systems. More recently they have started to establish and car video recorders and video cameras in the household.
Unfortunately, consumers are not entirely true, from what really determines the quality of the recording car DVR. Advertising has done its job, and many buyers are looking for megapixels. Almost all of the old matrix had 5MP resolution, and new 2.1Mp or 3.5Mp. And it is very difficult to explain to the buyer that the matrix with a resolution of 2.1Mp can generate higher image quality in low light conditions than the 5-megapixel.
Matrix, video matrix, sensor, video sensor - the main part of any camcorder or camera. It is this element converts light into electricity. There are two production technologies matrices. CCD (CCD) and CMOS (CMOS). Without going into details, we note that a good CCD-matrix better CMOS-matrix. But megapixel performance is very expensive and are used only in professional equipment. Therefore, producers and video cameras in the budget price range set only CMOS-matrix. In addition, CMOS technology is developing very fast. The difference in quality between the two technologies is not that big. And the price difference is huge.
How does the CMOS-matrix?
Many people know what a solar panel and saw its active element, which converts solar energy into electricity. CMOS-matrix technology is based on the same principle. The entire surface of the matrix consists of a miniaturized active elements - pixels. Each pixel in response to light, generates an electrical charge. Then, the charge is read out and processed to form a frame. Reading of the electric charge runs a special electronic circuit. We call this scheme "electronic shutter". On a bright sunny day, the charge in each pixel accumulates very quickly. Enough 1/1000 sec -1 / 10000. At night - on the contrary, the charge builds up very slowly. Typically, in a dark charge accumulates during 1/25 - 1/30 sec. This limitation is due to the frame rate. If we have a second 30 frames, then charge from the pixel to be read 30 times per second. Unfortunately, such a large charge accumulation time in the pixels in the 1/30 sec., It time to accumulate and "electronic" noise. Therefore, the image at night as "noise." Furthermore. Imagine that the camera is mounted on a car and driven at a speed of at least 30 km / h For 1/30 sec, the car can be several tens of centimeters. That's why, when driving at night image camera can turn out blurred. In order to be an image at night with minimal noise was and was not spread out, it is necessary to do so, that would charge accumulated in the pixel for a faster time. Not for 1/30 sec. And for example, 1/100- 1/120 sec. Of course, you can force to reduce the exposure time, but then the accumulated charge is not enough, and the image is very dark.
Manufacturers are addressing this problem in several ways:
Increasing the size of the pixel. More surface - will get more light.
Improving the quality of the material is done pixel.
Installing the front of each pixel of the microlenses. Allows you to more carefully focus the light output.
Installation of microlenses between pixels. Allows you to focus on the lenses additional light output.
The use of technologies that improve the "shutter".
You can also greatly improve the quality of the night recording quality using high-aperture lens in front of the matrix (as is done in TrendVision TV-Q2). But it depends on the manufacturers of car DVR, not matrices.
Key Features megapixel CMOS-matrix.
Aptina MT9P01 Aptina AR0330 OmniVision OV2710
Resolution 5Mp 3.5Mp 2.1Mp
Number of Pixels 2592H x 1944V 2304H x 1296V 1920H x 1080V
The size of the matrix 1/2.5 " 1/3 " 1/3 "
The pixel size 2.2μm x 2.2μm 2.2μm x 2.2μm 3.0μm x 3.0μm
Sensitivity 1.4 V / lux-sec 1.9 V / lux-sec 3.3 V / lux-sec
Dynamic range 70dB 72.4dB 69dB
Signal / noise 38dB 40dB 39dB
1. Resolution. Depends on the number of pixels on the sensor. The greater the number of pixels, the higher the resolution.
Most car DVR installed 5-megapixel American firm Aptina MT9P01. It has a pixel 2592H x 1944V. But it is worth considering that the video with a resolution of 1920x1080 FullHD enough pixel (2.1 megapixel). Historically, the first Car DVR FullHd with a resolution of 1920x1080 appeared by further developing common household miniature video cameras. These cameras budget price range do not place high demands on the quality when shooting in low light. In addition, for promotional purposes, such cells should be more and cameras. It is therefore established a matrix of 5 MP. The camera is able to make a video with a resolution of 2.1 megapixels and photos with a resolution of 5 megapixels. However, in the vast majority of car DVR set wide-angle optics, which does not have an optical resolution of 5 megapixels, the pictures were obtained poor quality. For video recording, the whole matrix is not used, and only the central part 1920H x 1080V pixel (2.1 megapixel).
For video recording in FullHD format enough matrix 2.1 megapixels (1920x1080 multiply = 2073600).
2. The size of the matrix. Usually expressed in inches diagonally. Clearly, the larger the matrix, the larger the size (area) of a single pixel. And this means more light falls on it.
Aptina MT9P01 old matrix has a size of 1/2.5. "This is a big matrix. But when you consider that this area is 5 million pixels, each pixel is obtained does not seem like much, just 2.2 microns.'s Large image sensor - but a small pixel. A Only if the active zone take the matrix of 1920 x 1080, to form the necessary signal, the size of this area will be less than 1/4 ". Enough. Conversely, on a smaller matrix 1/3 "OV2710 is only 2.1 million pixels (remember, this is enough to form a frame FullHD resolution of 1920 x 1080), and each pixel is obtained fairly large - 3.0 micron. And in the matrix uses the entire area.
Large, but 5-megapixel uses a video signal only part of the area the size of 1920x1080 pixel.
3. The size of the pixel. As has been stated earlier, the larger the surface area of the pixel, the more it gets light rays. Hence the matrix rapidly accumulate an electrical signal. There is a chance to reduce the exposure time, thus reduce the level of noise and the blur in low light.
4. Sensitivity. Depends on the size of the pixel. The higher it is, the more sensitive the matrix. As can be seen from the table, OV2710 has the largest and the largest pixel sensitivity. AR0330 also has a higher sensitivity than the old matrix, although the size of the same - 2.2μm x 2.2μm. But this is achieved by new technologies - improved microlenses in front of the pixels of the new system to remove the charge and other technologies.
5. Dynamic range and S / N ratio in all three matrices is roughly the same.
Our research and testing has shown that the new matrix OmniVision OV2710 and APTINA AR0330 are about equally good quality recording in low light conditions. OV2710 - due to the large pixel in size. AR0330 - by using the latest technologies. But the two matrices have a distinct advantage over the old 5-megapixel Aptina MT9P01 , which is still used by most manufacturers of automotive DVRs.
It should be noted that the video quality also depends on the lens that is mounted in front of the matrix. Cheap lens, which lets a little light, can easily bring all the benefits of a good matrix to the minimum. Conversely, a good fast lens can significantly improve the quality of video images at night, even with the old matrix.
http://translate.google.ie/translate?sl ... -sensor%2F
CMOS-sensor OmniVision OV2710 and Aptina AR0330
On this page of our website we will try to explain to customers what the usual "matrix" of how it works, features and differences of the different matrices. Yes, let us forgive professionals if our presentation will seem to them too primitive, and the terms are incorrect. We do not aim to write a scientific article, but we want to just communicate clearly convey the essence of the theme handily.
Now there are new modern matrix is significantly more sensitive and "low-noise". They are able to generate enough high-quality video at night compared to the old matrix. OmniVision OV2710 and Aptina AR0330 U.S. production. This matrix has been successfully used in professional video surveillance systems. More recently they have started to establish and car video recorders and video cameras in the household.
Unfortunately, consumers are not entirely true, from what really determines the quality of the recording car DVR. Advertising has done its job, and many buyers are looking for megapixels. Almost all of the old matrix had 5MP resolution, and new 2.1Mp or 3.5Mp. And it is very difficult to explain to the buyer that the matrix with a resolution of 2.1Mp can generate higher image quality in low light conditions than the 5-megapixel.
Matrix, video matrix, sensor, video sensor - the main part of any camcorder or camera. It is this element converts light into electricity. There are two production technologies matrices. CCD (CCD) and CMOS (CMOS). Without going into details, we note that a good CCD-matrix better CMOS-matrix. But megapixel performance is very expensive and are used only in professional equipment. Therefore, producers and video cameras in the budget price range set only CMOS-matrix. In addition, CMOS technology is developing very fast. The difference in quality between the two technologies is not that big. And the price difference is huge.
How does the CMOS-matrix?
Many people know what a solar panel and saw its active element, which converts solar energy into electricity. CMOS-matrix technology is based on the same principle. The entire surface of the matrix consists of a miniaturized active elements - pixels. Each pixel in response to light, generates an electrical charge. Then, the charge is read out and processed to form a frame. Reading of the electric charge runs a special electronic circuit. We call this scheme "electronic shutter". On a bright sunny day, the charge in each pixel accumulates very quickly. Enough 1/1000 sec -1 / 10000. At night - on the contrary, the charge builds up very slowly. Typically, in a dark charge accumulates during 1/25 - 1/30 sec. This limitation is due to the frame rate. If we have a second 30 frames, then charge from the pixel to be read 30 times per second. Unfortunately, such a large charge accumulation time in the pixels in the 1/30 sec., It time to accumulate and "electronic" noise. Therefore, the image at night as "noise." Furthermore. Imagine that the camera is mounted on a car and driven at a speed of at least 30 km / h For 1/30 sec, the car can be several tens of centimeters. That's why, when driving at night image camera can turn out blurred. In order to be an image at night with minimal noise was and was not spread out, it is necessary to do so, that would charge accumulated in the pixel for a faster time. Not for 1/30 sec. And for example, 1/100- 1/120 sec. Of course, you can force to reduce the exposure time, but then the accumulated charge is not enough, and the image is very dark.
Manufacturers are addressing this problem in several ways:
Increasing the size of the pixel. More surface - will get more light.
Improving the quality of the material is done pixel.
Installing the front of each pixel of the microlenses. Allows you to more carefully focus the light output.
Installation of microlenses between pixels. Allows you to focus on the lenses additional light output.
The use of technologies that improve the "shutter".
You can also greatly improve the quality of the night recording quality using high-aperture lens in front of the matrix (as is done in TrendVision TV-Q2). But it depends on the manufacturers of car DVR, not matrices.
Key Features megapixel CMOS-matrix.
Aptina MT9P01 Aptina AR0330 OmniVision OV2710
Resolution 5Mp 3.5Mp 2.1Mp
Number of Pixels 2592H x 1944V 2304H x 1296V 1920H x 1080V
The size of the matrix 1/2.5 " 1/3 " 1/3 "
The pixel size 2.2μm x 2.2μm 2.2μm x 2.2μm 3.0μm x 3.0μm
Sensitivity 1.4 V / lux-sec 1.9 V / lux-sec 3.3 V / lux-sec
Dynamic range 70dB 72.4dB 69dB
Signal / noise 38dB 40dB 39dB
1. Resolution. Depends on the number of pixels on the sensor. The greater the number of pixels, the higher the resolution.
Most car DVR installed 5-megapixel American firm Aptina MT9P01. It has a pixel 2592H x 1944V. But it is worth considering that the video with a resolution of 1920x1080 FullHD enough pixel (2.1 megapixel). Historically, the first Car DVR FullHd with a resolution of 1920x1080 appeared by further developing common household miniature video cameras. These cameras budget price range do not place high demands on the quality when shooting in low light. In addition, for promotional purposes, such cells should be more and cameras. It is therefore established a matrix of 5 MP. The camera is able to make a video with a resolution of 2.1 megapixels and photos with a resolution of 5 megapixels. However, in the vast majority of car DVR set wide-angle optics, which does not have an optical resolution of 5 megapixels, the pictures were obtained poor quality. For video recording, the whole matrix is not used, and only the central part 1920H x 1080V pixel (2.1 megapixel).
For video recording in FullHD format enough matrix 2.1 megapixels (1920x1080 multiply = 2073600).
2. The size of the matrix. Usually expressed in inches diagonally. Clearly, the larger the matrix, the larger the size (area) of a single pixel. And this means more light falls on it.
Aptina MT9P01 old matrix has a size of 1/2.5. "This is a big matrix. But when you consider that this area is 5 million pixels, each pixel is obtained does not seem like much, just 2.2 microns.'s Large image sensor - but a small pixel. A Only if the active zone take the matrix of 1920 x 1080, to form the necessary signal, the size of this area will be less than 1/4 ". Enough. Conversely, on a smaller matrix 1/3 "OV2710 is only 2.1 million pixels (remember, this is enough to form a frame FullHD resolution of 1920 x 1080), and each pixel is obtained fairly large - 3.0 micron. And in the matrix uses the entire area.
Large, but 5-megapixel uses a video signal only part of the area the size of 1920x1080 pixel.
3. The size of the pixel. As has been stated earlier, the larger the surface area of the pixel, the more it gets light rays. Hence the matrix rapidly accumulate an electrical signal. There is a chance to reduce the exposure time, thus reduce the level of noise and the blur in low light.
4. Sensitivity. Depends on the size of the pixel. The higher it is, the more sensitive the matrix. As can be seen from the table, OV2710 has the largest and the largest pixel sensitivity. AR0330 also has a higher sensitivity than the old matrix, although the size of the same - 2.2μm x 2.2μm. But this is achieved by new technologies - improved microlenses in front of the pixels of the new system to remove the charge and other technologies.
5. Dynamic range and S / N ratio in all three matrices is roughly the same.
Our research and testing has shown that the new matrix OmniVision OV2710 and APTINA AR0330 are about equally good quality recording in low light conditions. OV2710 - due to the large pixel in size. AR0330 - by using the latest technologies. But the two matrices have a distinct advantage over the old 5-megapixel Aptina MT9P01 , which is still used by most manufacturers of automotive DVRs.
It should be noted that the video quality also depends on the lens that is mounted in front of the matrix. Cheap lens, which lets a little light, can easily bring all the benefits of a good matrix to the minimum. Conversely, a good fast lens can significantly improve the quality of video images at night, even with the old matrix.
Last edited by a moderator:
