









There is several type of IP CAM in the market. In this topic, I will be quantification for these factors. For example, PTZ, MPEG-4/MJPEG, CCD/ CMOS and etc.
What kind of factor is more important? What is its relative value
I will depend on Axis IP CAM price to analysis the factor value using
Numerical analysis:
1. I get the Axis IP cam price from Axis’
2. I think these price lists will inform me two information
A. sort/priority of factor
B. The relative weight of factor. But not absolute weight
3. Finally, we can understand which factor is important and realize every factors’ relative weight.
First, I divided Axis IP CAM into 3 types.
Those are Low cost CMOS IP CAM, middle end CCD IP cam and PTZ/speed dome series.


The add value form high to low is as below.
1. PTZ platform :Fixed <>
2. Auto focus: fixed <>
3. resolution : VGA <>
4. D/N switch normal <>
5. Sensor : CMOS <>
6. Compression : MPEG4 <> We only can realize the factor rank (relative value) but NOT its quantification (absolute value ).
What is optical filter?
If you know Chinese character, you can reference the website as below.
http://bbs.cpsmedia.cn/frame.php?frameon=yes&referer=http://bbs.cpsmedia.cn/
Depending on that article, he inform me that Optical filter will improve two points
1. Infrared
2. Adjust the high frequency video image. (My English is not very good. So I do not use the “correct “ word to describe )
Some one asks me if we can use digital video Algorithm to replace optical filter?
I haven’t seen the Algorithm so far, especially adjust the Infrared
1. Infrared
In security or surveillance field, we want to cut infrared in daytime but we must need infrared in the night.
In daytime, we want to cut infrared because of beautiful picture.
We can see the comparison table as below.
The left picture is without IR cut filter. (that means the infrared will input the sensor )
The right picture is with IR Cut filter.
But in the night, we need capture the video by using extra infrared light source.
That is why we often hear Day & Night camera in the market.
The D&N camera has a switch IR cut filter machine controlled by motor. In daytime, the IR cut filter will auto be “with” IR cutl filter.
In the evening, the IR cut filter will auto be “without” IR cut filter.
If you have ability to let left picture to become right picture by using video Algorithm.
Be careful we will become rich guy.
2.Adjust the high frequency video image.
When I first to touch optical low pass filter (OLPF), I also think this thing is like stock’s moving average line (MA).
Compare with 1day MA and 10day MA, 10 day MA will be smoother than 1 MA.
That is trade off for detail information and trend (average).
The OLPF is like this. It will cut off Moiré (false color). But sometime it will loss a little resolution.

But the Day & Night camera will separate two parts, IR cut filter and OLPF.
Regarding IR cut filter, it will be controlled by CPU through Light Dependent Resistor signal.
Why using 3D De-interlace?
Now, most of signal that we use are traditional CVBS signal for NTSC or PAL, these type are for interlace scan. When it is through the compression chip to compress, the jagged will be occur without de-interlace process.
What kind of product need 3d-deinterlace?
They are IP camera, PC base DVR and video server.
That divided two parts to analysis. 1CH input and 4CH input
1. 1CH.
This application is for ccd interlace ccd ip camera.
There is no ic design house to design 3d-deinterlace solution for ip camera before.
All that you see what IP camera maker claim build in 3D de-interlace H/W base is DVD player chip set solution. Because of no solution, IP camera maker only use these solution.
Many IC design house start to notice this function now.
First generation: (DVD solution)
That is three chip solutions.
Second generation: (start to think 3d de-interlace for security market use)
The solution still needs two chips. But this 3d-deinterlace digital ic has embedded DRAM.

Third generation: (perfect type)
One chip solution

One IC design in
What kind of its quality?
It's been good so far.
Not only 3D-deinterlace, but also 3D Comb filter and 3D noise reduction.
Most of video decoder that using in Security market are 2D comb filter.
If our camera focus on a tight squeeze, the video will output false color (Moiré)
Please see comparison picture as below.

By the way, 3D noise reduction is different to 2D noise reduction. 3D noise reduction will depend on 3 frame pictures to reduce noise. So it do not loss the image resolution.
2. 4CH
Video server or PC Base DVR will use 4ch video decoder and need 3D-deinterlace process.
Some one think the deinterlace faction can be proceed on the video compression chip.
I do not agree this view point at all.
It is ok in the 1ch system but not in 4ch.
We can get the 8ch PC Base DVR card with video compression in the CONEXANT website.

They need extra FPGA to process signal interface between video and compression chip.
.
If 3D-deinterlace put on video decoder side and output progressive signal, it is very easy for video compression to process.
3D comb filter must process on video decoder side.
I think if video decoder can process 3d-deinterlace, 3d noise reduction and 3D comb filter, it will share the loading. The video compression will get good video to compress.
How about PC base DVR w/o compression card?
If computer want to process 3d de-interlace by software, it will use CPU computing power at least 50%
People that using PC base DVR w/o compression card want to pc to do more and more intelligence control. if we waste the CPU computing power so much, it will be less the computing power to do intelligence control
One company in








Athough SONY has strong promote its progressive scan ccd for IP camera market, the interlace is still at high market share now.
From the IP surveillance manufacturer view, they still get ccd camera module from traditional CCTV analog camera maker.
It is very easy to become CCD ip cam when connect video decoder.
Video decoder‘s input is CVBS signal and its output is ITU656. that is interlace.
ITU656 will become video compression’s input.
Because its is still interlace signa (NTSC 60field, PAL 50 field), the video output will occur jagged
In TV chip set field, the 3D-interlace is a basic Algorithm. Instead, the in the surveillance market, I do not see this ASIC now.
Although, some video codec soc claim they have de-interlace, the it is only 2D but not 3D.
I think 3D-deinterlace can be embedded in video decoder side.
The comparison picture is as below.
Before 3D-deinterlace

📱 我開發的「手機全螢幕投影器最終版本」:解決螢幕比例困境與多功能整合技術 我對手機投影到電視這件事情的熱衷,其實有個轉變的過程 。在 2022 年底,我根本不看好這塊市場 ,因為我覺得這只是個消費性的東西,沒什麼好做的 。畢竟,Apple 原廠的 Lightning ...