Measuring to Find the Best Video Processing and Compression Algorithms
by billreckwerdt on 2007-09-23
Video processing and compression algorithms change the characteristics of the original program in the quest of reducing the bandwidth needed to send the programming information to the home. The art is to do this without allowing the audience to perceive a change in video quality. Successful video processing and compression algorithms perform the desired modifications while presenting a result to the viewer that, subjectively, looks natural and realistic. This sounds difficult, but it is necessary to send many channels of high-quality programming.
Each broadcaster - traditional or web caster - must deal with rapidly changing varieties of programming, new video processing algorithms, and new compression algorithms. Many video processing and compression companies continuously invent sophisticated ways to reduce the huge bandwidth requirements to manageable levels.
Previous Options
Traditionally, broadcasters invite the various video processing and compression companies into their R&D facilities, and perform side-by-side tests also known as a "bake-off". Each vendor starts with the same source material, and does their best job to reduce the bandwidth while keeping the video quality high.
The test setup usually consists of source video programming stored on a VTR or a light-compressed video server. The source video is sent to the various video processing and compression systems. The original source and processed video sequences are saved for later viewing.
The broadcaster shows the results to a group of experts and asks them, which one is the best. This is termed subjective video analysis, and it measures the overall perceived video quality. The most commonly used video quality evaluation method is the Mean Opinion Score (MOS), recommended by the ITU. It consists in having several experts viewing a known distorted video sequences in order to rate its quality, according to a predefined quality scale. By doing this the expert viewers are trained to build a mapping between the quality scale and a set of processed video sequences. After the "training" is complete, the subjects are then asked to rate the new video processing algorithms. Although MOS studies have served as the basis for analyzing many aspects of signal processing, they present several limitations: a) very stringent environments are required; b) the process cannot be automated; c) it is very costly and time consuming to repeat it frequently.
Video Clarity Solution
The ClearView Video Analysis system simplifies the work flow, by combining the video server and capture device into one unit. By doing this, the original source and processed video sequences can be displayed - side-by-side, mirrored, or seamless split - on the same display. Further the operator can play any video sequence in any order, at any speed, for any duration with zoom, pan, jog, and shuttle capabilities.
ClearView applies various objective metrics to the video sequences, generates graphs, and calculates an objective score. While development of several algorithms is ongoing, ClearView is a hybrid system that takes into consideration subjective testing with objective measurements. ClearView can easily be programmed to display video sequences for the expert viewers; while recording the objective metric scores along with the MOS. While the MOS cannot be repeated, the objective metric can be easily and readily.
Benefits
* Repeatable tests, quantitative results, and a streamlined setup.
* Analyze 2, 1080P video sequences in real-time
* ClearView does not alter the original video sequences - video sequences are processed completely uncompressed.
* Multiple viewing modes are presented on 1 display - no need to calibrate 2 separate Television displays to compare video sequences.
* Analytics are displayed for each video sequence.
Implementation
ClearView takes advantage of the high-reliability of today's off-the-shelf computer platforms. This ensures that products are made with the latest and greatest hardware available, while at the same time avoiding the high cost of custom designs. ClearView is packages in a 4U, 20" deep Industrial Chassis, and is supported by VideoClarity.
Summary
The ClearView Video Analysis System provides broadcasters, video researchers, compression developers with the unique ability to capture, play-out, and analyze video sequences. Objective measurements are generated and logged for repeatable tests in the future.
Each broadcaster - traditional or web caster - must deal with rapidly changing varieties of programming, new video processing algorithms, and new compression algorithms. Many video processing and compression companies continuously invent sophisticated ways to reduce the huge bandwidth requirements to manageable levels.
Previous Options
Traditionally, broadcasters invite the various video processing and compression companies into their R&D facilities, and perform side-by-side tests also known as a "bake-off". Each vendor starts with the same source material, and does their best job to reduce the bandwidth while keeping the video quality high.
The test setup usually consists of source video programming stored on a VTR or a light-compressed video server. The source video is sent to the various video processing and compression systems. The original source and processed video sequences are saved for later viewing.
The broadcaster shows the results to a group of experts and asks them, which one is the best. This is termed subjective video analysis, and it measures the overall perceived video quality. The most commonly used video quality evaluation method is the Mean Opinion Score (MOS), recommended by the ITU. It consists in having several experts viewing a known distorted video sequences in order to rate its quality, according to a predefined quality scale. By doing this the expert viewers are trained to build a mapping between the quality scale and a set of processed video sequences. After the "training" is complete, the subjects are then asked to rate the new video processing algorithms. Although MOS studies have served as the basis for analyzing many aspects of signal processing, they present several limitations: a) very stringent environments are required; b) the process cannot be automated; c) it is very costly and time consuming to repeat it frequently.
Video Clarity Solution
The ClearView Video Analysis system simplifies the work flow, by combining the video server and capture device into one unit. By doing this, the original source and processed video sequences can be displayed - side-by-side, mirrored, or seamless split - on the same display. Further the operator can play any video sequence in any order, at any speed, for any duration with zoom, pan, jog, and shuttle capabilities.
ClearView applies various objective metrics to the video sequences, generates graphs, and calculates an objective score. While development of several algorithms is ongoing, ClearView is a hybrid system that takes into consideration subjective testing with objective measurements. ClearView can easily be programmed to display video sequences for the expert viewers; while recording the objective metric scores along with the MOS. While the MOS cannot be repeated, the objective metric can be easily and readily.
Benefits
* Repeatable tests, quantitative results, and a streamlined setup.
* Analyze 2, 1080P video sequences in real-time
* ClearView does not alter the original video sequences - video sequences are processed completely uncompressed.
* Multiple viewing modes are presented on 1 display - no need to calibrate 2 separate Television displays to compare video sequences.
* Analytics are displayed for each video sequence.
Implementation
ClearView takes advantage of the high-reliability of today's off-the-shelf computer platforms. This ensures that products are made with the latest and greatest hardware available, while at the same time avoiding the high cost of custom designs. ClearView is packages in a 4U, 20" deep Industrial Chassis, and is supported by VideoClarity.
Summary
The ClearView Video Analysis System provides broadcasters, video researchers, compression developers with the unique ability to capture, play-out, and analyze video sequences. Objective measurements are generated and logged for repeatable tests in the future.
About The Author: