Monitoring Image and Signal Degradation

Mike Hodson

Author: Mike Hodson

Published 1st August 2013

by Peter Lampel
Issue 79 - July 2013
OmniTeks latest waveform analysis software focused on assessing the degradation in video and audio quality and timing suffered by images as they are transmitted or stored. What approach has OmniTek taken to measuring these effects? In an ideal world, all broadcasters would love to have a piece of equipment that simply looks at a video signal and tells the operator what the quality of the signal is. Unfortunately this is an almost impossible thing to do, for many reasons. What is the meaning of quality, and how do you measure it? What happens if the video material has deliberately been distorted or de-focussed by the producer, for effect? Will this confuse the measurement system?
In practice, the only reliable way to assess the quality of a signal that has been through a lossy transmission path is to compare the received signal with the original transmitted signal, and look at the differences. This is called a full-reference quality measurement, and it is what the OmniTek PQA system does. With access to both the source and received signals, it is then possible to make a quantitative, repeatable assessment of the amount of degradation that has taken place.
Different full-reference comparisons can be made, varying from comparisons of two stored copies (used to assess the quality of a copy stored in a particular way), through comparisons of issued and received transmissions (used to determine the effect of the transmission path on the quality of the signal), to the comparison of the quality of the signal at different points along a transmission path (used to assess how the signal is degrading).

What metrics are used?

A range of different metrics are used for making Picture Quality assessments. The simplest and most ubiquitous is the Peak Signal to Noise Ratio (PSNR), which is a calculation of the actual differences in the pixel values between transmitted and received signals, summed over the whole screen area and expressed as a ratio in dB. This gives a quick and easy comparison of the two images, but is not very good at determining subtle differences in images that the human eye is more receptive to.
The OmniTek PQA also offers an improved algorithm called the Compensated Signal to Noise Ratio (CSNR), where the pixel difference values are modulated by an edge map of the source image (to highlight compression artefacts) and also a luminance characteristic curve (to mask errors where the image is very dark or very bright). This technique gives results which more accurately reflect a viewers perception of the image quality.
Of course, Picture Quality isnt the only factor to be considered in relation to transmission and storage of video. Theres also Audio Quality which, in accordance with ITU-R BS.1387-1 and the PEAQ audio model, is assessed in terms of an Objective Difference Grade, a Distortion Index, and loss of Amplitude. Also important are the delays that become applied to the audio and video streams, which when different give rise to Lip-Sync problems. How does OmniTek measure Lip-Sync delay?
OmniTek offers two different methods for measuring lip-sync delay. The first technique involves the use of a special video sequence, which consists of a test pattern plus an audio channel blip. This sequence is played out from a VTR or file server, through a transmission path: when received by any OmniTek measurement system, this displays the lip-sync delay. Calculation of lip-sync using this technique is accurate to one audio sample (typically 21 microseconds).
The latest version of OmniTek software now also includes an optional in-service lip-sync delay measurement system, which does not require any special video sequence and is at least as accurate if not slightly more accurate than the test sequence technique. As long as the OmniTek equipment is provided with signals from both the transmission and the reception paths, the total loop delay of both video and audio components (and hence the relative A/V lip-sync delay) can be calculated and displayed.
Where did OmniTek gain its expertise in Waveform Analysis?
OmniTek has been creating high-resolution waveform monitoring systems for over 10 years. Our in-house design team have a strong background in audio and video signal analysis and digital image manipulation, and generating waveforms is a special application of these signal processing techniques.
How has compliance monitoring changed over the years?
The test and measurement business has actually changed quite radically over the past 25 years. The days of engineers in white coats measuring diff phase & gain on composite video signals have largely gone: nowadays, in an all-digital transmission system, the primary requirement for signal delivery is one of monitoring rather than measurement of signals.
For this reason, a lot of the work we do at OmniTek is integration of our waveform and monitoring software into third-party equipment such as flat-panel displays, editing systems, switchers etc. For many applications, all that is required is a green light / red light indication that the signal on an SDI cable is within the accepted SMPTE (or other) standards. OmniTek provides both the monitoring systems and the key processing software that enables third parties to offer monitoring functions in a wide variety of different systems.
Changes in technology and government measures such as the CALM Act in the US to limit differences in loudness between programmes and adverts have also led to changes in the range of analyses we offer. The use of digital rather than analogue transmission has also introduced the need for eye diagrams and jitter displays as part of the monitoring
package. Has monitoring Loudness presented any special challenges?
OmniTek provides a comprehensive set of measures of audio amplitude on all its products from PPMs mimicking industry-standard metering systems, through True Peak displays, to a range of displays showing Momentary Loudness (calculated over 400ms), Short-Term Loudness (calculated over 3s) and Integrated Loudness (calculated over the whole of a programme). The primary challenge has been keeping pace with all the industry standards, which have been a bit of a moving target over the past 5 years! Happily OmniTeks software-based system architecture has meant that we have been able to keep pace with all the changes, and be first-to-market when a new specification emerges.
What problems are helped by precise measurement of eye patterns and digital jitter?
The key factor here is Quality of Service. How can a broadcaster guarantee that his signals are being distributed throughout his installation without introducing errors? The answer is to monitor the signal integrity at as many places as possible, and make sure that the SDI signals are well within specifications for amplitude, rise/fall time, and jitter. Eye diagrams can be generated on an oscilloscope but at a cost, particularly for 3G video where a suitable oscilloscope could easily cost £10,000. By incorporating eye pattern and jitter facilities in a waveform monitor such as OmniTeks OTM 1000, the broadcaster gets accurate, easy-to-read displays and instant analysis which can moreover be recorded for subsequent analysis.

Related Articles

Related News

Related Videos

© KitPlus (tv-bay limited). All trademarks recognised. Reproduction of this content is strictly prohibited without written consent.