The Broadcast Knowledge

Video: A Forensic Approach to Video

Posted on 2nd March 2020 by Russell Trafford-Jones

Unplayable media is everyone’s nightmare, made all the worse if it could be key evidence in a crimnial case. This is daily fight that Gareth Harbord from the Metropolitan Police has as he tries to render old CCTV footage and files from crashed dash cams playable, files from damaged SD cards and hard drives readable and recover video from old tape formats which have been obselete for years.

In terms of data recovery, there are two main elments: Getting the data off the device and then fixing the data to make it playable. Getting the data off a device tends to be difficult because either the device is damaged and/or connecting to the device requires some proprietary hardware/software which simply isn’t available any more. Pioneers in a field often have to come up with their own way of interfacing which, when the market becomes bigger, is often then improved by a standard way of doing things. Take, as an example, mobile phone cables. They used to be all sorts of shapes and sizes but are now much more uniform with 3 main types. The same was initially true with hard drives, however the first hard drives were so long ago that osolecence is much more of an issue.

Once you have the data on your own system, it’s then time to start analysing it to see why it won’t play. It may play because the data itself is of an old or proprietary format, which Gareth says is very common with CCTV manufacturers. While there are some poular formats, there are many variations from different companies including putting all, say, 4 cameras onto one image or into one file, running the data for the four cameras in parallel. After a while, you start to be able to get a feel for the formats but not without many hours of previous trial and error.

Gareth starts his talk explaining that he works in the download and data receovery function which is different from the people who make the evidence ready for presentation in a trial. Their job is to find the best way to show the relevant parts both in terms of presentation but also technically making sure it is easy to play for the technically uninitiated in court and that it is robust and reliable. Presentation covers the effort behind combining multiple sources of video evidence into one timeline and ensuring the correct chronology. Other teams also deal with enhancing the video and Gareth shows examples of deblurring an image and also using frame averaging to enhance the intelligability of the picture.

Gareth spends some time discussing CCTV where he calls the result of the lack of standardisation “a myriad of madness.” He says it’s not uncommon to have 15-year-old systems which are brought in but, since the hard drives have been spinning for one and half decades, don’t start again when they are repowered. On the otherhand the newer IP cameras are more complicated whereby each camera is generating its own time-stampped video going into a networked video recorder which also has a timestamp. What happens when all of the timestamps disagree?

Mobile devices cause problems due to variable frame rates which are used to deal with dim scenes, non-conformance with standards and who can forget the fun of CMOS videos where the CMOS sensors lead to wobbling of the image when the phone is panned left or right. Gareth highlights a few of the tools he and his colleagues use such as the ever-informative MediaInfo and FFProbe before discussing the formats that they transode to in order to share the videos internally.

Gareth walks us through an example file looking at the how data can be lined up to start understanding the structure and start to decode it. This can lead to the need to write some simple code in C#, or similar, to rework the data. When it’s not possible to get hold of the data in a partiular format to be playable in VLC, or similar, a proprietary player may be the only way forward. When this is the case, often a capture of the computer screen is the only way to excerpt the clip. Gareth looks at the pros and cons of this method.

Watch now!
Speakers

Gareth Harbord
Senior Digital Forensic Specialist (Video)
Metropolitan Police Service

Video: 5 PTP Implementation Challenges & Best Practices

Posted on 28th February 2020 by Russell Trafford-Jones

PTP is an underlying technology enabling the whole SMPTE 2110 uncompressed ecosystem to work. Using PTP, the Precision Time Protocol, the time a frame of video, audio etc. was captured is recorded and so when decoded can be synchronised with other media recorded around that same time. Though parts of 2110 can function without it, when it comes to bringing media together which need synchronisation, vision mixing for instance, PTP is the way to go.

PTP is actually a standard for time distribution which, like its forerunner NTP, was developed by the IEEE and is a cross-industry standard. Now on version IEEE-1588-2019, it defines not only how to send time onto a network, but also how a receiver can work out what the time actually is. Afterall, if you had a letter in the post telling you the time, you’d know that time – and date for that matter – was old. PTP defines a way of working out how long the letter took to arrive so that you can know the date and time based on the letter and you new-found knowledge of the delivery time.

Knowing the time of day is all very well, but to truly synchronise media, SMPTE ST 2059 is used to interpret PTP for professional media. Video and audio are made from repeating data structures. 2059 relates these repeating data structures back to a common time in the past so that at any time in the future, you can calculate the phase of the signal.

Karl Khun from Tektronix starts by laying out the problems to be solved, such as managing jitter and the precision needed. This leads in into a look at how timestamps are used to make a note of when, separately, video and audio were captured. The network needed to implement PTP, particularly for redundancy and the ability of GPS allowing buildings to be co-timed without being connected.

Troubleshooting PTP will be tricky for many, but learning the IT side of this is only part of the solution. Karl looks at some best practices and tips on faultfinding PPT errors which leads on to a discussion of PTP domains and profiles. An important aspect of PTP is that it is bi-directional. Not only that but it’s much more than a distribution of a signal like the previous black and burst infrastructure. It is a system which needs to be managed and deserves to be monitored. Karl shows how graphs can help show the stability of the network and how RTP/CC errors can show network packet losses/corruptions.

Watch now!
Speakers

Karl J. Khun
Principal Solutions Architect
Telestream/Tekronix

Video: LCEVC – The Latest MPEG Standard

Posted on 27th February 2020 by Russell Trafford-Jones

Video is so pervasive in our world that we need to move past thinking of codecs and compression being about reducing bitrate. That will always be a major consideration, but speed of compression and the computation needed can also be deal breakers. Millions of embedded devices need to encode video which don’t have the grunt available to the live AV1-encoding clusters in the cloud. Further more, the structure of the final data itself can be important for later processing and decoding. So we can see how use-cases can arise out needs of various industries, far beyond broadcast, which mean that codecs need to do more than make files small.

This year LCEVC from MPEG will be standardised. Called Low Complexity Enhancement Video Coding, this codec provides compression both where computing is constrained and where it is plentiful. Guido Meardi, CEO of V-Nova, talks us through what LCEVC is starting with a chart showing how computation has increased vastly as compression has improved. It’s this trend that this codec intends to put an end to by adding, Guido explains, an enhancement layer over some lower-resolution video. By encoding a lower-resolution, computational processing is minimised. When displayed, an enhancement layer allows this low resolution video to be sharpened again to bring it back to the original.

After demonstrating the business benefits, we see the block diagram of the encoder and decoder which helps visualise how this enhancement might be calculated and work. Guido then shows us what the enhancement layer looks like – a fairy flat image with lots of thin edges on it but, importantly, it also captures a lot of almost random detail which can’t be guessed by upsamplers. This, of course, is the point. If it were possible to upscale the low-resolution video and guess/infer all the data, then we would always do that. Rather, downscaling and upscaling is a lossy process. Here, that loss is worth it because of the computational gains and because the enhancement layer will put back much of what was once lost.

In order to demonstrate LCEVC’s ability, Guido shows graphs comparing LCEVC at UHD for x264 showing improvements of between 20 and 45% and image examples of artefacts which are avoided using LCEVC. We then see that when applied to AVC, HEVC and VVC it speeds up encodes at least two fold. Guido finishes this presentation showing how you can test out the encoder and decoder yourself.

The last segment of this video, Tarek Amara from Twitch sits down to talk with Guido about the codec and the background behind it. Their talk covers V-Nova’s approach to open source, licensing, LCEVC’s gradual improvements as it went through the proving process as part of MPEG standardisation plus questions from the floor.

Watch now!
Speakers

	Guido Meardi CEO & Co-Founder, V-Nova
	Tarek Amara Principal Video Specialist, Twitch

Video: Extension to 4K resolution of a Parametric Model for Perceptual Video Quality

Posted on 26th February 2020 by Russell Trafford-Jones

Measuring video quality automatically is invaluable and, for many uses, essential. But as video evolves with higher frame rates, HDR, a wider colour gamut (WCG) and higher resolutions, we need to make sure the automatic evaluations evolve too. Called ‘Objective Metrics’, these computer-based assessments go by the name of PSNR, DMOS, VMAF and others. One use for these metrics is to automatically analyse an encoded video to determine if it looks good enough and should be re-encoded. This allows for the bitrate to be optimised for quality. Rafael Sotelo, from the Universidad de Montevideo, explains how his university helped work on an update to Predicted MOS to do just this.

MOS is the Mean Opinion Score and is a result derived from a group of people watching some content in a controlled environment. They vote to say how they feel about the content and the data, when combined gives an indication of the quality of the video. The trick is to enable a computer to predict what people will say. Rafael explains how this is done looking at some of the maths behind the predicted score.

In order to test any ‘upgrades’ to the objective metric, you need to test it against people’s actual score. So Rafael explains how he set up his viewing environments in both Uruguay and Italy to be compliant with BT.500. BT.500 is a standard which explains how a room should be in order to have viewing conditions which maximise the ability of the viewers to appreciate the pros and cons of the content. For instance, it explains how dim the room should be, how reflective the screens and how they should be calibrated. The guidelines don’t apply to HDR, 4K etc. so the team devised an extension to the standard in order to carryout the testing. This is called ‘subjective testing’.

With all of this work done, Rafael shows us the benefits of using this extended metric and the results achieved.

Watch now!
Speakers