From cameras to the court: How to make full video integration a reality

David Spreadborough

Certified Forensic Video Analyst & International Trainer, Amped Software

September 11, 2017

Get the IFSEC Global newsletter

The latest security and fire news, reports and resources

Download

VESDA protects thousands of tourists at the largest timber-framed church in Europe

Video has always been a rather touchy subject in the criminal justice system, and it’s often hard for many people to appreciate the difficulties.

Looking back, even in the VHS tape era, there were many court clerks scratching their heads at why the tape was playing too fast, or why it was flicking from one camera view to another.

I remember many years ago, one such clerk taking about 20 attempts at hitting pause at the right time to show the shoplifting suspect ‘in the act’. Although we had a standard recording format, no one had processed the time-lapsed, multiplexed video for court.

Then came video DVD, and the legal system thought their prayers were answered. The challenge though came to the police and defence experts. In many places, it was their responsibility to prepare this old analogue, tape-based footage and get it onto a digital disk.

For playback, it was easy: you had a standard player that could be purchased for peanuts and a standard format that allowed some degree of control.

For creation though, it was a legal nightmare and one that created all sorts of problems. Getting the footage onto that disk, whilst retaining integrity and evidential worth required hardware, software and people with competency. It was very easy to create a bad video DVD with poor quality and little integrity.

Manufacturers needed an ‘open’ system. Great for them – but even better for a beleaguered video analyst like me

And just to add in another complication, at that time we were right in the middle of the big VHS/DVR switchover. First and second-generation DVRs were purely VHS tape replacements.

They were terrible, with very little thought about how to get the digital data out in an evidential manner. Many different solutions were implemented, but again the requirements on those having to deal with the surveillance footage increased.

Many companies that had considered the export process often did not fully understand worldwide legal frameworks and would fail to give any ability to interpret or work effectively with the footage that had been recorded.

So, we find ourselves circa 2010, where a video analyst in the UK needed around three desks worth of equipment to cope with the average weekly workload. Signal capturing, downscaling, screen grabbing, file conversion, reformatting: all just to get an image or a video into the criminal justice system.

Due to some of the technical requirements, it was even easier to get things wrong. Getting it wrong in the legal world could have some dire consequences.

It was a challenge dealing with the input format, and now a challenge with the output format, mainly due to video DVD sometimes being completely unsuitable for the images/video to be presented but the courts only accepting that format.

Early days of IP

Very slowly, one or two system manufacturers started to talk together. We were in the early days of IP and they needed to have a video format that could start in a camera and end up being exported.

They needed an ‘open’ system. Great for them – but even better for a beleaguered video analyst like me.

I now had video data that I could independently validate and interpret. Not only on the HDD but the export as well.

There were even one or two manufacturers that kept with stream formatting specifications, allowing individual cameras to be easily managed.

There are quite possibly a great number of people working within the video surveillance industry thinking that you all do this. All footage can be played and exported and then reviewed.

I estimate that there are between 7,000 to 10,000 different types of video surveillance recording devices currently in use

Well, within your own world it can – but bring many formats into my world, and that’s where things start to get interesting.

In just the same way as manufacturers have started to integrate with others, ensuring the video signal gets from camera to DVR/NVR, we now need further integration to allow the video to carry on into the legal system.

The regulations, guidance and forensic workflows of dealing with video evidence ensure that it can be trusted and relied upon. It is vital that the original recordings are retained, interpreted correctly, and if enhancement is required, then it’s completed in a repeatable and reproducible manner.

Getting the correct perspective of objects: An example of Amped Software’s work

In the most recent UK National Surveillance Camera Strategy for England and Wales, it states that there is a challenge within law enforcement and investigations because of an estimated 900 different formats. As we know, this is nothing new, it’s just that national digitisation projects are now identifying the problem that many investigators were shouting about over 10 years ago.

This number can be a little confusing so let me give you some estimates, based on my experience of investigating CCTV systems worldwide over the past 15 years. I estimate that there are between 7,000 to 10,000 different types of video surveillance recording devices currently in use.

Within those systems, there are less than five different recording codecs, with H264 being the current top dog. Finally, we have more than 900 different formats.

The format relates to how the data is structured and the use of a specific file extension. It also includes the method of storing and decoding date and time information.

DVR A and DVR B

To put all of this into some form of perspective, let me introduce you to DVR A, and DVR B. DVR A is great. DVR B is, well let’s just say that it’s not so great.

DVR A records in a standard H264 format. Its date and time format is stored within an open standard that can be decoded easily. It stores this time against every frame and includes the milliseconds.

Each camera is a separate stream, and the manufacturer has retained the standard for stream mapping and identification. The HDD can easily be removed and slotted into a caddy.

The manufacturer retains all instructions of how to review, analyse and export, within documents easily found on its website. It also has all players and clients available to download. If clips are exported from the DVR, these clips are standard video files with standard subtitle tracks.

Measuring height: Example of Amped Software’s work

The video files can be analysed at all levels: format, codec, GOP (group of pictures – frames related between them), frame, macroblock (block of related pixels in a frame), pixel. The video files are given a unique checksum using an open algorithm upon export. Any video will play in any standard player.

Analysts can drop these videos and subtitles into forensic video software, such as Amped FIVE (Forensic Image and Video Enhancement), and immediately begin their investigative process knowing that this was the original recording and not a transcoded or recaptured copy.

It’s fantastic to see some manufacturers writing SDKs and APIs that allow other software to decode their formats when they use a proprietary recording method

File authentication is simple with these exports as they each have a unique, standard mathematical value that is created at time of export. The method of authentication is transaperent, verifiable and most importantly, it is not reliant on a proprietary technology.

With DVR B, it’s a different story. For starters, the USB export ports are on the back of the DVR so they cannot be accessed easily.

If all the footage was required, the internal HDD is secured inside the case so it’s not a simple extraction. The manufacturer only retains make and model support for current hardware and places those behind an installer login page.

When removed, the data on the hard drive is non-standard, with the video data and time information stored separately using a proprietary method. Although recording in H264, the header information is not held on the drive.

Exports can be created by the DVR but they transcode each new video to place the date and time information on the footage. The transcode process rewrites the files.

Native files cannot be viewed or interpreted or validated by other software. The proprietary player requires installing on a computer and installs proprietary codecs within the system. Although the software has the ability to check file integrity, it relies upon a closed technology.

Most systems fall somewhere in between these two extremes but I hope you get the picture – which I often don’t when dealing with exports from DVR B!

Bridging the gap

I train police officers and staff worldwide in forensic video analysis and in image authentication. Software such as Amped FIVE forms part of the final stage in a surveillance video’s journey.

It’s not the DVR or NVR, and it’s not the export player. Legal systems around the world now need your help to bridge the gap and ensure that the challenge of these 900 plus formats can easily be overcome.

If you are a CCTV owner, can you quickly export 10 mins, 10 hours or your entire system? For those with large installations, do you have a major incident recovery plan in place? For installers, are you installing equipment where footage can be exported, analysed, interpreted and managed if an incident occurs?

Can it be played and reviewed without the need to install software? For manufacturers, are you working alongside the forensic video and investigative community to ensure that the footage your devices captures can be easily managed and integrated within any legal system?

On a final note, it is fantastic to see some manufacturers writing specific SDKs and APIs that allow other software to understand and decode their formats when they use a proprietary recording method.

I always enjoy the look on students’ faces when I show them that I can simply drop in a piece of footage and it plays without a problem. After spending years sourcing players and codecs, and then being forced to screen capture the footage, direct decoding is like a breath of fresh air.

There are no losers in integration. By working together and ensuring a full native workflow, the camera to court ideology becomes a reality.

Please get in touch if you want to learn more.

CFVA: info@ampedsoftware.com

 

 

Free Download: The key to mitigating cybersecurity risks

Exploiting IoT technology without creating cybersecurity vulnerabilities is one of the defining challenges in today’s security landscape.

This report will help you to see why third parties should adhere to ‘secure by design’ principles and why the necessary convergence of IT and security departments demands a holistic approach.

Download now

Related Topics

Leave a Reply

1 Comment on "From cameras to the court: How to make full video integration a reality"

Notify of
avatar
Sort by:   newest | oldest | most voted
Doktor Jon
Guest

Excellent insights into the world of the FVA.
Many of the points raised regarding recording management and export have long been key components in the TRUSTED CCTV Operational Standards.
Just a shame that despite many years of discussion, we still don’t have any form of industry standardisation, for recognised compliant recording devices.
Hopefully in the not too distant future ….
Doktor Jon

wpDiscuz
Topics: