Featured Post

Welcome to the Forensic Multimedia Analysis blog (formerly the Forensic Photoshop blog). With the latest developments in the analysis of m...

Thursday, February 28, 2019

The importance of audio

It's been about 4 years now since the body-worn camera really exploded on the scene. Citizens demanded accountability and transparency from their police and the market responded with cameras of all types.  Cameras have been deployed on the officer, on the vehicle, and in the air (drones / helicopters). Citizens like the real-time view that theses cameras bring in reviewing officers' interactions with the public. BUT ... these cameras often come with a very sensitive microphone that can clearly capture not only the words of the officers but also every single sound within about 25' radius. This can be a problem in light of the various privacy laws that exist in the US.

This mind-blowing amount of data is sitting in servers at police stations and in the cloud via platforms like Evidence.com. The majority of the cloud vendors offer a solution to accommodate records act releases in light of the many privacy laws that require compliance. But, as users are finding out, pennies saved on features not purchased are costing them dollars when dealing with redaction and other requests for the video they store.

When an agency goes through an equipment purchasing process, they often expend all of their resources and are locked into the solution for years - sometimes a decade or longer. Thus, the test / validate process becomes vital. It's equally vital that all stakeholders participate.

Many agencies purchased camera and storage solutions, but skipped on a redaction plan to come under spending targets. Or, they had a legal structure that let them deny requests for data at the time - but that legal assurance no longer exists (California's new laws are an example of this). They're now faced with new challenges as laws have been enacted / changed to assure that disclosure of videos is the norm, not the exception. Given the guidance in SWGDE's Video and Audio Redaction Guidelines, that redaction must include the audio, video, and metadata, agencies are scrambling to cobble together solutions to fulfill records requests in a timely and cost effective manor.

Video redaction is rather straight forward. Many tools have offered the basic blur, mosaic, and/or solid shape redactions for video content for quite some time. The COTS tools from Adobe, Apple, Avid, Magic all offer some form of global or selective video redaction. There are purpose built redaction tools from Axon, Ikena, etc. that offer redaction as well. Tools like Amped's FIVE have the ability to redact video. BUT ... it's the audio / metadata redaction where many tools fall short.

Amped SRL's CEO promised audio redaction in FIVE in 2106 at the LEVA Training Symposium in Scottsdale in support of Amped's partnership with Axon (who were both present at the Symposium as vendors). To date, audio redaction still has not been implemented in FIVE. Sure, if audio is present in the file, FIVE will pass it through or strip it out. But FIVE isn't an audio editing / redacting tool. For redactions with FIVE, I suggest using FIVE to separate the audio and using a tool like Audacity to edit / redact the audio. At leas Audacity is free and easy to use. If an agency has multiple people working on redaction, it really does make sense to split up the redaction tasks anyway (audio / video / metadata), so I guess the lack of audio support in FIVE for the redaction task isn't that big of a deal.

Ikena's Spotlight suffers similarly. Though it supports audio redaction, it's lack of support for proprietary / secure video file types makes it a tough sell to agencies. In a comparison of FIVE vs Spotlight for redaction, it's a close call. FIVE gives you support for odd file types as well as a full suite of restoration / clarification filters. Spotlight gives you a dedicated redaction platform for audio and video. Both fall short on identifying / redacting specific elements of a file's metadata.

I've been quite the cheerleader for Amped's products over the years, and have noted their strengths for a while now. I've been teaching their product line for over 7 years now, creating and delivering a curriculum that's rather specific to forensic science. But, the lack of audio support has always been the elephant in the room. For example, with the older Video Mixer and new Multiview filters, audio is simply dropped. Users expect to mix video views to create compelling demonstratives - that include an audio source. But these filters do not allow the user to select an audio source for the resulting file. With this, they end up processing the file in a non-linear editor. This issue is where the lack of audio support becomes a big deal and brings us back to that "one tool to fix everything" discussion. It's quite obvious by now that Amped SRL wants to have the best video / image tool, and doesn't want to concern themselves with audio / metadata.

This brings us back to redaction, and the other parts of the multimedia container. If you're going to need an audio editor anyway, and if you're going to have to assemble the separately redacted parts as a final step, why not just do everything in a COTS tool and save the money?

For these reasons, and a few others, the decision was made to offer our redaction courses in many flavors. In standing up training and professional services as a separate organization this year, we can offer training in performing redactions with the relevant COTS platforms as well as a more generic redaction course for supervisory / quality control staff. The initial roll-out of our redaction classes will happen in March 2019 and feature the Adobe Creative Suite tools, Magix Software tools, and Amped SRL's FIVE paired with Audacity - each offered as a separate class. Also the generic course will highlight the different tools, focussing on the process without diving too deeply on the technology.

Recent court cases involving body-worn cameras have highlighted the importance of the devices' audio. Audio is a vital piece of the container. The audio track is often used to confirm / refute elements of complaints, uses of force, and other events that unfold before the device. Because it is present in the container, it must be redacted carefully and accurately. Because audio must be redacted, your redaction solution must include the ability to redact audio. Because it must be redacted, our training sessions feature a variety of options from the big vendors. Stay tuned ...

No comments: