|
|
| Digital security video can cause proprietary headaches |
| by Wayne Cole |
|
|
|
|
Social web
| To ShareThis, click on a service below: |
Post
| To ShareThis to your profile or blog, click a service below: |
|
|
Unlike VHS systems, DVR-based security solutions may store imagery as still sequences, multiplexed audio and video streams, separate audio and video streams, or video streams with multiple camera angles interleaved into a single file. Popular compression schemes include JPEG, MPEG-1, MPEG-2, and (increasingly) MPEG-4. Even content in these standard formats is often organized into proprietary storage schemes that require a proprietary player to view it in a logical manner.
When investigators request the video, they generally receive a CD-R or DVD-R with the video of interest in its proprietary storage format with the proprietary player for playback on a computer. Players can have some inscrutable controls, weak to nonexistent help features, and minimal toolsets for extracting stills or short sequences from individual cameras. Manufacturers of these security systems are loath to provide documentation on how to operate the player/extractor tools to non-owners, leaving investigators with few choices for viewing, analyzing, and presenting security video evidence.
Extracting Clips
For example, one multi-camera DVR system from DLA provides an off-line player called Backup Viewer with video extracted from the system. The media files are proprietary files with inscrutable numeric names and a .ddf extension.
Backup Viewer allows investigators to view single or multiple camera views synchronously along with visible camera number, mode, and date/time stamps. Extracted stills show this information on the picture. But if you use Backup Viewers AVI extractor (the only option for extracting video streams one camera at a time), no camera, mode, or date/time stamp appears in the clip. This lack of data can be all it takes to get the extracted clips tossed out at trial.
To work around this problem, investigators can use a screen capture tool to grab the video as it plays in the viewer. Ocean Systems DVR dCoder provides the ability to capture an entire screen, single window, or user-defined region to an AVI or QuickTime file. The Windows Media Encoder free download from Microsoft provides the same ability. However, with Windows Media Encoder you will need to create a custom set of capture and encoding parameters to capture high-quality, full frame-rate video in your Windows Media file output.
***image1***
Some Sanyo DVRs are equipped with an internal application that outputs each camera view as a series of JPEGs arranged in hierarchical directories. For each hour of recording from each camera, the top directory was named by camera, start date, record start time, and end time of the interval. Within that directory were a series of sequentially numbered sub-directories that contained JPEGs with sequential numeric filenames starting with 00000001.JPEG.
Video from these units comes with proprietary players as an HTML page with embedded JavaScript in each top-level directory. The HTML page shows a viewing area, transport controls, play speed, and picture size controls. Below the picture output area, data fields show the frame number, date/time stamp, and originating camera for each displayed frame. Playback speed can be set in five steps from 1 fps to about 15 fps.
The players JavaScript contained a table that provided the correspondence between the date/time stamp and JPEG. The table was generated from metadata striped into the top 18 visible lines of the half-D1 (720x248) pictures.
Written in a manner similar to vertical interval time code, these barcode-like stripes dont appear in the scaled-up images produced by the HTML player. As with Backup Viewer, if you want to capture a standard video clip with date/time stamps, you will need a screen capture tool with high-quality output.
There is another way to produce a video clip with this storage scheme, albeit without readable date/time stamps. Using Adobe Photoshop CS3 Extended, the JPEGs can be imported as an image sequence into a video or animation layer, resized to standard video dimensions and pixel aspect ratio, then exported to QuickTime in any codec supported by your system.
The drawback to these methods is that clips often do not playback in real time. Each camera may have different numbers of frames recorded during each second. Playback (and thus screen capture) usually occurs at some arbitrary constant frame rate. To do something like synchronize a camera view to an independently recorded audio track, youll need a system with key-framable motion effects so that each frame plays at the proper time relative to the audio.
Picture Quality
While most DVR-based security systems can record higher quality pictures than EP-mode VHS tape systems, the end product can be as bad or worse. Because there is no law requiring security video to be archived at the same resolution with which it was recorded, many unsophisticated operators will choose the highest compression and smallest frame size available to save on storage media.
In distributed operations, remote locations may submit all their security video for archive to a central server over the Internet or a VPN. Bandwidth may dictate another lossy high compression pass before submission. The result can be video that is useless for identification purposes or even event timeline analysis.
Increasingly, investigators who request security video from a user take delivery via e-mail. In these instances, the archival video is highly compressed yet again to fit down the Internet pipe, instead of being copied directly to DVD or other media for delivery by courier or mail. The investigators often dont realize something isnt right until they show up in court and get challenged with much higher quality video from the same cameras than that which they received electronically.
The simplicity of security video on VHS is rapidly being overwhelmed by DVR systems with a multitude of recording and storage formats, as well as offline viewing/access methods. Investigative teams without digital video expertise increasingly need to get expert help early in the process to ensure that they obtain and preserve the best available value from relevant digital security video sources.
MORE INFO
Adobe adobe.com
DLA dladvrs.com
Microsoft microsoft.com
Ocean Systems oceansystems.com
Sanyo us.sanyo.com/industrial/security
The right questions
With all the proprietary systems, storage schemes, and metadata setups, investigators can no longer simply ask for a copy of the video -- especially from an adversary in the case. In fact, an entire line of questions need answers so investigators can harvest the highest quality security video:
- What is the make and model, and who is the supplier of the DVR security system?
- Is the video of interest still on the DVR or has it been archived?
- What process is used to move video from the DVR to archive?
- What storage format is used for archive and/or offline viewing?
- Does it require a proprietary player application, and if so, what platforms does the player support?
- For multiple camera systems, is the video recorded to a single stream/file with all the cameras multiplexed (interleaved) within the stream?
- If a proprietary player is required, can the system user supply it with the video? Does the player support de-multiplexed camera views, export of stills and clips to standard video formats with/without date/time stamp and camera/channel number displayed?
- Will the provided video be accompanied by a written statement or affidavit that certifies the video to be the highest quality version available, if not a direct bit-for-bit duplication of the original recording?
|
|
Social web
| To ShareThis, click on a service below: |
Post
| To ShareThis to your profile or blog, click a service below: |
|
|
|
|
|
|
|
|