Software and Result Verification

The Verify Process

A lot of discussion recently, on forums and sites, has been on the importance of verifying your tools (software/hardware) and the verification of your results after any video process. Some people seem to have this nailed down completely. Some people have nothing. Most though are somewhere in between.

Taking a number of people’s views into account, and merging their methods into one,  I am attempting to pull together a set of guidelines that may help others looking into this process.

The capturing and transcoding of Digital CCTV is one of the more common processes asked of people working with video. There are many different methods to achieve the task and all depend on the footage origin, type, content, desired output and importantly, the requirement for forensic integrity.

In my opinion there are three different times when the verification of a process could take place.

  1. Before using a tool or software in order to verify its proficiency.
  2. When it is necessary to use a piece of software you have no experience or knowledge in.
  3. After a process, in order to assess the quality of the result.

Before using a tool or software in order to verify its proficiency.

The purpose of the initial verification process is to test the software/hardware method used and then document the results. These results can be filed and then produced if the specific method is questioned.

It is possible to create your own video clip for the purpose of testing. One that I have used for a while consists of a 10 second uncompressed avi file. The video is progressive (non-interlaced), with time-code, moving graphics, text and colour bars.

verifi

However, I have also started to use the test video’s that are included with Virtualdub. The benefits of using these are that they are freely available to anyone and you can adjust them to fit your purpose.

vdub

To test for loss in clarity I often use the Zone Plates.

zoneplates

You can find them in Tools > Create Test Video

The four main tasks that would benefit from verify reports are:

  1. Scan Conversion
  2. Screen Capture
  3. HD Screen Recording
  4. Transcoding into another format

Scan Conversion

Using a TVone Corio Card, Model Number C2-160.

The output from the computers graphics card (RGB 1024×768) are input into the Corio card and then downscaled to standard PAL resolution and output via Y/C (S-Video). For this test I have fed the signal directly into a Video DVD Recorder. This would be similar to playing Digital CCTV footage and recording it directly to DVD without any enhancement or editing. I have used my test verification clip within Virtualdub and this has been recorded by the DVD Recorder.

After conducting the tests and creating a ‘Verify’ DVD, the resulting files can be analysed and compared against each other. The analyst should report exactly how this has been done and with what software.

The conclusion should detail any image artefacts introduced or loss in visual clarity. Finally, suggestions should be made as to examples where this method of creating a Video DVD may be suitable along with warnings of when not to use such a process.

Screen Capture

Any screen capture software, recording into any format.

After conducting the capture and creating a ‘Verify’ video file, resulting images and video can be analysed and compared against the original. The analyst should report exactly how this has been done and with what software.

The conclusion should detail any image artefacts introduced, loss in visual clarity and importantly for video, evidence of dropped frames. Finally, suggestions should be made as to examples where this method of creating a Video file may be suitable along with warnings of when not to use such a process.

HD Screen Recording

For this example I used a Grass Valley Pegasus, High Definition Digital Recording System. This card enables the output from any video device, including DVI and RGB from a computer graphics card, to be recorded in its original resolution.

After completing a recording, the analysis and comparison of the two files can be made and reported on. This would be in the same manner as in the above two examples.

Transcoding

There are so many methods to conduct this process, that having some good initial verification reports will help others in using the right software, with the right settings, for the right purpose. Having a good understanding of compression and transcoding technologies is vital in the reporting stage. A book that I often refer to is Video Compression by Andy Beach but there are numerous available.

For the transcoding reports, it’s best not to be too specific. This will restrict a user’s ability to choose the correct settings for the task at hand. If you are asked to quickly send over a video clip to be presented at a meeting and they need to play it on a Mac, there would be little use sending a Windows Media Video as that was the only method available to them.

The reporting for transcoding could be task, or software specific.

Task Examples:

Video for Powerpoint

Video for preview

Archiving

Software Examples:

Expression Encoder

Squeeze

Free Video Converter

FVC is a GUI for AVCONV but comes loaded with editable presets

FVC is a GUI for AVCONV but comes loaded with editable presets

If a new task is identified that is going to crop up regularly, then a new report should be made on that task.

If a new piece of software is purchased or used, then a new report on how best to utilize it should be made.

Results

There is no point in using the Verify technique if you are not going to document the results. These results should be stored safely and then used if your capture method is called into question or, if you believe that it is failing to reproduce the image to a required standard.

If you are displaying your images on different systems / monitors then you will need to calibrate these and ensure that all the monitors are using the same colour profile.

When presented with a piece of software you have no experience or knowledge in.

Verifying your output is just as important when dealing with the myriad of CCTV Software players that we all have to use. It is imperative that you understand and assess the software’s ability to present the video to you in its correct form. You are then able to assess it’s ability to output the correct still images or video. Think of the amount of times that a player presents one thing but it actually recorded another. They will all be on a job by job basis, and you may not complete one every time. Once you have understood it for one job, you may not do it again until you are presented with another piece of unknown software. The process may be a separate document that can be referred to in your notes, or detailed directly in your initial analysis report.

After a process, in order to assess the quality of the result.

After any video file or Video DVD has been created, perhaps by using the Scan Converter to DVD recorder or by transcoding a h264 to WMV, it is important that your new video is verified. This may be as simple as watching it and ensuring that it shows the details required. Those details may be large so a slight reduction in video quality would not cause any loss in evidential quality. There are times though that after verifying the output, an unacceptable loss in quality is noticed. This is why, when it comes to changing video formats, it is not a good idea to have a single format and pre-set. You will also have to deal with output file size. If you need to urgently email a clip, then having this at 200mb might cause a few issues.

There is a fine line walked to get the perfect balance of file size, pixel size (height x width), and quality, all in a format that is acceptable for the end user and purpose.

If possible, you need to ensure that you test any output inside the software that is to be used by the end user. I have a number of examples where (for some unknown reason) a group of people were only using VLC to preview video. WMV files played fine in Windows Media Player but suffered with stuttering within VLC.

Disclaimer!

It would be naive not to mention that for Forensic Image Analysis, using a transcoded or incorrectly captured piece of video is probably not a good idea! The verify process is to ensure that any need for format shifting can be completed, and in a way that’s fit for its intended purpose. If the intended purpose is Image Analysis, ANY change to the make-up of the file can jeopardise its integrity. When format shifting for a specific purpose, some people add an intro at the start of the video to highlight a change in format in order to avoid any confusion or misunderstanding further down the line.

Advertisements
By Spreadys Posted in EEPIP

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s