Key Abdi video unreliable, expert tells officer's manslaughter trial

The surveillance video that captured the violent altercation between Abdirahman Abdi and Ottawa police can't be used to judge speed, timing or force, a forensic video expert explained at the manslaughter trial of one of the officers this week.

Const. Daniel Montsion has pleaded not guilty to manslaughter, aggravated assault and assault with a weapon in Abdi's death in July 2016.

• Why none of the videos showing Abdi's arrest may ever be shown in court
• Montsion trial on month-long hiatus while judge ponders charter question

Surveillance video from the lobby of 55 Hilda St. shows Montsion appearing on the scene to find another officer, Const. Dave Weir, his baton extended, facing Abdi.

Montsion, who was wearing reinforced gloves,punched Abdi twice before Weir pushed Abdi to the ground. Montsion punched Abdi twice more in the head while the officers worked to get him into handcuffs.

Abdi lost vital signs soon after, and officially died in hospital the next day.

The video has the potential to become a key piece of evidence at the trial, butaccording to forensic video analyst Grant Fredericks, its poor quality leaves key elements,including the speed and force of the officers' actions, open to misinterpretation.

The defence plans to argue that the video is so misleading, it shouldn't be entered as evidence.

WARNING: Video contains graphic content

6 years ago

Duration 5:51

A longer version of the video secured as digital evidence in the trial of Const. Daniel Montsion. It appears to show Abdirahman Abdi being kicked and punched by police officers in front of the apartment building at 55 Hilda St.

Different opinions

Fredericks, an instructor with the FBI National Academy,is a certified forensic video analyst who teaches video science to police chiefs around the world.

Montsion's lawyers have retained him to challenge the findings of the Crown's video expert, Ed Segeren, a broadcast post-production editor, who testified earlier in the trial.

Segeren and Fredericks agree the original surveillance video has a variable frame rate, which makes the action on camera appear jerky.

Segeren told the court in May that he was able to reconstruct the video by taking a still image of each frame and stitching them together to produce a true and accurate depiction of the eventsas they happened.

WARNING: Video contains graphic content

6 years ago

Duration 0:26

The Crown's video expert deconstructed the surveillance footage by creating an image out of each individual frame, then used video editing software to stitch the images together into this new video file.

No way to salvage video, expert says

But the underlying data associated with each frame of the original file shows that the images appear to have beentaken at random intervals, Fredericks explained.

While some frames were captured roughly 30 milliseconds apart the frame rate you would see on TV others were capturedas few as 10 milliseconds apart, or as many as 31 seconds apart.

Among other issues, Segeren's method forced the video to play at the wrong speed throughout, Fredericks said.

"It could lay the seed for somebody watching it to misinterpret it, which I think could be dangerous," Federicks said.

In fact, given how the surveillance video was recorded in the first place, there's no way to accurately recreate the speed of events, he said.

CBC Explains: Here's why some experts say the surveillance video is unreliable

5 years ago

Duration 0:26

A forensic video expert says the surveillance video can't be used to judge speed, timing or force because of its variable frame rate.

Fredericks has not yet been cross-examined by the Crown. It's not clear when his testimony will resume.

Once his testimony is complete, the defence plans to argue the video should be thrown out as evidence.