Video forensics: How easy would it be to fake a Rob Ford video?
A Q&A with forensic analyst David McKay
According to the Toronto Star and the gossip website Gawker.com, there is a cellphone video for sale that purportedly shows Toronto Mayor Rob Ford smoking what appears to be crack cocaine.
CBC News has not seen the video and has not been able to validate any of the claims being made.
On Friday, Ford told reporters that the "allegations are ridiculous. It's another story with respect to the Toronto Star going after me and that's all."
However, the claims have now been widely reported all over the world, and Toronto city councillors called on Tuesday for Ford to respond to the allegations directly. But the mayor has continued to decline to do so.
On CBC Radio's Metro Morning, Coun. Peter Milczyn, a Ford ally, called on the mayor to give his side of the story. "Either he's adamant that it's untrue or he has another explanation," Milczyn said.
On Friday, Deputy Mayor Doug Holyday told CBC's Andrew Nichols, "We know that with the technical opportunities today people can make videos of anything," adding that "you're dealing with drug dealers and heaven knows what's going on here."
Holyday is asking people to reserve judgment until the actual video is produced. He also stated, "It's more important that the video be analyzed to see whether or not it's made up or whether it's really a true video."
So how is a video's veracity determined? For that, CBC News turned to David McKay, the president of Blackstone Forensics Ltd. in Vancouver.
McKay spent six years with the RCMP and has testified in court as an expert witness on video forensic analysis. He is also the manager of the B.C. Institute of Technology's Forensic Video and Surveillance Technology lab. The following is an edited version of our conversation.
CBC News: How easy is it to produce a phoney video that would pass a serious forensic analysis?
David McKay: It would be difficult to do that. You'd have to have the right software tools, and skill as well. You'd have to know what you'd need to do.
You'd also have to have some kind of knowledge of what kind of analysis would be done on the video so that you would know what areas you needed to conceal.
It sounds like you'd have to be pretty technically sophisticated to pull off a fake video that would get past serious forensic analysis.
McKay: Yes, you'd definitely have to have some good skills.
In the case of a video shot on a mobile phone, how important is it to have access to the phone it was shot on?
McKay: That's important, because you're looking at the technical data of the video and you want to be able to match that with the recording device, and that's going to be a pretty important part of the [verification] process.
With the Toronto video, the people who have seen it have said they saw it on an iPhone only. How easy would it have been for them to tell if that's the real thing?
It's going to be difficult just to visually observe a piece of video and say 'that's authentic.'
It's possible to do, in some cases, if you have seen a lot of video and you know the way video works and you understand it, it would be easier for you to pick up certain things.
So for forensic analysis, getting the information about who made it and so on is important.
How would you analyze the video for authenticity?
McKay: Once you've received the information, you're going to need to find out as much as you can about the recording, how it was made, who it was made by, the device it was made with, and then you're going to need to match that information with the recording you have.
Some of the things you're going to look at is the format that it's recorded in. For example if it's a common format like an .avi or a .mov format, we know that those types of formats can be created by numerous types of editing programs.
But if you've got it in, say, some type of iPhone specific format or other type of proprietary format, then we know it's going to be very difficult to create a fake that way.
What are some other clues that you look for when you're doing an analysis?
McKay: The metadata [that should come with the video] is important. You're matching up the recorded video and the settings of that video with the device that actually did the recording itself.
This is a video with audio, so how important is the audio in a forensic analysis?
McKay: It's going to be important. Less so from the technical sense, where you're looking at the metadata of the created file.
It will be important if you are trying to do voice analysis, comparing the voice that you have on the recording with the known voice of a particular individual.
If there's been compression, if the audio's gone through any kind of filtering, it can be difficult to make a conclusion on that.
The other thing to look at is how it is synced together. Is it just audio in the background? Or is the camera moving around so much that you can't actually see the words coming out of the person's mouth, but you're looking at the video and it appears they are coming out of the person's mouth?
So you need to have that clarity as well, too.
As we understand it, the Toronto video in question was one shot, no edits. Does that make it easier or harder to produce a fake?
McKay: If you've got one continuous shot and you know for a fact that there are no edits, then that's going to make it a lot more difficult to produce a fake than if you do have, what we would call, an obviously edited video.
At your lab, what equipment do you use for forensic video analysis for determining authenticity?
McKay: We use what's called an Avid detective system. An Avid editing system is kind of our main base of operations.
Then we have a bunch of different software programs that allow us to observe the metadata. We've got different tools that we can use to graph visual data of audio and video that allow you to look at spots that could have been edited or spliced.
How interested would you be in doing the analysis of the Rob Ford video?
McKay: We would definitely be interested in having a look at it.
How long would it take to do the analysis?
McKay: It's going to vary. You may have all the things you need to do an authentication right away, so there's no extra questions that need to be asked, no extra information that needs to be found out.
In many cases to do an authentication it's pretty straightforward if you look at the information that you have and all the data's there.
There are a series of technical checks you need to go through, however just because a video may fail a particular check, this does not necessarily mean it is fake, you just may not be able to confirm its authenticity at that point.
Then you need to explore the reasons why this technical check failed and on the balance of probability how impactful is that particular aspect on the overall authentication of the video as a whole.