Redaction by reduction

Pictures never tell the whole truth. Whatever truth they do offer is limited by technical qualities the general public, and most people in the legal community, simply do not understand. Most people think ‘a picture is a picture.’ Not so.

With the release of George Zimmerman’s video ‘re-enactment’ of the events leading to the shooting, and the recordings of his statements to police investigators, the people still paying attention to this case (i.e. not the mainstream media) are generally moving on to the contradictions between the GZ’s police call and his various statements, the changing details from one statement to the next, and so on. There’s plenty of interesting material here, to be sure, but the evidence from the first discovery dump remains quite relevant to the case, and there’s plenty there yet to be analyzed.

My argument here is similar to my umbrage over the release of poor quality audio of DeeDee’s interview with Attny. Crump when a clearer version must exist, since several sound bites from it were played on numerous national newscasts. AFAIK, no one besides me has made a point of this, and whonoze is probably one of the most obscure and least-read blogs on Martin/Zimmerman. (Help me out here folks! Spread the word!). The issue in this post is even more subtle, but yeah it really matters.

The still photographs from the first discovery dump have been posted on a few websites, usually in the form of web slideshows. In making these slideshows the photos have been greatly reduced in size to conserve bandwidth. For example, the photo of George Zimmerman’s jacket and shirt I included in the earlier post on Frank Taaffe is only 600 pixels wide.

I don’t know what camera was used to take this photo, or what settings, but even the cheapest ($80) brand-name digital camera available at Best Buy right now has a 10 megapixel sensor, meaning it can take pictures at 3648 x 2736 pixels. $100 gets you 16 megapixels, or 4608 x 3456 pixels. So I feel quite confident in assuming the original of the image above is a LOT wider than 600 pixels. As I said in the earlier post, this resolution seems high enough to confirm that GZ did not have any large bloodstains on the front of his clothing – no sign of a gusher-type bloody nose. But what about little drops of blood? Especially against that red jacket fabric.  People are going to look at a photo like this and say, “See, there’s no blood on Zimmerman’s front!” But this photo doesn’t show that at all. It’s inconclusive.

Lets look at the difference in another photo. I found this 8 megapixel image of Google guy posing with a University President on the left and NYC Mayor Bloomberg on the right. I reduced it to 600 pixelswide, which is what WordPress likes.

Now lets look at Bloomberg’s pins, blown up from this 600 pixel version so that the detail is also 600 pixels wide.

We still know it’s a flag pin, since the general shape is so distinctive, but if there’s a small spot on his suit jacket of some sort, you couldn’t see it. Now lets look at the same area at the original resolution:

Note that in the first pic, you can’t even tell the flag pin is next to a button hole.

But it’s not just a question of resolution, there’s also the amount of compression applied by the .jpg algorithm which is scalable in quality. All .jpg files are significantly compressed, but at the higher quality end (less compressed) the difference is so slight as to be unnoticeable. At higher compression/lower quality, you get a lot of artifacts. That pic of GZ looks fine at native res, but blow it up and you’ll see the artifacts. Here’s a blowup of the pic of GZ above.

See all those weird shadows, and the streaks of orange to the left of the zipper. Those are compression artifacts.

The resolution situation is even worse with the stills of the crime scene taken in daylight on a rainy day. One example here:

These photos were shot with a 35mm film still camera, which captures much more detail than even the best digital cameras. It certainly doesn’t look like this photo, taken a good number of days after the shooting, would have any secrets to reveal if we could view it in more detail… OK everybody, go rent the DVD of Antonioni’s Blow Up and watch it right now! The point is we can’t know if there’s something important in the detail unless we can see the detail.

I don’t think there’s any attempt by anyone to hide anything here. I think people just don’t get how media technologies work. I wonder if the State gave the defense original quality copies, or maybe intermediate quality copies. I remember seeing a post on some site (I forget where alas) where someone with a bit of legal background noted the reduced resolution of the photos with some annoyance but said they were “good enough for discovery.”

SAY WHAT? If I’m on trial for a felony, I want my attorney to have access to every bit of detail in the State’s evidence file. If someone mugs me and goes on trial for assault, i want the State to work with the most detailed evidence they can gather. I don’t want blurry, artifact filled lo-res .jpgs of photos. To attempt an analogy: if this is ‘OK for discovery’ then saying you ought to be  able to can submit abstracts or Reader’s Digest abridgments of written statements – you know, just give us the topic sentences in outline form, and whack out the rest of the content of each paragraph.

I wonder if the quality of AV materials released in discovery has ever been raised in a criminal trial, and how the court might have ruled on the issue.

Of course, I save the biggest joke for last: The release of a crime scene photo in the form of a Xerox:

The thing is, they don’t it’s a joke, and neither do most of the people commenting about the case.

Uff da! and Oy Vey!