decker labs

Recently I did some playing around with video and thresholds, mostly just for fun. I first experimented with just changing the threshold of the video, essentially converting it to black and white. Then I tried comparing the current frame threshold and the previous frame threshold to get the difference. The result looks like an outline/edge effect. It’s an interesting way to visualize movement. Then, just for kicks, I decided to display the previous comparison and the current one, colorizing the current difference as red and the previous as cyan. The result is a very pseudo-3D outline effect. This fake 3D technique can also be applied to to straight up video, showing the previous frame as cyan and the current as red and screening the two. Since this all based on movement and the idea that objects closer to the camera appear to move greater distances due to their perspective, it’s very easy for it to display incorrectly.

Check out the first demo. You’ll need a webcam. Click on the video to change the mode, and the two scroll bars control the threshold level and amount of blur applied to the video before any image processing.

I then noticed that while changing the threshold, it almost seemed like it was displaying different depth passes on the video. Of course, this has nothing to do with depth, but rather brightness. Still, I wanted to see what it would look like if I broke out multiple thresholds of the same frame and layered them in 3D. The result, although unpractical, still looks pretty cool.

Check out the second demo. There’s different options on the left and two scroll bars at the bottom that control layer transparency and amount of blur applied to the video before any image processing.

Recently I’ve been working on making a front end profanity filter. It makes more sense to have a filter be on the server side and return an accepted or rejected response, but since my strength is in AS3 I’ve decided to do it this way. There’s two parts I want to explain, the first being the reason for a profanity filter, and the second being the actual code. You can check out the demo here.

Why

If you have a site that lets users submit text and you want to control what is displayed, you’re going to need some sort of moderation system. There are two main options. The first is a profanity filter, which is a script that looks for certain words or word patterns, and allows or blocks them based on it’s parameters. The strength here is that it is automatic, but the weakness is that it can’t always catch everything, due to new slang, misspellings, context, and creative punctuation. Additionally, a filter that tries to block too much might end up blocking actually appropriate submissions. For example, for a website that lets users submit their favorite books, the filter might block “My Big Hairy Dick” but it might also block “Dick Tracy” just because it picked up on the word “dick.” Context is a major weakness, which brings us to the second option, which is a live moderator. The strengths of having a live person allow and reject is that they can easily see past misspellings, have a better idea on new slang, and most of all, they can interpret context. The downside is that the content cannot go live until it is approved, and if there are lots of submissions, that means more work for the moderator, meaning more time and more money.

My solution is to use both. First, a profanity filter is not enough, unless you don’t really care what gets by. But in most cases, clients want to make sure nothing offensive is posted to their site, both in terms of vulgarities and  brand defacement. The trick with writing a profanity filter is not necessarily blocking offensive words, instead it’s allowing appropriate words. For example, when to allow the word “ass.” You can block the word ass alone, but what about ass tied with other words, such as asshole, assface, asshat, etc. You can combine ass with many nouns to create something sounding [absurdly] offensive. And you cannot just block all instances of ass+word, because then you block assistance, assets, assist, etc. When writing  a profanity filter, the filter should fire little to no false positives.

But if there’s a person moderating content, then why bother at all with a front end profanity filter? The reason is to make that person’s job easier. The profanity filter’s job is to cut down on the amount of submissions by rejecting the without a doubt non-acceptable ones. If a site gets thousands of submissions and one person needs to go though them one by one and approve, that can take a long time. But if we can reject a handful from the start that will make that person’s job much easier. The profanity filter should never “assume” that something is offensive, – if something might be offensive (such as the “Dick” example above), then let it pass and the moderator will decide based on it’s context. I worked on a site that used this technique (in this case it was just a word list and a php script that matched literals), and we tracked the number of times users clicked submit and the number of times the profanity filter returned true (meaning the submission was rejected). The result was that over one third of submissions were blocked by a very basic filter, which gives me reason to believe that having both systems is defiantly a good idea. Now, as with the previous case, the filter could just be a gigantic word list congaing every variation of certain words, but it’s more fun to try and make a more “smart filter” that can take care of leetspeak and word variations on it’s own.

How

Rather than looking for exact words, the filter I wrote looks more or less for word patterns. This was a good case to dive into some regular expressions. The filter has three word lists (which can be set by the user, as this content should be external). The first word list are words to flag no matter what – words such as “fuck” that in the terms of the site are never appropriate or a part of any word. The second set is a list of words that are only to be flagged if written on their own, so we detect “ass” but not “assist” or “assface.” I also test for plurals and other suffixes automatically, so not only would “shit” be matched, but “shits,” “shitting,” etc. The third list is an allow list, words that might be exceptions to the previous list. For instance, block “crap” and “crappy” but don’t block “craps,” as craps is a casino game.

The first step is converting the word lists into regular expression patterns. The first step is that each word from the profanity list is split up so that various characters can be replaced. For instance, asshole can be written as a$$hole. So rather than match /asshole/ we replace letters with boolean values.  Instead we have something that looks like this: /a(s|$)(s|$)hole/.  When the regular expression is run, the ignore case flag is included so that capitals and lowercase are matched automatically. However, this does not work for accented characters (such as É and Ü), which I will mention some more about later.

In order to test for plurals and other suffixes, we add an optional ending to the pattern. This looks something like ((s|ing|er)?) and would test for [word], [word]s, [word]ing, and [word]er. We can also try and look for [word]y but first we must make sure that only y can be appended on. In some cases, usually when the last character is a certain letter (b,d,f,g,l,m,n,p,r,t,v,z)  and the character before that is a vowel, we duplicate that last letter. So “shit” becomes “shitty” instead of just “shity.” Of course, this is a very broad rule and won’t work in every case, which again is why there is a person moderating. Additionally, in rare cases by adding on the suffix will create a valid word. This is where that third list comes into play, were we can make sure we don’t match “craps” for crap.

Once our word list is build, it’s simply running the different RegEx patterns on the given text. Even with large bodies of text, this is still relatively fast, although I’m sure it would be tons faster on the server side. However, I don’t think there would be many cases where users would be submitting huge paragraphs of text. The class I wrote has two options: a normal validate function and a quickValidate function. The quick one simply keeps running patterns, and if it gets a match it terminates and returns true. The other validate function runs though each word and keeps track of the matches. It then returns a result object, that contains the validation status (true or false), as well as three arrays: a list of matches as they appear in the submitted text, the index start and end values for those matches, and a list of matches as defined in the filter list. This is useful if you need to do any syntax highlighting or anything other than just simply validating.

One issue I ran into building my demo was involving accented characters. Whenever there was a non-regularly used character, it would throw off the index position of that character. It seems for one reason or another, certain characters are considered to take up more than one index position according to RegEx. For instance, the em dash character is interpreted to take up three index positions. So if there is an em dash at index position of 5, and the letter “a” at 6, RegEx will think that the letter “a” is at index position of  8 instead. The TextField class does not think that these characters take up more positions than they do, so as in the previous example, if you tried to highlight the  letter a based on the returned index position, it would be three characters off (or if it were at the end it would throw an error). One workaround is right before validation run another RegEx that matches all these accented characters and replaces them with similar looking letters or just with an asterisk or something else – this is just for running the RegEx test, as the displayed text is left alone. This allows you to at least get the correct index values, and also will match É for E since the ignore case flag doesn’t ignore accents.

I created a demo that shows the matching though highlighting, as well as lets you edit the word lists. Keep in mind, this probably could have more uses than just for profanity.

Quick Example:

(Either JavaScript is not active or you are using an old version of Adobe Flash Player. Please install the newest Flash Player.)

Did a quick test this weekend. I wanted to see if you have a player 9 swf and load it into a player 10 swf, if there could be any issues. Mainly, I wanted to see if there could be problems with the new 3D properties in player 10. So I just created  a new player 9 swf that had some five3D elements in it, so essentially I had an extension of the Sprite class with properties such as z, rotationX, rotationY, and rotationZ. Since player 10’s Sprite class contains these properties nativity,  my theory was that importing the player 9 swf would throw errors because now those 3D properties of the five3D elements would not be properly overridden. This is exactly what happened.  Good to keep in mind in case you ever run into a situation where you might have to load an older flash 9 swf into a newer flash 10 shell.

I don’t know how this escaped me for so long, but the other day I realized that Flash doesn’t understand HTML character entity reference codes. Codes other than & and a few others won’t render out to their appropriate symbol. So for example, if I need to display a special character in an HTML page, I would normally use the &+code+; in order to display it. So “copyright ©2009 eric decker” would be “copyright © 2009 eric decker”. So it would make sense that if I bring in this copy into flash, set it to a textfield with the set htmlText method, that it would have the same result. However, it doesn’t. &copy will read exactly as &copy.  Flash does, however, understand numeric reference codes, which look like ©. Personally, I find the name-based entity codes easier to use and seem like they are more widely used (although I could be wrong about that, I’m not a HTML guy).

Anyways, determined to be able to use © instead of the more ambiguous © I wrote a class that is able to understand the entity codes. The reasons for this were 1. as I just mentioned, the name based codes are easier to read and 2. sometime you’re not in charge of your source copy, and it might contain entity based symbols instead of numeric and of course 3. developers are a little neurotic like that and sometimes ‘do it for the sake of doing it.’ (That could be  a T-Shirt.) I was able to write a pretty simple converter using a regular expression and a lookup dictionary. You can download the class or view it as text.

The class contains one accessible static function that you pass your raw text to and it returns the new formatted text: myTextField.htmlText = HTMLTextUtils.formatHTMLTokens(myString);  Better yet, if you have a custom TextField class that you’ve extended, you can override the set htmlText function to have it always use this function.

Grant Skinner’s RegExr is a huge help when doing anything with RegEx.

So I haven’t had much time to work on any personal projects recently, as my current project at Firstborn has picked up and I’ve been focusing 100% on that. Additionally, last weekend I was out in The Land that Law Forgot, better known as Las Vegas. I like to say the reason is because I am a professional sports photographer, but in reality it’s because I knew how to use a camera and won an inner-office Twitter contest.

Every year my boss along with a bunch of 35+ year old basketball fans fly out to Vegas to attend Michael Jordan’s Sr. Flight School. MJ  teaches them ways to improve their game as well as gives them all a detailed analysis of Space Jam. This year the camp didn’t happen, so the die-hard fans put on their own show and organized it themselves, sans Michael Jordan. I went along with some other Firstborners to help document the occasion though photo, video, and interpretative dance. If you’re interested, the photos are on flickr.  And yes, they actually did call it “Old Balls ‘09.”

We rented a Nikon D90, which was a great camera. It can shoot HD video in up to 5 minute segments, and while it doesn’t replace having a video camera it’s a nice addition to have. The feature that blew me away the most was how well it handles low light. Yes, shooting with and ISO of 6,400 becomes fairly noisy, but really what can you expect. Additionally, the noise on the high levels of the D90 are not that much worse then the higher levels (1,600 I think being the max) on my D70. And from what I’ve seen with the D700, it handles low light and extreme ISO even better. I can’t imagine what it will be like in another five years.

Oh, we also managed to steal a Bugatti Veyron,  hijack someone’s bachelorette party, get in a fight with a mob all dressed like Elvis, and ‘accidentally’ knocked a chair off of the top of the Palm’s Ghostbar. Okay, those things didn’t happen. Probably…

There’s been a lot of buzz around the Brazilian public service campaign “Pee in the Shower” (or Xixi no Banho). As far as I can tell, the campaign was created by Saatchi & Saatchi and features both TV and web. (It was featured on FWA as Site of the Day over a month ago.) The idea is that if you pee while in the shower, as opposed to finishing your shower and using the toilet like normal, the average household can save 4,380 liters (that’s 1,157 gallons for us US folk). Saving water saves the rain forest.  Sounds like a good plan, and the ads have gone viral (you’ve probably seen them online somewhere by now, right?) And it makes a very valid point – as soon as I get in the shower with the water running, I get that urge. But still, there’s a lot of negative feedback.

There are lots of people giving really negative feedback, through blog post and comments. Many claim it’s gross. Why? Why is urinating into a drain filthy, when the water is running away from you. (Yet somehow soaking in your own filth in a bath is totally fine.) Besides, the slogan is “Pee in the Shower” and not “Pee on your Leg.” Urine is actually sterile, and is harmless unless you have certain serious health issues, such as hepatitis. Another sad thing I’ve seen is a lot of people making derogatory remarks against Brazilians, calling them dirty and disgusting. Seriously, that’s really disturbing.

Besides, I think people might be missing the point. I doubt that the main goal of the campaign is to get people to actually pee in the shower. If you do, that’s great, but what the message here is that even little changes can add up and make a difference.  Think about it, how insignificant is peeing in the shower, yet doing so saves you  flush, and those flushes ad up. Tiny actions eventually add up. Turn off lights when you’re not in the room, turn the AC off when you’re out, we’ve all heard this before – this is just the first time we seen it presented in a such an effective or creative way. It’s a way to spread the message. So go ahead, pee in the shower, it’ll help save the rain forest. Plus, it’s fun…

So over the last week or so, I restructured my 3D engine. Nothing that would be visibly noticeable, but I feel better about where it is at right now. What I like about it, though, is that it is still not what would be probably considered “the right way” to do it. The reason I am happy about this is that it makes it my own, and I understand exactly why and how it works. Once I have a really solid understanding, then I can move towards doing it the “correct way,” but for now, my little hacked together experiment makes me happy.

One of the changes I did make is to the shaders. I added a PseudoGouraudShader, which uses gradients to draw each polygon face to try and smooth out the hard lines created by using a flat shader. I doubt this method would ever produce 100% accurate results that are identical to a gouraud shader, but it makes for interesting results. The method I am using is pretty slow, as it needs to draw each face four times, three of them being gradient fills. Each vertex’s normal is calculated from the sum of all the faces it is a part of, and that brightness value is used to draw a radial gradient on that face, the center point being the vertex. (Two examples: RGB example and dot example, the circle areas are ‘painted’ on via gradient fill) This can produce some fun effects too when played with.

I had an example of the PseudoGouraudShader last week, but it was off, it looked like alternating faces had different brightness values. Just playing around last night I found the issue, it lies with how I was calculating my normal.  When getting the cross product of the vectors, it is better to normalize the result instead of normalizing the vector before. The example below shows the two results and code snippets:

Still not perfect, but I’m not sure how close I can get using this method. I think there might still be some slight errors in the core math, which could be producing the results that look similar to the image on the left (although less intense). You can check out the live demo here.

On a last note, calculating normals and centroid over and over again for rendering can get pretty intense. In the case like this, the same normal gets calculated multiple times a render even though it hasn’t changed (because each vertex needs to get the normal of each face it is connect to). In order to boost speed, I created a caching system. When the normal is calculated, it saves it to a “stale normal” and updates an object (which I called the cached face) that tracks the position of each vertex in the face. Then, the next time the normal is called, if that cached face is not different then the real one, it just uses the stale normal instead of re-calculating.

Another experiment recently did was compress three grayscale images into one file. The reason for this would be of course file size and load time. A real-case scenario where you might need to load 3 grayscale images of the same size is loading textures for a 3D app. For instance, your model might be color dynamically, but you need to load a diffuse map, a specular map, and a bump map – all of these would be the same dimension and would be black  and white.

So all you need to do is take your three grayscale images into photoshop and layer them on top of one another. Then use the gradient map image adjustment to have each image range from red, green, or blue to black. Have the base layer be flat black, and set each of the three layers to screen. Then, once that image is loaded into Flash, you can extract each color channel and set it back to black and white with the ColorMatrixFilter. You just need to set the matrix so that each of the R,G,B rows are [0,0,0,0,255] and the last row (the alpha) is [1,0,0,0,0] for red, [0,1,0,0,0] for green, or [0,0,1,0,0] for blue. The just smoosh that transparent white image into a black bitmap, and you’re set.

Again, not entirely useful for every day cases, and it does cut down on how easy and dynamic the loaded content is, but if you’re hurting due to file size and load time, this could be something to keep in mind. In case you’re wondering – the layered color image I used in my example is 188 KB, and the blue channel saved as its own grayscale file is 132 KB.

RGB layered image and the 3 channels as their own grayscale image

Playing around in Flash making a ‘image encoder’ that converts text into image, or String into BitmapData. Basically, I wanted to convert each character into binary, and then represent it as an image. So the first thing I did was have each character assigned a number, and then convert that into binary. Since I wanted to control what characters could be encoded, I created an array of the characters I wanted (instead of using the charCode) and used the index of the character as it’s id. Each character is converted into a 7-digit binary value (this value was based off of the array length) and then each digit is written to bitmap data, 0 as black and 1 as white.

I tried the same thing, but except with a color image. Instead of having 7 pixels representing a character, I used one pixel per character, converting each charCode into a hex value. The result is that the image is all blue, as the common used character codes are all under 255. Not a very useful exercise, but it was still sort of fun, as was converting a number into binary “manually.” (It involves a lot of use of modulo. )

The image bellow is the first chapter of Lewis Carroll’s Alice in Wonderland encoded into an image. Note: I chose Alice in Wonderland because it is copyright free, not because of the new trailer for Tim Burton movie. There’s a larger example here with the color image as well.