Mental Masturbation, Musings, and Methods » Miscellanious

Convert Quicktime Screencasts with FFmpeg

Alex — Wed, 28 Jul 2010 00:42:05 +0000

I was trying to convert a Quicktime screencast (saved as .mov) to a format my professor could more readily use. For those who don’t know, FFmpeg is an amazing tool with a wide range of applications, including converting between video formats and ripping audio from a video. Anyway, this took me a little to get quite right (and maintain output quality) so I thought I’d post it here both for others and a place to save it for myself. Although it looks really simple in hindsight:

ffmpeg -i screencast.mov -f avi -vcodec mpeg4 -sameq -ac 2 -ab 128k screencast.avi

It is necessary to specify a video codec because AVI can be used with multiple codecs. To maintain video quality use -sameq. When converting screencasts, it may also be useful to crop out parts of the screen since Quicktime only does full screen capture. For this, you can use the -croptop, -cropbottom, -cropright, and -cropleft arguments. Last, make sure the arguments come before the output file; otherwise they are not applied.

Maybe I will later do a more full post of the different ways I use FFmpeg. But for now, that is all.

Hilbert Curve with the Canvas element

Alex — Thu, 01 Oct 2009 22:56:48 +0000

For my discrete math course (CPS 102) this semester, I had to create a space-filling curve for one of my homework assignments. I decided to take this as an opportunity to play with HTML5′s canvas element since I’ve yet to had a legitimate use for it in any programming yet. I think it came out decently well (and it’s kind of fun to play with), so I thought I’d post it. Here is a link to it on my Duke account http://duke.edu/~amb79/fillSpace.html Like I said, nothing huge but fun to program and cool to watch.

I am on fall break now, and will be spending the weekend at home. Hopefully I’ll do something cool while I’m home and will post if anything comes out good.

Image Aliasing of Plane Propellers in Photos and Video

Alex — Fri, 24 Jul 2009 05:33:27 +0000

Recently, while flying between Durham and New York, I decided to take a picture of the propeller during take off. I don’t know what prompted to me to do this, but when I pointed my iPhone camera at the propeller out my window I noticed very odd pattens. I took multiple pictures throughout the flight with my iPhone, and in some cases at the same time took pictures/video with my Canon camera. To clarify the specs of the two cameras: I have an iPhone 3G (8 GB) and a Canon SD700 IS. As you will see in the pictures below, the iPhone’s picture simply makes no sense with the blades appearing to be disconnected vertical lines. The photo from the Canon seems reasonable, but as you will see in the video the propeller appears to be going fairly slowly, when in real life all I could see was a blur because of how fast it was moving.

Plane Propeller Photo from iPhone 3G

Canon SD700 IS Plane Propeller Photo

Update: here is a video taken from my iPhone 3G of the propeller.

This got me thinking about why the images looked the way they did. I assumed they were related to frame rate and as a result wondered if I could work backwards to find the speed of the propeller. When I got back to Duke, I talked it over with some of my friends. We discussed one theory that explained the slow spinning in the video from my Canon. I later found out is similar to the Wagon-wheel effect, and will go into more detail on this later. We also discussed possible distortions and other possible causes but couldn’t figure out exactly what was causing the odd pattern in the iPhone photo. Eventually, I decided to search around the internet and found that I was not the only one to notice this oddity. I also stumbled upon this great post explaining clearly and graphically why the iPhone photo looks the way it does. In short, most cameras save pixels from all parts of the lens at once, while the iPhone camera goes through the lens saving pixels linearly over time, similar to watching a scanner save the page of a book. (If anyone has an iPhone 3GS and can snap a picture and/or video of a propeller, please let me know. I’m very intrested to see if Apple’s upgraded camera still behaves this way. Please let me know if you find out.) The writer of the post also demonstrated this in a cool video made from a Matlab script:

Personally, I think this is really cool so I decided to play around with it some more. As I mentioned earlier, my friends and I had discussed the wagon-wheel effect or the stroboscopic effect, so I thought I’d modify his script to demonstrate it more clearly since the Wikipedia pages on the two topics both seemed a bit lacking. First, it is necessary to explain the general concept. The stroboscopic effect explains if you only see flashes or quick individual frames of a moving object at regular intervals you can get a distorted view of the motion, especially if the motion of the object that is being watched is cyclic or has a frequency. The relationship between the frequency of the flashes or frames and the frequency of the motion is what determines the perceived motion. The most common case of this is seeing a wheel which you know to be rolling forward, but it looks to be spinning backwards slowly. I modified the previous Matlab code to create a basic example of this with the propeller (download the code here). As you will see, the propeller on the left will spin counterclockwise at a moderate speed and every 8 frames the image on the right will update with a copy of the image on the left. This out of sync updating will make the propeller on the right appear to rotate clockwise.

Click the image to view the video

I also decided to show this more mathematically by plotting these movements in Mathematica. In the following examples I graph both the y-coordinate of the tip of the propeller and the angle that the blade has rotated in radians. In the graph of y the coordinate, there are three sinusoidal functions representing 3 propellers, each offset from the previous by 120 degrees (thus equally spaced). These lines and their colors correspond to the colored lines on the radian vs. time graph. The red lines mark a flash or frame and the dot on the line shows the top-most propeller. As you vary the spacing of the lines and thus change the frame rate, you can see the propeller will appear to stay still, then go backwards, then flip back and forth with no perceived motion, then go forwards, until it appears to stand still again. Below are pictures showing each of these stages, followed by a video fluidly showing these variations.

Here is the propeller appearing to stand still. This is due to the fact that the spacing of the frames is twice the period.

The graphs below show the propeller appearing to go backwards. Here we see that the frame rate is just below two periods of the propeller. The downward slope on the radian graph is a simple quick way to see the direction of the propeller rotation.

Propeller appearing to flip repeatedly:

This example shows the propeller appearing to rotate forwards. Again, the upwards slope of the points on the radian graph is an easy way to see the direction of rotation.

Last, here is the propeller appearing to stand still again. Unlike the first case of the propeller standing still, the spacing of the frames is not a multiple of the period. Rather, we see here that it is a multiple of 1/3 the period. It is only necessary that the spacing be a multiple of 1/3 the period because there are three blades and the blades are indistinguishable.

And finally, putting it all together:

The Mathematica code can be viewed here or downloaded here, and a printout of the entire Mathematica worksheet can be found here. Last, if one had a camera which could vary the frame rate just so slightly as to find two nearest frame rates to show the propeller standing still, the frequency of the propeller could be calculated as shown below.

As for the iPhone picture, a single image (and knowing the number of blades) is enough to roughly approximate the speed of the propeller as shown below.

However, just as a warning both of these are back of the envelope calculations and it is very possible I made a mistake (if so please let me know). That is all for now. Nothing ground breaking, but as the title suggests, it was a good chance for some mental masturbation, playing with Mathematica and Matlab.

Duke LDAP on the iPhone

Alex — Sun, 19 Jul 2009 23:11:41 +0000

I got an e-mail this morning from someone who saw my post on Twitter about using LDAP on the iPhone here at Duke. The person said they followed the tutorial but couldn’t quite get it to work. So I figured it may be useful to post a few quick instructions on setting it up. As the post on Twitter explains, the basic instructions for getting to adding an LDAP account can be viewed here and Duke’s LDAP server settings can be found here. Once you are set to add the LDAP account, simply set the server to ldap.duke.edu, no need for a username or password. Once you click next you will want to turn off SSL and change the Search settings for duke,edu to have a base value of “ou=People,dc=duke,dc=edu” (leave the scope at subtree). And that’s it. To use it, which turned out to be the issue for the person emailing me, go to write a new email and start to type in the name of the person to search. You should see a throbber show up at the top of your screen, showing that it is connecting to the server. There is a bit of a lag but it should eventually return the contacts for people by that name. Below are some pictures of the settings and the search. Let me know if you run into any issues.

Chromium on OS X

Alex — Wed, 18 Mar 2009 06:43:48 +0000

I tried this over break but didn’t have enough room on my internal hard drive, but now that I’m back at school downloaded the source for chromium to my external (took up about 10 gig) and built it (about another 5 gig). And it actually runs! Its definitely still in development, crashes on all sorts of little things, but its cool to see it progressing and it is definitely lightweight right now. Odd thing you’ll notice in the screenshot is that half of the Chromium processes in Activity manager show as being Not Responding. Anyway, here’s a screenshot just because I thought it was cool. Click for full size.

Catching up

Alex — Thu, 19 Feb 2009 19:32:05 +0000

Haven’t posted in a while and having some free time for the first time in a long while, I thought I’d write something. After having a few friends becoming obsessed with Twitter at The Chronicle (Duke’s student newspaper that I am partially involved with – I’ll write about this at some point), I can’t bring myself to write a full blown 140 character tweet but should be able to write a good few paragraphs here. Anyway, I think I’ll delve a little more than I have previously into my academics. I am currently taking Physics 181 (Intermediate Mechanics), Physics 176 (Thermal Physics), Math 107 (Linear Algebra and DiffEq), CS 104 (Comp. organization, essentially assembly), and cultural anthropology 181, which is about musical protests. Interesting mixture of courses and constantly changing my view on the different subjects. I have to declare my major within the next two weeks and am planning on officially declaring as a double BS in physics and computer science. However, I am not sure what that will actually result in, as I would like to shape my CS degree to be more oriented towards CS theory and less on systems and large projects. On the other hand, shaping my major this way means I would need to either not get an official CS degree although I’d have more than enough courses, or getting an interdepartmental major, which the DUS of physics seems to be very against. Neither seems ideal, but I figure I’ll take what interests me and what helps me further my work and knowledge, and let that philosophy take me where it may; I figure it can’t lead me that far astray, especially if I don’t even know what path I am straying from. Anyway, that’s where I am at right now.

Yesterday at work (my job with Duke OIT) we finished up my code for the new version of Duke’s online AFS interface – Webfiles, which should be tested and rolled out in coming weeks. I believe we will be making it open source, which I think is exciting. This led me to my next endeavor today, which was to finally install Git and integrate it with Textmate. As usual when I do small new things like this that change what I am used to, I realize how little I know about some parts of computing, especially Unix. I am thinking about forcing myself to use Vim for a little while and properly learning Unix. New great command: which. After installing things with macports like mysql or git, and with my not great knowledge of Unix systems this command is really helpful for integrating it with the rest of my OS X installation (such as letting Textmate plug into it).

Ironic, I know that I started off the post explaining I don’t like CS systems courses and finish the post by explaining the cool new Unix command I learned, but I think I am comfortable with that paradox. Random Unix commands and little tidbits like that seem more like something I will continue to learn in practice, not something I should spend a semester on in one of my few semesters left here. Anyway, this was a nice exercise in public introspection. As I always say, I’ll try to write soon; hopefully I’ll actually do that one of these times.

Skype and Verizon Wireless

Alex — Tue, 23 Dec 2008 20:24:28 +0000

At the start of this break I was frustrated with the fact that my house is a dead zone for Verizon and as a result I missed any and all calls while I was home. To deal with this I decided to finally properly configure Skype to work in sync with my cell phone. I forked over $20 and got a 3 month plan for unlimited calling through Skype, a phone number for my account and voicemail. I also set it up to display my cell phone number when I call from Skype so as not confuse people. To have it work in sync literally all I had to do is dial *72 followed by my Skype phone number and all calls are automatically forwarded to my computer. And when I leave the house all I have to do to have calls go back to my cell is dial *73
and everything goes back to normal. It is simple and it works.

To my surprise just a few days layer I switched to AT&T and got an iPhone (from which I am currently typing this post). So service in the house is no longer an issue, but I still occasionally use the Skype calling. Overall, I definitely recomend this Skype configuration if service is an issue for you. I will post more about the iphone soon.

Break from Studying

Alex — Fri, 12 Dec 2008 22:28:58 +0000

In the middle of studying for some finals, I thought I’d take a break, install the new WordPress 2.7. Of course, this required I actually write a blog post to give the new version a legitimate spin. So just wanted to post that my new favorite command (after rsync) has been wget, installed via macports, which I used to download all of my notes for a few classes off my teachers’ websites. Again, another small amazement in Unix, and thus OS X. Maybe over break I’ll pick up a book and really learn its full potential. Anyway, I’ll try to write a few more blog posts over break with whatever I’m working on or any other exciting news. Back to studying.

Welcome!

Alex — Mon, 10 Dec 2007 19:30:02 +0000

So I decided to start a blog, mostly just to post things I find while wasting time on the internet; don’t expect to read much about my life, it isn’t that exciting. I have a good amount of catching up to do as far as posting things I find interesting online, so hopefully in the next few days, assuming I don’t become completely consumed by my math final, I will post some interesting things. Enjoy.