GenuineJD

Flex Ant flexTasks, Manifests and Library Includes

Wed, 14 Dec 2011 23:46:00 -0500

I recently was working on a framework project where one of our front-end developers had written several UI components in Flash Profressional and compiled them into a SWC for me to leverage in an ActionScript project in Flash Builder. Aside from the compiled SWF, we also had to deliver a SWC to third-party vendors that were developing content against our framework. The SWC had to include a small subset of event classes from the ActionScript project along with the UI components library from my front-end dev.

The first step was to leverage the flexTasks Ant library included with the Flex SDK to use the compc task. This is simple enough by specifying the classpath to the flexTasks.jar. The next step was to create a manifest file specifying the classes in my ActionScript project that should be included in the SWC. This involves defining a namespace for the manifest, then including the namespace in the compc task. The final step was including the externally compiled SWC for the UI components. As simple as it turned out to be, this was a much more difficult task to accomplish, primarily because of the lack of good documentation and examples from Adobe.

Below are examples of the manifest and ant builder files I ended up using for the project. Keep in mind some variables Ant variables are defined outside of the builder file itself as Ant properties in the builder launch configuration setup in Eclipse.

manifest.xml

<?xml version="1.0"?>
<componentPackage>
<component class="com.genuinejd.events.NavEvent"/>
<component class="com.genuinejd.events.LayoutEvent"/>
<component class="com.genuinejd.events.ReferenceEvent"/>
<component class="com.genuinejd.events.ModalEvent"/>
</componentPackage>

builder.xml

<?xml version="1.0" encoding="UTF-8"?>
<project name="SWC Builder" basedir=".">
<property name="SOURCE_DIR" value="${basedir}/src" />
<taskdef resource="flexTasks.tasks" classpath="${FLEX_HOME}/ant/lib/flexTasks.jar" />
<target name="main">
<compc output="${OUTPUT_DIR}/Lib.swc" >
<source-path path-element="${SOURCE_DIR}/"/>
<namespace uri="http://genuinejd.com/2011" manifest="${basedir}/manifest.xml" />
<include-namespaces>http://genuinejd.com/2011</include-namespaces>
   <compiler.include-libraries dir="${SOURCE_DIR}/../libs" append="true">
   <include name="UI_controls.swc" />
</compiler.include-libraries>
</compc>
</target>
<target name="clean">
<delete>
<fileset dir="${OUTPUT_DIR}" includes="Lib.swc"/>
</delete>
</target>
</project>

Building FFmpeg for Windows on Linux - Part 2 - Setting Up the Build Environment and Extra Libraries

Thu, 07 Apr 2011 15:23:00 -0500

In the last article, I explained how to install Ubuntu 10.10 64-bit using the VirtualBox software and get the instance of Ubuntu all up to date. This article will focus on getting our build environment set up for compiling FFmpeg. This part is arguably the most difficult as it involves not only setting up the compilation tools, but also cross-compiling additional libraries to be used with FFmpeg that are not included by default. If you using the same exact software and versions as described in my introduction, this should be as easy as following directions step by step. Let's get started...

Open a terminal window in the Ubuntu guest system (Applications -> Accessories -> Terminal). Most of the work will be performed here for the rest of this series.
Most of the hard work has been done for us by a gentleman who maintains the site at arrozcru.org. To use his libraries for cross-compiling, download and import his gpg key and update your packages list
- Download and import the gpg key to your system
  
  ~$ gpg --keyserver www.keyserver.net --recv-key 0x25E635F9
  ~$ gpg --export --armor 0x25E635F9 | sudo apt-key add -
- Add the apt repositories to your sources.list file
  
  ~$ echo "deb http://apt.arrozcru.org ./" | sudo tee -a /etc/apt/sources.list
  ~$ echo "deb-src http://apt.arrozcru.org ./" | sudo tee -a /etc/apt/sources.list
- Update your package list
  
  ~$ sudo apt-get update
Set up and install the cross toolchain. If you are working with a version of Ubuntu other than 10.10 or the 32-bit version, be sure to follow Option 1 to recompile the toolchain. If you are using Ubuntu 10.10 64-bit, be sure to follow Option 2.
- Option 1
  1. Create cross toolchain directory and install dev tools
    
    ~$ mkdir cross
    ~$ cd cross
    ~/cross$ sudo apt-get install dpkg-dev debhelper autotools-dev libmpfr-dev libgmp3-dev libppl0.10-dev libcloog-ppl-dev libmpc-dev
  2. Install the mingw api and runtime
    
    ~/cross$ sudo apt-get install mingw32-w32api mingw32-runtime
  3. Recompile and deploy mingw binutils
    
    ~/cross$ apt-get source --compile mingw32-binutils
    [lots of output not shown here]
    ~/cross$ sudo dpkg -i mingw32-binutils_*.deb
  4. Recompile and deploy mingw gcc 4.4
    
    ~/cross$ apt-get source --compile mingw32-gcc-4.4
    [lots of output not shown here]
    ~/cross$ sudo dpkg -i mingw32-gcc-4.4_*.deb
  5. Install mingw binutils and gcc
    
    ~/cross$ sudo apt-get install mingw32-binutils mingw32-gcc-4.4
- Option 2
  
  ~$ sudo apt-get install mingw32-w32api mingw32-runtime mingw32-gcc-4.4 mingw32-binutils
Patch the cross toolchain with Arrozcru's patches
- Install patch to patch the cross toolchain
  
  ~$ sudo apt-get install patch
- Download patches from Arrozcru
  
  ~$ mkdir mingw32-patches
  ~$ cd mingw32-patches
  ~/mingw32-patches$ wget http://fate.arrozcru.org/mingw32/patches/tempnam.diff
  ~/mingw32-patches$ wget http://fate.arrozcru.org/mingw32/patches/strcasecmp.diff
  ~/mingw32-patches$ wget http://fate.arrozcru.org/mingw32/patches/nomt.diff
- Patch the cross toolchain
  
  ~/mingw32-patches$ cd /usr/i686-mingw32
  /usr/i686-mingw32$ sudo patch -p2 < /home/corpapps/mingw32-patches/tempnam.diff
  /usr/i686-mingw32$ sudo patch -p2 < /home/corpapps/mingw32-patches/strcasecmp.diff
  /usr/i686-mingw32$ sudo patch -p2 < /home/corpapps/mingw32-patches/nomt.diff
Compile extra libs as needed
- Get and build libfaac
  1. Download and extract faac-1.28.tar.bz2
    
    ~$ wget http://downloads.sourceforge.net/faac/faac-1.28.tar.bz2
    ~$ tar -xvf faac-1.28.tar.bz2
  2. Configure and build libfaac
    
    ~$ cd faac-1.28
    ~/faac-1.28$ ./configure --host=i686-mingw32 --prefix=/usr/i686-mingw32 --enable-static --disable-shared --with-mp4v2=no
    ~/faac-1.28$ make
    [lots of output not included here]
    ~/faac-1.28$ sudo make install
Install extra libs from Arrozcru's apt repository
- You will need to determine what libraries to install using
  
  ~$ apt-cache search mingw32-
- Install each library one at a time using sudo apt-get install mingw32-<pkg-name> or copy/paste the string below
  
  ~$ sudo apt-get install mingw32-ocaml mingw32-libmp3lame mingw32-bzip2 mingw32-libcloog-ppl mingw32-libsdl mingw32-libppl mingw32-libx264 mingw32-libopenjpeg mingw32-polarssl mingw32-libmpc mingw32-libfaad2 mingw32-librtmp mingw32-libtheora mingw32-libxvid mingw32-libmpfr mingw32-openssl mingw32-libgmp mingw32-libschroedinger mingw32-liborc mingw32-libvorbis mingw32-libogg mingw32-libspeex mingw32-pthreads mingw32-directx mingw32-zlib mingw32-libgsm mingw32-liboil mingw32-libopencore-amr

That's it! The hardest part of cross-compiling FFmpeg for Windows is in the setup; getting the cross-compiler and the libraries ready. One thing to notes is that several times when trying to download the gpg key for arrozcru, the keyserver timed out. I did have luck removing the keyserver argument and attempting to download it with the following command

~$ gpg --recv-key 0x25E635F9

In the next and final article, I will outline how to setup the directories we'll use for downloading the FFmpeg source, configuring and building it.

Building FFmpeg for Windows on Linux - Part 1 - Setting Up the Virtual Environment

Mon, 14 Mar 2011 15:27:00 -0500

The first part of this tutorial will cover setting up the virtual instance of Ubuntu 10.10 on a Windows 7 64-bit laptop using VirtualBox. The steps below will outline the entire process to get your virtual evironment ready to compile FFmpeg.

Download and install VirtualBox for Windows hosts
http://www.virtualbox.org/wiki/Downloads
Download Ubuntu Desktop Edition 10.10 64-bit ISO
http://www.ubuntu.com/desktop/get-ubuntu/download
Create a new VirtualBox virtual machine and install Ubuntu from the downloaded ISO
- VM Name and OS Type
  - Name: Ubuntu 10.10
  - Operating System: Linux
  - Version: Ubuntu (64 bit)
- Memory: 512 MB (Default) - I increased this to speed up the process a bit. If you have more RAM to allocate to this VM, go ahead and do so.
- Virtual Hard Disk
  - Boot Hard Disk: Checked
  - Create new hard disk: Selected
    - Storage Type: Dynamically expanding storage
    - Virtual Disk Location: Default
    - Virtual Disk Size: 8.00 GB (at least 6 GB is recommended)
- Add the Ubuntu ISO as a CD Drive
  1. Select the newly created VM and click Settings
  2. In the Settings dialog, select Storage from the left side menu
  3. In the Storage Tree, select Empty under IDE Controller, then click the folder icon to the right of the CD/DVD Device drop-down to open the Virtual Media Manager
  4. Click the Add button to add the Ubuntu 10.10 ISO that you downloaded above
  5. Select the ISO that you just added and click the Select button
  6. Verify that the ISO has been added to the CD/DVD Device and click the OK button to close the Settings dialog and accept the changes.
- Start the VM and follow the Ubuntu installation instructions
  1. When Ubuntu first loads from the CD (ISO), it will prompt you to either Try Ubuntu or Install. Select Install.
  2. Preparing to install Ubuntu
    - Download updates while installing: Unchecked. We will install updates after the installation
    - Install this third-party software: Unchecked. This is not necessary for our purposes.
  3. Allocate drive space: Erase and use the entire disk (use defaults on next screen [select drive])
  4. Click the Install Now button
  5. During installation, select the appropriate options for timezone and keyboard layout
  6. Who are you? Create a username and password, be sure to either write down this information or remember it so you can login to Ubuntu later.
  7. Sit back and wait for Ubuntu to finish installing, it takes about 10-15 minutes depending on the allocated resources of the VM.
  8. Once the installation has finished, you'll be prompted to restart. Go ahead and restart, and if it prompts you to remove installation media, do so by opening the menu Devices -> CD/DVD Devices and click Ubuntu 10.10 ISO to un-check it.
  9. After restart, login to Ubuntu with the username/password entered during installation
Run all Ubuntu updates, either via Update Manager or apt-get. Either option will take some time to download and install updates. Get and install updates until there are no updates left.
- Update Manager
  1. Open the Update Manager in Ubuntu under the System -> Administration menu (it may open automatically after having been logged in for a minute)
  2. When the Update Manager opens, click the Install Updates button and enter your password when prompted to Authenticate
  3. Wait for the updates to install, this may take several minutes
  4. After the updates are installed, restart and log back in
- apt-get
  Open a terminal (Applications -> Accessories -> Terminal) and enter the following commands
  
  ~$ sudo apt-get update
  ~$ (output removed)
  ~$ sudo apt-get upgrade
  ~$ (output removed)
Install VM Additions - this will make it a little easier to interact with your guest operating system
1. In the VM host window, select Devices -> Install Guest Additions... to "insert" the VM Additions Cd into the VM
2. Mount the CD in the guest VM by selecting Places -> VBOXADDITIONS_x.x.xx_xxxxx
3. This should cause a "Places" window to open which indicates "The media has been detected as 'UNIX software'. Click the Open Autorun Prompt, then click the Run button when prompted to confirm
4. Enter your password if prompted to perform administrative tasks and click the OK button
5. VM addittions will open a new terminal window and be uncompressed, built and installed
6. When it is finished, press Return (Enter) to close the build window
7. Restart to finish enabling VM additions
8. After restarting and logging back in, you can "eject" the VM additions CD by right-clicking the desktop icon, and selecting Eject, then under the VM host window, select Devices -> CD/DVD -> VBoxGuestAdditions.iso to un-check the ISO.

Your virtual instance of Ubuntu has now been installed and updated. At this point, you can do with it what you like. In the next part of this series, I will cover the steps needed to set up the build environment and all required libraries for FFmpeg.

Building FFmpeg for Windows on Linux - Introduction

Wed, 09 Mar 2011 19:04:00 -0500

At work, we deal with a lot of video production. Those videos need to constantly be reviewed and approved by various legal staff. To support this workflow, we decided to implement a video reviewing and management application. The web application was the easy part. Our technology stack was based around a lot of Adobe products, so we decided on Flash Media Server for our streaming infrastructure. The difficult part was getting the video files in a format to stream using our streaming technology. Encoding solutions that could scale easily were expensive. So after evaluating several options, we decided to write our own solution to handle video encoding.

At first we started using an automated binary build of FFmpeg from http://ffmpeg.arrozcru.org/ and just wrapped some C# code around it (thanks to a suggestion from my brother-in-law). This worked fine in the beginning, but as our format and encoding option needs changed, we could no longer rely on the generic automated builds. We needed to be able to build our own binary of FFmpeg with specific libraries included. Additionally, at the time of this writing, that site no longer hosts the automated builds.

This ended up being a far more complicated task than I imagined. There was some documentation in various places, but nothing that specifically walked through the process step-by-step for someone unfamiliar with the process. Once I was finally able to successfully compile and build FFmpeg, I decided I needed to document the entire procedure. This series of blog posts will attempt to document and identify the entire process I went through to accomplish this. Some steps may seem obvious, but I will include information for those that may have no knowledge of compiling and building software such as FFmpeg.

This process is written based on the following specifications:

Dell Latitude E6500 Intel C ore 2 Duo P8700 2.54 GHz 4.00 GB RAM
Windows 7 Professional 64-bit (using a 32-bit OS should be fine, but you will need to recompile the cross toolchain)
Oracle VM VirtualBox Version 3.2.10 r66523 (http://www.virtualbox.org/)
Ubuntu 10.10 64-bit (http://www.ubuntu.com/)

Stay tuned for the first set of details on setting up the virtual environment using VirtualBox and Ubuntu.

Cr-48 Chrome OS Beta Update 0.10.156.46

Thu, 03 Mar 2011 23:28:00 -0500

I was lucky enough to receive a Cr-48 just before Christmas (thanks Santa!). I don't do anything really crazy with it. You know, email, surfing, the occasional Google Docs editing, etc. Having the 3G has come in very handy, though. Anyway, I just got the Beta channel update for 0.10.156.46, which includes an update to Google Chrome 10.0.648.116. So far, it sounds like mostly bug fixes, none of which I have experienced. After the last update, I noticed the startup time and general usage seemed to slow down a bit. I'm hoping I'll see a little performance boost.

Here's the post regarding this update: http://googlechromereleases.blogspot.com/2011/03/chrome-os-beta-channel-update.html

Additionally, I recently turned on 2-step verification security for my Google account. So far so good. I do feel a bit better knowing there's an extra layer of security around my account. That's all fine and good, however, on my Cr-48, I was seeing a Sync Error and my Chrome bookmarks and plugins weren't syncing. Pressing the button that was supposed to let me enter a password did nothing.

After this update, I was able to correct the issue by opening Settings -> Personal Stuff and selecting to sign-in. After entering my password, I had to generate an application specific password, and I was good to go!

Google Voice 0.2.8 for Android 1.5 on HTC Droid Eris

Wed, 06 Jan 2010 17:49:00 -0500

After having a BlackBerry for about three years and loving it for work email, calendar, contacts, etc, I recently decided to make a switch and picked up the HTC DROID Eris for Verizon. I'm a big fan of most things Google, so I knew I would love this device. I won't go into the details of why or give my own review, but I did want to touch on one thing that really frustrated me.

I have a Google Voice account, which I also love. It no longer matters if I switch phones or don't get service. I have one number, and I can set it up to ring anywhere. Plain and simple. The downside was not having my Google Voice number show up on peoples' caller ids. Not a big deal, but irritating. Enter the Google Voice app (available on many platforms).

When I first installed the app (version 0.2.7) it worked great. I could make calls from my Eris and people would see my GV number instead of my Verizon number. After an updated 0.2.8 was released, suddenly, I was receiving, "network busy" or "network unavailable" messages and I couldn't make any calls.

Sure I could browse to the GV webpage and initiate calls from there, have GV call me on my device, then connect the call, but what a huge pain. After no responses from a Google rep on a thread in the GV help forum, someone figured out that the app was putting a + in front of the bridge number, which was causing the call to fail.

With no way to configure this and still no response, it seemed unlikely that I'd be able to use GV on my device any time soon. Someone else finally mentioned that there is a way to remove the extra + in front of the number: clear the dialer app data.

Menu -> Settings -> Applications -> Dialer -> Clear Data

Simple as that. Periodically, I find the app resorts to adding the + in front of the number, but clearing the dialer app data seems to fix it each time. While this is only a workaround, and hopefully Google will fix this in a near-future release, I am able to use my Google Voice number again, and it works great!

Adding Windows 7 64-bit PDF Indexing

Thu, 31 Dec 2009 13:04:00 -0500

I have been using Windows 7 64-bit since its official release in October and have loved everything about it so far. I was surprised at first when my PDFs weren't getting indexed, but was happy to know that it took very little work to get it to index them in just a few easy steps.

Verify that PDF files are selected for indexing
- Open the Indexing Options by clicking Start and typing, "indexing options" (without the quotes) and hitting enter
- Click the Advanced button
- Select the File Types tab
- Scroll down to .pdf, which should already be selected, but probably says, "Registered IFilter not found"
Download and install the Adobe PDF iFilter 9 for 64-bit platforms

That's it! You may need to rebuild the index (click click the Rebuild button in the Indexing Options), but afterwards, you should be able to search on PDF content. 32-bit users should be able to index PDFs out of the box.

Reading Embedded XMP Packets using ColdFusion

Fri, 20 Feb 2009 12:29:00 -0500

In working more and more on our Digital Asset Management initiatives, I've become more involved in working with XMP. Mostly as an experiment, I wanted to see if I could read and write this embedded metadata from and to a file. The Adobe XMP Toolkit indicates how this can be done based on the structure of specific types of files. It also indicates that packet scanning techniques can be used when file structure is unknown, however, is not encouraged.

I recognize it might not be the BEST way to read the XMP packet, however, all is fair in love and ColdFusion experimentation :) It's important to note that this code looks only for the FIRST instance of the string <x:xmpmeta and if there are multiple instances of XMP packets, subsequent instances are ignored. Adobe also indicates that if there are multiple packets, there is no way to know which packet is the correct one.

Needless to say, using the code below, I was able to read and parse XMP packets from all the Adobe sample files included in their toolkit. It may not be the end-all solution, and may not work for writing back to the file, but it's a start!

<cfparam name="URL.source" default="xmp-asset.jpg">
<cffile action="readbinary" file="#ExpandPath(URL.source)#" variable="data">

<cfset hex_data = BinaryEncode(data,"hex") />

<cfset xmp_string_begin = "3C783A786D706D657461" />

<cfset xmp_string_end = "3C2F783A786D706D6574613E" />

<cfset idx_start = FindNoCase(xmp_string_begin,hex_data) />

<cfset idx_end = FindNoCase(xmp_string_end,hex_data,idx_start) + Len(xmp_string_end) />

<cfset xmp_hex = Mid(hex_data,idx_start,Evaluate(idx_end-idx_start)) />

<cfset xmp_string = ToString(BinaryDecode(xmp_hex,"hex")) />

<cfset xmp_xml = XmlParse(xmp_string) />
<cfcontent type="text/xml">
<cfoutput>#xmp_string#</cfoutput>

Be sure to check out the Adobe XMP Developer Center for more information.

Wiimote Whiteboard Lab

Sat, 24 Jan 2009 10:37:00 -0500

Having been very intrigued by the experiments of Johnny Lee using the Nintendo Wii remote, I decided to give one of them a try. I decided the experiment that would be most useful to me was the interactive whiteboard.

Following several instructions by different people online to build an infrared LED pen, I went to Radio Shack, spent about $10, and had a (somewhat) working version after about 5 hours worth of work. Had I known what I was doing, it should have only taken me about an hour, if that.

The first problems I ran into were due to a lack of understanding of some fundamental principles of electricity. Initially, I was using a battery housing that holds 3 AAA batteries (thinking that much power was needed) end-to-end outputting 4.5 volts (1.5 volts each). Each IR LED I used to complete the circuit ended up overloaded and burned out. Finally, with some advice from my father with a background in electricity, I learned that keeping the positive and negative ends of the batteries separated (laying them side-by-side instead of end-to-end), I could increase the output without increasing the voltage. Thereby powering the IR LED without burning it out.

That hurdle overcome, I started discovering that sometimes the IR LED was getting power, and sometimes it wasn't. Had I done some more research online, I could have saved myself a couple hours. I finally learned that with LEDs, polarity matters. In other words, the positive and negative charges had to be correctly paired with the tips of the LED. So, the cathode tip of the LED must be matched up with the positive charge of the battery. After discovering this, I learned that one AAA battery provided sufficient power for the LED, which made construction of my circuit much easier.

Finally, the last major problem I had was reading. Yes, reading the packaging on wire I bought to connect the components in the circuit. Had I read the package, I would have seen that the wire I bought was covered in a protective coating. So in order for it to conduct the electricity of the circuit, I had to scrape the coating off the tips to make a good contact with the battery and LED and other components.

Back to one of my earlier statements; I had it (somewhat) working. Was my IR LED pen able to "transmit" data to my Wiimote? Yes. Was it consistent? Somewhat. Was it as easy to use as a real whiteboard? No. It was quite difficult for me to find a position for my Wiimote where it could "see" the IR beam or reflection. This led me to believe that either the IR LED I purchased wasn't providing sufficient light to be seen by the Wiimote or that something about the circuit I constructed wasn't optimal. Either way, not being knowledgeable enough about the Wiimote's capabilities, the electrical circuit I was attempting to construct, or the properties and capabilities of the IR LED with which I was working, I just wasn't getting the expected results.

I decided that for my experiment I was at least successful even if my construction was not optimal. In the end I decided to purchase an inexpensive IR LED pen online that used the same brand and model of IR LED as originally recommended by Johnny Lee (NOT available at my local Radio Shack). It worked like a charm! I also found that using the Wiimote with the IR LED pen worked the best with a projector instead of on a computer monitor.

So, if you consider yourself anything of a hobbyist with some basic electrical knowledge, you should give this a shot yourself. If you just want to see and use a virtual whiteboard with a Nintendo Wiimote and infrared LED pen, save yourself some time and frustration, and buy one online!

Ubuntu 8.10 and Microsoft Virtual PC 2007

Mon, 15 Dec 2008 13:06:00 -0500

After finally getting Cygwin installed on my Windows XP Pro SP 2 laptop, I discovered how difficult it can be to get large open-source projects to correctly compile under Cygwin. This was exteremely irritating for obvious reasons.

So I decided to try and install Ubuntu 8.10 under Microsoft Virtual PC 2007. I had previously installed older versions of Ubuntu this way, but was having some trouble getting 8.10 installed. After some searching, I found this wonderful post and comment thread.

The short story is, I had to add vga=791 noreplace-paravirt to the grub command line for the initial boot. Then noreplace-paravirt to /boot/grub/menu.lst after installation and any subsequent updates that modify this file.

Cygwin Setup Finally Working

Tue, 02 Dec 2008 12:19:00 -0500

I had been trying to get my environment set up to play with Adobe Alchemy, but kept running into problems with my Cygwin setup. After several frustrating hours of installs, re-installs, uninstalls and some out-of-control Perl threads, I think I've got it ready to go.

The problems I was running into were largely due to two things: 1) The path of my home directory containing spaces ("C:\Documents and Settings\...") and 2) my inexperience with Cygwin.

In regards to #1 above, I think this was causing my home directory (/home/username) to not get created properly. That meant additional difficulties in customing my environment by editing my .bashrc file. After the initial learning curve and interpreting the less than specific Adobe Getting Started on Windows instructions for Alchemy, I was able to get the sample library compiled and working correctly.

I did run into a few other problems, but I was able to get through them. I'll be doing a write up for getting started with Alchemy and Cygwin for Windows users with very little Unix experience when I get some more time. Of course, now that I've got Cygwin running, I'm hoping to do try out something more significant than Hello World :)

Adobe Alchemy: Initial Setup Impressions

Sun, 30 Nov 2008 16:02:00 -0500

My first thoughts on the Alchemy project were of doubt. What would be the benefit of running C/C++ code on the AVM2? It seemed more like a novelty, like running old school NES games on the Wii. The more I started to think about it (and get past my bad memories of Computer Science C/C++ classes in college), the more I started to realize just how powerful this could be. There are so many open source C/C++ libraries for a wide variety of applications that, when paired with Flash, could have an amazing impact on Flash and AIR.

I thought I would tinker with it a bit. Of course, coming to the realization that Alchemy is still limited by the Flash player sandbox was a little disheartening, but I suppose understandable. I found the initial download and setup quite cumbersome. One of my biggest roadblocks was working with Cygwin. I admit, that using Cygwin has its advantages, but configuration thus far has been a beast. Additionally, the getting started section for Windows section left something to be desired in terms of details. I'm the first to admit I'm no UNIX or Linux expert, but I'd like to think I know my way around a bit.

After a few hours of trying to just get my environment set up, something more important required my attention, so I have yet to get the provided examples working. I hope to get back to this more in the next couple of weeks. I've got a few open source C++ libraries I'd love to test out. I'm extremely intrigued, but until I get my environment set up, I'm still skeptical.

Branden Hall has written a good overview of Alchemy.

Mango Blog Skin: Lotus Flower

Sat, 22 Nov 2008 18:39:00 -0500

I've ported a few free css templates to Mango Blog skins, but this is the first one I'm posting for public consumption :) I used lotus flower from Free CSS Templates for my personal blog/website (picked by my wife, I have to add) and just swapped out the header image.

Download the skin and extract it to the "skins" directory in your Mango Blog installation. Then just select it in the Mango Blog administration. That's it!

Adobe Media Orchestrator: Video Asset Management?

Wed, 19 Nov 2008 10:02:00 -0500

One of my colleagues at Adobe MAX in San Francisco sent me a text message last night. He was at the sneaks session and indicated that there was a preview of an Adobe product called Media Orchestrator. This AIR application, in conjunction with existing products like Premiere and LiveCycle, will facilitate video production workflow, from production to review, feedback and rights application.

Considering that we have been spending most of our time lately at work trying to write a custom video asset management solution using Adobe products, this announcement is really bittersweet. It's fantastic that Adobe is (or considering) releasing this as a product, however, it most likely won't be available any time soon and we need it right now.

More information and pictures on Serge Jespers and Andrew Shorten's blogs where they were live blogging the sneaks event.

can it be?

Carbon Coder at the Core of Adobe Flash Media Encoding Server

Tue, 18 Nov 2008 11:59:00 -0500

Ever since September 10, I've been quite excited about the announcment that Adobe was finally releasing an encoding server solution called Flash Media Encoding Server. I was able to secure a pre-release version through work to take it for a test drive.

Initial results were great. Of course, being a pre-release, there were several challenges. The first being that the software required a USB drive for a USB key. In a time when virtualization is sweeping the corporate world in an effort to reduce costs, the requirement for something as simple as a physical USB port was very irritating. We overcame that with a hardware/software solution, not the most desirable (spending money just for trial software), but oh well.

The next challenage was the documentation, or rather, the lack thereof. Yes, the basic steps for installing the software were included, as well as the basic steps for setting up watch folders. One of the big features I was anxious to test, though, was the SDK. Surprise, surprise. No SDK documentation. We're hoping to put FMES (or some transcoding agent) in the front of our video asset management workflow to generate proxies of video assets for review within out applications. Most of our applications are homegrown and based on Adobe technologies (ColdFusion, Flash/Flex, AIR, etc). So the ability to leverage a transcoding solution via SDK is a vital element.

Aside from the lack of documenation with the pre-release trial, I did find that several of the source files I tried to run through FMES couldn't be encoded. Ok so they were odd codecs (WMV3, muxed mpeg, etc), but still. The error messages gave me little to go on besides the error message, "There is a video decoding error." Using some other transcoding tools, I was able to encode the problem files, however, so why couldn't FMES process them?

Finally, my biggest surprise is that in testing FMES, I discovered that at its core, it's really a product by Rhozet called Carbon Coder. After a little quick digging, I discovered this press release indicating Adobe's intentions. It seems to me that, basically, Adobe has taken Rhozet's product, swapped out the words "Rhozet Carbon Coder" with "Adobe Flash Media Encoding Server" and limited the outputs to only those that are compatible with Flash. The licensing model is slightly different than Rhozet's, however, it is more or less the same product.

At this point, I'm trying to decide why we would want to go with FMES over Carbon Coder. If we're going to spend several thousand dollars to a transcoding solution, why not spend a few more and not be limited to only Flash-enabled media outputs?

Now that FMES is available for purchase, Adobe will have more documentation or white papers available that can make a good argument, in addition to not choking on certain input formats. Granted, video transcoding experts can probably explain why some of these problems exist and how to solve or get around them, but for a product that seems as if it should be a turnkey solution from Adobe, I would have hoped my intermediate skills in working with video transcoding should be more than enough to work with FMES.

More to follow in the coming weeks, I'm sure...