The first step was to configure a media server that would be visible from our TiVo. Using pyTivo was the ideal choice, and this was incredibly easy to configure. There were a few key settings to optimize it for our Tivo Premiere. Using the output pixel width/height the max resolution was bumped up to take advantage of the HD resolution, and ffmpeg_pram was used to maximize a dual-core dual-CPU server configuration:


[Server]
port=9032

# Full path to ffmpeg with options
ffmpeg=/usr/bin/ffmpeg
ffmpeg_pram=-threads 4

# Output Pixel Width
width=1440
height=720

# Use 16:9 aspect ratio
aspect169=true

[Movies]
type=video
path=/export/video/Movies

Increasing the number of threads for transcoding allows the pyTivo server to make HD quality video instantly accessible from my TV. You will need to exercise caution, however, if your server has less hardware then you need to adjust accordingly.

In this configuration, getting videos to the Tivo is as simple as copying them to the /export/video/Movies directory. While this is possible to do directly, automating this process for videos and movies in my iTunes database is more ideal. For movies tagged with meta-data it would be even better to publish this so that this information is displayed on my TiVo.

My first step was to relocate my iTunes database to my pyTivo server. These files were placed into /export/video/iTunes and includes Music, Movies and TV Shows. For the purposes of this only the Movies are going to be considered for publishing to Tivo. As videos are added to my iTunes database, they are copied into a subdirectory under /export/video/iTunes/Movies so my script simply watches this directory for new content.

Typical execution is something like this:

perl AtomicSage.pl monitor /export/video/iTunes/Movies /export/video/Movies/


A few things are going to be needed for AtomicSage, both AtomicParsley and the Perl library AtomicParsley::Command need to be installed. Additionally, the script assumes a Linux environment to monitor the iTunes movie directory for changes. When a new movie appears in iTunes, AtomicSage copies it to /export/video/Movies and organizes the files by genre. A meta file is created for each video, inside this file things like movie title, description and genre are stored – this information is displayed on your Tivo screen.

You can use this with home movies and something like MetaX as well. Once your videos are built, run MetaX and you can edit the meta-data within the M4V container. Using AtomicSage, once videos are added to iTunes the meta data is extracted and displays on your TV as you browse media from your movie collection.

If you’re ready to go, download AtomicSage and get your movie database organized.

A few factors govern the total size of your DVD media. Obviously, if you have a lot of video footage you are going to immediately be in trouble – the typical rule of thumb is a video will minimally require approximately 1 GB for 1 hour of footage. Since a DVD is limited to 4.37 GB this pretty much restricts you to about 4 hours of video with some flexibility for a main menu or 4+ hours if you’re willing to sacrifice a little.

Audio tracks can have significant impact on the size of the media. Surround sound formats like AC-3 require additional streams to store data, most motion pictures include a number of audio tracks usually with a minimum of a stereo AAC and a surround sound AC-3 data stream. For a small video with both of these audio tracks, there might be 500 MB of audio and 900 MB of video information for approximately 1 hour. For the audio data, my example has 350 MB of AC3 sound and 150 MB of AAC stereo audio – if you’re looking to conserve space, you can imagine which one you will want.

The other factor is arriving at a way to shrink your video. On a DVD disc the standard is pretty well established, it must be MPEG-2 and the screen resolution is fixed at 720×480 so you can’t shrink your videos by decreasing the resolution. On the other hand, you can alter the video bitrate (kbps) to control the amount of video data being rendered. I highly recommend checking out the DVD-HQ: Bitrate & GOP Calculator to help you determine the optimal settings for this. Once you arrive at a number for your video bitrate, use SmallDVD to convert into a disc image.

A few other commercial solutions exist for creating DVD discs on a Mac, but most don’t offer the flexibility to adjust the video bitrate. If you are trying to squeeze the most out of your discs and aren’t looking for movie studio quality results then this is the best way to go.

Unfortunately, Oracle acquired Sun and has since discontinued the project and will return the OpenOffice framework to the developer community. There are a few new projects that have cropped up to address the gaps Oracle OpenOffice may be introducing, including the popular LibreOffice, but in the near term it’s not really clear what the prospects for OpenOffice.org are going to be.

So, despite my better judgment, after more than 10 years with OpenOffice I’m planning to standardize on Microsoft Office for a while. It has only been in the last few years that I have moved from a Linux desktop to Mac, which makes this transition even easier. However, I have a significant number of documents in OpenOffice format that need to get converted into MS Word, Excel and PowerPoint. Fortunately, I can use OpenOffice to do this for me.

My first step is to create some macros for automating the conversion, you can find this under Tools – Macros – Organize Macros – OpenOffice.org Basic.

Here you can see the macros that I have added under Module1, you may choose to create a new module or simple amend as I have done here. For each core document type, I have added a macro that is able to convert that file.

' Save document as a Microsoft Word file.
Sub SaveAsDoc( cFile )
   cURL = ConvertToURL( cFile )
   oDoc = StarDesktop.loadComponentFromURL( cURL, "_blank", 0, (_
            Array(MakePropertyValue( "Hidden", True ),))

   cFile = Left( cFile, Len( cFile ) - 4 ) + ".doc"
   cURL = ConvertToURL( cFile )

   oDoc.storeToURL( cURL, Array(_
            MakePropertyValue( "FilterName", "MS WinWord 6.0" ),)
   oDoc.close( True )

End Sub

' Save document as a Microsoft Excel file.
Sub SaveAsXls( cFile )
   cURL = ConvertToURL( cFile )
   oDoc = StarDesktop.loadComponentFromURL( cURL, "_blank", 0, (_
            Array(MakePropertyValue( "Hidden", True ),))

   cFile = Left( cFile, Len( cFile ) - 4 ) + ".xls"
   cURL = ConvertToURL( cFile )

   oDoc.storeToURL( cURL, Array(_
            MakePropertyValue( "FilterName", "MS Excel 97" ),)
   oDoc.close( True )

End Sub

' Save document as a Microsoft PowerPoint file.
Sub SaveAsPpt( cFile )
   cURL = ConvertToURL( cFile )
   oDoc = StarDesktop.loadComponentFromURL( cURL, "_blank", 0, (_
            Array(MakePropertyValue( "Hidden", True ),))

   cFile = Left( cFile, Len( cFile ) - 4 ) + ".ppt"
   cURL = ConvertToURL( cFile )

   oDoc.storeToURL( cURL, Array(_
            MakePropertyValue( "FilterName", "MS PowerPoint 97" ),)
   oDoc.close( True )

End Sub

Function MakePropertyValue( Optional cName As String, Optional uValue ) _
   As com.sun.star.beans.PropertyValue
   Dim oPropertyValue As New com.sun.star.beans.PropertyValue
   If Not IsMissing( cName ) Then
      oPropertyValue.Name = cName
   EndIf
   If Not IsMissing( uValue ) Then
      oPropertyValue.Value = uValue
   EndIf
   MakePropertyValue() = oPropertyValue
End Function

The next trick is to automate this, and I did that with a quick shell script. It’s really nothing fancy, and could easily be improved on – the idea is to find all the corresponding document types (ODT / ODS / ODP) and then execute the right macro from OpenOffice to convert from the OOO format into an MS Office file:

#!/bin/sh
export PATH=$PATH:/Applications/OpenOffice.org.app/Contents/MacOS

find ~/Documents -name \*.odt -print -exec soffice -invisible "macro:///Standard.Module1.SaveAsDoc({})" \;
find ~/Documents -name \*.ods -print -exec soffice -invisible "macro:///Standard.Module1.SaveAsXls({})" \;
find ~/Documents -name \*.odp -print -exec soffice -invisible "macro:///Standard.Module1.SaveAsPpt({})" \;

That’s it. With a final backup from TimeMachine, I then went ahead and removed the remaining OpenOffice documents from my Document directory.

As an Ubuntu user, I followed these directions with the exception of the step to read my iPhone device identifier. Since my iPhone directory had a sufficiently difficult name, it was easier to browse to that from a command prompt (~/.gvfs/[iPhone Name]) and issue the following:

sudo lsusb -v | grep 'iSerial' | awk 'length($0)>=68'
ipod-read-sysinfo-extended [id] ~/.gvfs/[iPhone Name]

Following a reboot, after plugging in the iPhone to the USB port the icons for both the camera and music device appeared on my desktop. Launching Rhythmbox displayed my music library with a device representing my iPhone. It was then possible to drag and drop music to the device, as well as remove existing songs and offload some audio clips.

Unlike iTunes, when working with media on my iPhone everything appears as a filesystem. While it’s not always possible to directly modify files, you can read anything off the device and Nautilus includes convenience buttons to launch native applications for writing music and photos.

When F-Spot opens for the photos you can view any of the pictures that have been taken on your device. There is native support for managing the iPod photo library as well, but you will need to run gtkpod and performance can be a little sluggish if you have a lot of pictures.

Keep in mind this project is still in development, so you may find this can supplement your iTunes synchronization process but is not quite ready yet as a full replacement. Sync operations can be slow, even after 2 hours with a Sync in Progress on my iPhone, it still needed another 15 minutes before the music library became available for playback on the device. There was very little feedback during this period, the standard sync screen disappears and the iPod application on the iPhone shows only that it is Syncing without any media visible in the library.

When everything completes, the selected music along with album art are available on the iPhone using the iPod program. There are some unsual nuances to the application; for example, ringtones appear in Rhythmbox alongside regular music. Be careful not to delete all of your music on the iPhone as this will also delete the other media on your device (videos, movies and ringtones).

Since we share a common music directory on my network, it is nice to have the library up-to-date automatically in Rhythmbox. On a Windows computer with iTunes, it is necessary to run a lengthy batch process to sync the music files with the filesystem. Even with the extra time for synchronization on Ubuntu, it still takes about the same amount of time to complete a transfer of around 100 songs.

The progress on native iPod Touch and iPhone synchronization now makes this a viable alternative for getting music onto your Apple device. It still would be nice if Apple could provide a native iTunes for Linux, there are a number of core hardware functions (eg: operating system upgrades) that still require iTunes on Windows or Mac. Overall, the current state of iPod support for the Touch and the iPhone is exemplary and can already provide Linux users with most of the function needed to synchronize their device with a music library, photos, videos and ringtones.

The answer might be as simple as marketing. In fact, the term Open Source (OSI) was intended to be a marketing slogan that replaced the loosely termed “free software” movement.

Unfortunately, OSI does not have the deep pockets of a large commercial establishment. As Steve Yegge pointed out in a recent podcast, people don’t really care how good your product is – it’s only about the brand association they have made with it. This is why Kleenex is so popular or that New Coke failed despite the fact that it actually tasted better. Once consumers have accepted your brand it’s very difficult to break that connection.

Why is this a problem for Open Source?

Consumers have already made a connection with commercial software, they know who Microsoft is and accept the products offered by that brand. Apple is experiencing a resurgence primarily as a result of brand, consumers have accepted Mac as a high-end computer that “just works”.

Normally, brands can try to re-establish themselves by choosing new names. To some extent, the free software movement already attempted to do this by calling themselves Open Source. The problem is, open source has no visibility in the consumer marketplace because it is free by design. You won’t see any glossy ads in Wired for LInux or flashy TV promo spots for the latest version of Firefox. In many ways, it’s amazing that open source has made any inroads at all.

Just because something is good doesn’t mean that consumers will use it. You might have the best software in the world, but if it isn’t promoted properly your market will never materialize. Of course when you product is free then consumers have no reason to buy it, in which case the market for your product is somewhat difficult to define.

To some extent the Internet has been an equalizer in this regard, if you visit the Firefox homepage you will find that the product looks attractive and the site may even convince you to install the program. Unfortunately, things are still somewhat muddy for potential users of open source solutions since it’s unclear how you get the product when there is nothing to buy.

Maybe the only way open source can take a foothold is if users can be convinced that software is suppose to be free? Unfortunately, for this to happen it will take significant financial investment and an extensive advertising campaign – neither of which is something that free software can afford to do. So ultimately it may be that the only way free software can reach people is by making it something you pay for rather than give away.