Building Raspbian images for Raspberry Pi

I recently bought a Raspberry Pi, which is a credit card sized computer with an ARM processor. I'm using it as my TV frontend, running Raspbian and XBMC. I'm building my own packages for XBMC since it requires the latest development version.

I initially installed my Pi with the foundation image, but found that it included a lot of packages which I didn't need. Since I have a slight obsession about doing things as efficiently as possible, I decided to build my own image with XBMC from scratch.

I implemented a script in Bash, mkraspbianxbmc.sh which does this. It uses debootstrap to install a minimal Raspbian system in a chroot. It then installs XBMC and a couple extra packages, and does some necessary configuration. Finally it creates an image file with the necessary partitions, creates the filesystems, and copies the installation into the image file. The resultant image fits onto a 1GiB SD card. You can download a pre-built image from this page.

The script can be modified to build images with different packages, or even a very minimal image which fits onto a 512MiB SD card.

Working with GPX files in Python

Extracting specific track segments from a track

I have an i-Blue 747A+ GPS logger which I use to track my runs (amongst other things). Afterwards I use BT747 to retrieve the data from the device and create a GPX file of the run, which I then upload to Endomondo which gives me nice graphs and statistics.

I need to modify the GPX file slightly before I can do so however: I use the button on the device to mark the beginning and end of the run, which appear as waypoints in the GPX file. BT747 creates separate track segments (within a single track) between each waypoint, but Endomondo ignores these. I therefore need to extract the single track segment covering the actual run and create a new GPX file with just that segment. I therefore wrote a script in Python, splittrack.py, to do this. It uses the gpxdata library to parse the input file, locates any track segment which match a set of time, distance, displacement and speed criteria1, and outputs a new GPX file with just those.

% splittrack.py mgorven-20121015_0109.gpx > run-20121014.gpx
Reading mgorven-20121015_0109.gpx
<TrackSegment (23 points)> covers 21m over 0:00:22 at 1.0m/s average speed with 4m displacement
<TrackSegment (904 points)> covers 3018m over 0:15:03 at 3.3m/s average speed with 8m displacement
Adding <TrackSegment (904 points)>
<TrackSegment (4 points)> covers 3m over 0:00:03 at 1.3m/s average speed with 3m displacement

Integrating heart rate data

I then recently bought an Oregon Scientific WM100 heart rate logger. It listens to the broadcasts from a heart rate strap2 and records the measurements every 2 seconds. I retrieve the data using the wm100 driver for Linux which writes a CSV file like this:


In order to get this data into Endomondo, I needed to combine the GPS trace with the HRM data into a single file format which Endomondo accepts. I initially started implementing a library for the TCX format3, but then discovered that there is a GPX extension for including heart rata data which Endomondo accepts. So I wrote a script in Python, wm100gpx.py, which reads the input GPX and CSV files, merges the heart rate measurements into the GPX records, and outputs a new GPX file.

% wm100gpx.py 2012-10-14T18:08:27.csv < mgorven-20121015_0109.gpx > run-20121014.gpx

The entries look like this:

<trkpt lat="37.392051" lon="-122.090240">

  1. I actually initially wrote this to find tracklogs of runs amongst all my tracklogs. 

  2. I use a strap from an entry level Nike triax C3 heart rate monitor watch. 

  3. Which is quite exhaustive... 

XBMC packages for Raspberry Pi running Raspbian

I recently got a Raspberry Pi, which is an ARM based device which runs Linux. My goal is to use this as my HTPC running XBMC, so that I can move the fileserver out the lounge.

Edit: I've moved the latest information and updates about these packages to this page.


I installed Raspbian on my RPi, which is basically a rebuild of Debian Wheezy specifically for the RPi (targeting the ARMv6 architecture with hard float). I found various instructions on how to build XBMC for Raspbian, but none of them were in the form of deb packages, and installing software without packages just makes me queezy. So I went off and built it myself.

Since the RPi is relatively low powered, I built the package on my laptop using qemu-user, which emulates binaries with a different architecture. I based the packaging on the XBMC package in Wheezy, and the source is from the xbmc-rbp branch on GitHub. I made the modifications to the source as per this forum post and added an initscript so that it can automatically start at bootup.


The easiest way to install the package is to add my archive to your system. To do this, store the following in /etc/apt/sources.list.d/mene.list:

deb http://archive.mene.za.net/raspbian wheezy contrib

and then import the archive signing key:

sudo apt-key adv --keyserver keyserver.ubuntu.com --recv-key 5243CDED

You can then install it as you would with any other package, for example, with apt-get:

sudo apt-get install xbmc

(If you don't want to configure my archive you can download the packages manually, but you'll have to deal with all the dependencies. Note that it requires a newer libcec package which is also in my archive.)

The user which you're going to run XBMC as needs to be a member of the following groups:

audio video input dialout plugdev


To run XBMC, run xbmc-standalone from a VT (i.e. not under X). XBMC accesses the display directly and not via Xorg.

If you want XBMC to automatically start when the system boots, edit /etc/default/xbmc and change ENABLED to 1:


You also need to set the user which XBMC should run as (the xbmc user is not automatically created at the moment). Run sudo service xbmc start to test this.


The following settings in advancedsettings.xml decreases the CPU usage while showing the UI. Disabling the RSS feeds also helps with this.



If you want to rebuild this package with a different source (e.g. a later Git revision), you need to prepare the source by running ./bootstrap before creating the orig.tar.gz. You obviously need all the build dependencies (which should be listed in the packaging), as well as the VideoCoreIV libraries in /opt. You probably want to set DEB_BUILD_OPTIONS=nocheck parallel=4 to disable the tests (they failed to build for me), and speed up the build by using more than one core. You can find the packaging in my archive or my Bazaar repository.

Migrating SMSes from Nokia to Android

My wife recently got a Samsung Exhibit II 4G Android phone to replace her aging Nokia E63. Migrating her contacts was accomplished fairly easily by exporting them to CSV with Nokia PC Suite and then importing them into Google Contacts. Migrating SMSes was not so trivial however.

Other approaches

There are a couple methods floating around the web, but none were suitable. This one uses Gammu to retrieve the SMSes from the Nokia, and then a script to convert them to an XML format readable by SMS Backup & Restore. It turns out that Gammu doesn't work on Symbian S60v3 devices however. This script can convert SMSes exported by the MsgExport app to XML for SMS Backup & Restore, but I didn't feel like paying for it or dealing with the Ovi Store. VeryAndroid is a Windows application which can convert SMSes from Nokia PC Suite CSV format and sync them directly to an Android device, and Nokia2AndroidSMS can convert SMSes from the OVI Suite database to XML for SMS Backup & Restore. I didn't want to deal with more Windows software though, so I just decided to write my own.


I already had the Nokia PC Suite installed and was using it to migrate contacts, so I decided to work with the CSV output it generates for messages. A received SMS looks like this:

sms,deliver,"+16501234567","","","2012.06.13 19:13","","Leaving now"

and a sent SMS looks like this:

sms,submit,"","+16501234567","","2012.06.13 19:11","","Where are you?"

The fields are:

  • 0: "sms" for SMSes (MMS is presumably different)
  • 1: "deliver" for received messages, "submit" for sent messages
  • 2: Sender's phone number (blank for sent messages)
  • 3: Recipient's phone number (blank for received messages)
  • 5: Date and time in "YYYY.MM.DD HH:MM" format and local timezone
  • 7: Message body

Fields 4 and 6 are always empty for SMSes (they are probably used for MMSes, one being the message subject).

I also decided to generate XML for the SMS Backup & Restore app. The XML format looks like this:

<?xml version='1.0' encoding='UTF-8' standalone='yes' ?>
<?xml-stylesheet type="text/xsl" href="sms.xsl"?>
<smses count="1">
    body="Leaving now"
    readable_date="Jun 29, 2012 8:03:04 PM"

but can be reduced down to this:

<?xml version="1.0" encoding="UTF-8"?>
<smses count="1">
        body="Leaving now"

The attributes of the <sms> element are:

  • protocol: Always "0" (possibly different for MMS)
  • address: Sender or recipient phone number
  • date: Date and time in milliseconds since 1 January 1970
  • type: "1" for received message, "2" for sent messages
  • body: Message body
  • read: "1" if the message has been read
  • status: Always "-1"

The script

I implemented a script called [nokia2android.py] in [Python] to convert one or more CSV files to this XML format.

./nokia2android.py received.csv sent.csv > android.xml

The XML file can then be transferred to the Android device (using USB or Bluetooth) and stored in /sdcard/SMSBackupRestore. It will then be presented as an option after selecting the Restore button in the app.

Clicking on Apple Macbook Pro trackpads under Ubuntu Precise

Apple trackpads don't have separate buttons, the entire trackpad is itself a clickable button. The default OS X behaviour is to treat clicking the pad with two fingers as the right button, and clicking the pad with three fingers as the middle button. Enabling tap to click (i.e. touching the trackpad but not actually clicking it) tends to result in false positives since the trackpad is so big. I therefore setup this behaviour under Ubuntu Oneiric.

When I upgraded to Ubuntu Precise two finger clicks started registering as the left button. (Three finger clicks still worked.) It turns out that this is due to the new clickpad support in Precise. The solution is to disable the ClickPad attribute. My Synaptics configuration now looks like this:

TapButton1              = 0
TapButton2              = 0
TapButton3              = 0
ClickFinger1            = 1
ClickFinger2            = 3
ClickFinger3            = 2
ClickPad                = 0

Encoding video from a Canon Vixia camera

I recorded some presentations at work recently using a Canon Vixia HF R21 video camera, and needed to encode the videos to lower resolution. It took me longer to figure this out than I expected, so I thought that I'd document the details.


The camera exposes an MTP interface when plugged in using USB, and the video files are stored under /AVCHD/BDMV/STREAM/. The M2TS container format is used, but the camera splits a single stream into multiple files such that each file is no bigger than 2GiB. These files need to be concatenated together before playing or re-encoding. Video is encoded using H.264 at an anamorphic resolution of 1440x1080 (i.e. it must actually be displayed at 1920x1080). The frames are interlaced at 50.96fps1. The bitrate can be set at 5, 7, 12, 17 or 24mbps. Audio is encoded using AAC with 2 channels at 48kHz, with a bitrate of 256kbps.


The best codecs in terms of storage space efficient seem to be H.264 and AAC, so that's what I chose to encode the output in. I wanted a good quality version at 720p, and a lower quality version at 480p. Since these were presentations, single channel audio is sufficient. I tried using both libav (previously ffmpeg) and mencoder and had trouble with both, but eventually got libav working.

Good quality

The command I used for good quality was:

avconv -i "$INFILE" \
    -acodec libfaac -ab 128k -ac 1 -coder ac \
    -vcodec libx264 -filter yadif -r 29.97 -s 1280x720 -aspect 16:9 \

Audio is reduced to mono and encoded in AAC at 128kbps. Video is resized to 1280x720, deinterlaced, and encoded in H.264 with the default quality settings. To adjust the quality one can use the --crf parameter (which defaults to a value of 23). At 30 the quality loss was easily noticeable.

On my Intel Core2 Duo P8600 this encoded at 10fps with a video bitrate of 700kbps and audio bitrate of 83kbps. This puts an hour of video at 339MiB. Adding -threads 2 allows both cores to be used, at the expense of slightly worse motion compensation. If this is too slow, one can select a faster x264 preset using the -pre parameter.

Low quality

The command I used for low quality was:

avconv -i "$INFILE" \
    -acodec libfaac -ab 128k -ac 1 -coder ac \
    -vcodec libx264 -filter yadif,crop=848:480 -r 29.97 -s 854x480 -aspect 53:30 \

Audio is encoded the same as above. Video is resized to 848x480, deinterlaced and encoded in H.264. Apparently codecs are more efficient when the resolutions are a multiple of 16 which is why the video is resized to 854 and then reduced to 848.

This encoded at 20fps with a video bitrate of 300kbps and audio bitrate of 83kbps. This puts an hour of video at 165MiB.

  1. The camera has options to use 30p and 24p in the menu, but this doesn't seem to have any effect.