This would have been pretty easy, but I decided to put it in an old retro 80s ghetto blaster case, namely this one, a Phillips Roller radio cassette player. My parents had one when I was a kid and I came across one on eBay and thought it would be perfect.

So, I ripped out the guts, including the rather measly speakers and cut out loads of the inner plastic that was supporting the amp, radio and cassette parts and went to work to put in my own contents.

This took quite a while and the added difficulty of a young baby and a two year old who needs constant entertainment only added to the challenge (using a dremel when the wife has put your 6 month old son to sleep doesn’t go down too well ).

Once all the insides had been cleaned out, I then went to work to put my own contents inside. Alongside the Raspberry Pi, I also put in:

USB WiFi dongle with external aerial (so I can replace the one you see from the outside)

Amp from Sure Electronics (I wanted a tripath, but this seemed the easiest)

Car Speakers (3-way speakers for better range, but at 5 1/4″ these aren’t gonna rock the house too much)

Simple USB sound card (I’d read the internal sound chip in the Pi is a bit basic if you are using the 3.5mm output, hence this)

The car speakers were chosen since the old ones were really small and not capable of any bass or offering high power output, the amp because i thought it had a volume control – but to get it to work needed some ninja soldering skills that I don’t have, the sound card I believe is an improvement on the in-built chip (but I might be wrong) and finally the WiFi dongle uses an external antennae, so I’ll still get a good service with all the different devices crammed inside the stereo.

The next step was getting the Pi to play music. Luckily, I tend to listen to all my music from a central server running the rather awesome MPD to stream my music out to as many connected MPD clients as I want. So, all I had to do was install MPD on to a fresh version of the latest Raspbian and I was almost ready to go. I wrote a small script which, on boot, checks to see if my MPD server is available, if so it starts the MPD daemon on the server playing the stream, then starts the MPD client on the Pi playing the stream from the server. If it doesn’t find the server, it adds music from the small local library stored on the SD card and plays that instead., via MPD.

Here’s the inside of the Roller, with the Pi, amp, soundcard and speakers in place:

A view inside the Raspberry Roller.

I stuffed the whole thing with cotton wool, to help boost the bass (but I know next to nothing about proper speaker design, it is one of those things I plan to do in my retirement – build massive bass-thundering speakers).

The Pi is at the bottom, the amp in the middle (you can see the amp’s fan sticking out) and the sound card is at the top.

The finished Roller in all it’s red raspberry goodness.

And here is the finished article (playing at the time, but you can’t tell when it is on apart from the music playing).

It sounds pretty good considering. The bass could be better, but you can’t expect much from such relatively small speakers. However, it does go pretty loud for something so small and it sounds really clean and crisp, so I am happy with it.

For the other Pi I have lying around, I plan to use it (when Piface comes out) to control the lights in the dolls house we got my daughter this Christmas. Eventually she will hopefully get in to using Scratch herself to control the lights in her dolls house (and maybe some other things like a little mini stereo or even a small screen in her dolls house). I would imagine when she sees that we can plug her dolls house into the TV and it is a computer, she will be a little excited and intrigued!

Many people have said I should try to cash in on this, but as others have pointed out, what I have done here isn’t rocket science, it isn’t complex, but I guess what is different is that I took the time to write about it and I plan to explain how to do it in non-techy speak. I also hope to write up a more detailed version, which will take time and I might charge a tiny amount for it, but that would simply be due to the time taken to write it up and because it would contains some code and scripts that I have created myself.

Kudos to WordPress for handling the server load. I use to host this site on my own sever at home, which would have melted by now!

I guess normality will resume tomorrow, but it was fun whilst it lasted. 

Wanting to simplify the approach and make it easier to set up, I came up with a slightly more refined solution that works much better. I’ve been using it successfully for about a year and it works with multiple cameras and screens. Now I have a second bundle of terror it has proved very useful.

Rather than trying to reinvent the wheel I decided to use products already out there, but ones that aren’t actually designed to monitor babies with and thus don’t have the crazy high price tags. I use an IP webcam, that connects to a WiFi network, this can then be viewed on any laptop or android device you have. It should also work with the iPhone as well, but not being a fan of Apple’s walled garden, I haven’t yet investigated this in any great detail.

The most important part of all this is the camera and I’ve stuck with the same camera I used last time, the IP Night Vision webcam. These can be found all over eBay, or there are also variants available on places like ebuyer.com and dabs.com. It has a WiFi and an ethernet network connection, night vision up to a couple of meters, two-way audio (so you can both listen to and talk to your baby – though I don’t suggest the latter, that is one big brother step too far) and it can be panned and tilted around so you can ensure the baby stays in view even if it moves. Most of these cameras are made by Chinese manufacturers, but some are of better quality than others. Foscam is a really good brand it seems, I have two and they have been running non-stop for the last year with no problems what so ever. The model I have used and would recommend is the Foscam FI8918W.

Foscam FI8918w

It is relatively simple to set up, you just need to connect it to your wireless network following the instructions included in the box. Once this is done you can then view the camera in many different ways. There is an internal web address that you can use to view it from any laptop connected to your wireless network using a web browser. You can also use the free media software VLC to view the video feed, the advantage being this approach also gives you sound and will work on pretty much all platforms supported by VLC (so Windows, Linux and Mac OS). I have written a number of scripts that can be run on Linux (Ubuntu) and Windows that launches the necessary video feeds and restarts the video every hour, since the audio can sometimes get out of sync with the video.

However, the best way to view the video feed I have found is via an Android app, called IP Cam Viewer. There is a free version which works perfectly for my needs, but if you want to say thanks to the developer, you can get the Pro version which is only a few pounds and gives you a few more features. I run this app on my Android tablet, an Android set top box and my Android phone. It allows me to view the the video feed of the two sleeping terrors on a TV in the kitchen using the Android set top box, anywhere in the house on my Android tablet or whilst out and about pretty much anywhere in the world (with a mobile data connection) on my Android phone. The app allows you to move the camera around if you need to and you can put widgets on your home screen that update as fast as once every second, so you don’t even need to run the app to see what is happening.

IP Cam Viewer in action

There is some technical cleverness that needs to be carried out to allow you to view the web cam outside of your wireless network on your phone, but this is very simple to set up and can be done for free.

All in all, this set-up works really well. So much so that the BBC just interviewed me about it! It is certainly a much cheaper and more flexible approach than the dedicated solutions on the market currently and rather than only being able to view the video feed on one tiny screen, you can view it on pretty much any screen you wish pretty much anywhere you want.

If you want to do this yourself, do get in contact via the Contact page. I might put together some more detailed information into a PDF, maybe with a “Pro” version that contains simple step-by-step instructions for a small fee or something. It seems all the rage these days

UPDATE: Gosh, so the story has got a bit of traction on BBC News and I’ve got quite a few comments and emails to read through. I’ll read and respond to all your comments over the weekend and I am sure I can produce a PDF with instructions.

Glad other people are keen on this idea!

This solution uses Scripting Layer 4 Android (SL4A) and Tasker alongside a custom ROM with rsync (any ROM should do so long as it has rsync). For those that don’t know, rsync is an awesome application that allows for remote backup and sync across machines. It turns out you don’t even need a ROM with rsync built in, as you can install an app that provides rsync, the app is called rsync backup for android and can be found here: https://market.android.com/details?id=eu.kowalczuk.rsync4android&hl=en

The only issue is you can’t call rsync from the command line using simply “rsync” since it isn’t in your systems path. However, if you use the following string instead,replacing calls to “rsync” with the following, the scripts still work: /data/data/eu.kowalczuk.rsync4android/files/rsync

SL4A is used to set out what to do via a script. You can write scripts in various languages in SL4A but I am using Bash as I am familiar with it. Rsync is used to actually handle the sync / backup and Tasker is used to launch the scripts when certain conditions are met.

I have created two scripts in SL4A, one backups my photos folder to my main photo folder on my server. The server runs the rsync daemon which rsync on the phone connects to. The other script does the reverse and copies a remote folder in my server that contains a bunch of music to my phone.

Tasker is set up with a profile that activates when my phone is plugged in and it’s between midnight and 7.00 am. This then connects to my WiFi network and then runs the two scripts via the SL4A plugin. Since I charge my phone each night this is effectively automatic.

The key here is getting permissions correct with rsync during the file transfer, as the memory card uses fat32 it hasn’t got any permissions. The rsync daemon doesn’t like this and errors out, hence the need for various settings. The second key here is exporting your password as an environmental variable. This is inherently insecure but since my server has multiple redundant backups and is only locally accessible I don’t care much. I could use trusted keys but I’m too lazy.

Here are the two scripts. First the music script that syncs from server to phone:

#rsync sync
export RSYNC_PASSWORD=password
DATE=$(date)
LOG=/mnt/sdcard/rsyncmusic.txt
echo rsync started $DATE > $LOG
TRY=1
rsync_com ()
{
DATE=$(date)
if [ $TRY = 15 ]; then
echo rsync failed, quitting on $DATE >> $LOG
exit
fi
sleep 10
echo rsync attempt $TRY started $DATE >> $LOG
rsync --progress -vHrltD --chmod=Du+rwx,go-rwx,Fu+rw,go-rw --no-perms --stats --password-file=/mnt/sdcard/scrt prupert@prupert::amusic /mnt/sdcard/amusic >> $LOG 2>&1
EXIT=$?
TRY=`expr $TRY + 1`
echo exit code is $EXIT >> $LOG
echo "********************" >> $LOG
}
rsync_com
while [ $EXIT != 0 ]; do
rsync_com
done
echo rsync finished $DATE >> $LOG
exit

The second script syncs the phones photos folder to my server:

#rsync sync photo
export RSYNC_PASSWORD=password
DATE=$(date)
LOG=/mnt/sdcard/rsyncphoto.txt
echo rsyncphoto started $DATE > $LOG
TRY=1
rsync_com ()
{
DATE=$(date)
if [ $TRY = 15 ]; then
echo rsync failed, quitting on $DATE >> $LOG
exit
fi
sleep 10
echo rsync attempt $TRY started $DATE >> $LOG
rsync -vHrltD --chmod=Du+rwx,go-rwx,Fu+rw,go-rw --no-perms --stats --password-file /mnt/sdcard/scrt /mnt/sdcard/DCIM prupert@prupert::apics >> $LOG 2>&1
EXIT=$?
TRY=`expr $TRY + 1`
echo exit code is $EXIT >> $LOG
echo "********************" >> $LOG
}
rsync_com
while [ $EXIT != 0 ]; do
rsync_com
done
echo rsync finished $DATE >> $LOG
exit

I have put some logging in to check progress and also some retry code that retries the sync if it timesout. It seems my HTC Hero’s WiFi connection claps out after a while so the script retries up to 15 times to run successfully based on the rsync exit code.

Since there is no startup script for the get_iplayer web pvr program, I created one. You can find it here:

https://github.com/pruperting/code_snippets/blob/master/get_iplayer

So, clearly, I have let this blog fall by the wayside. I haven’t written a new post since April 2011. Bad bad me.

As it’s time to renew my hosting, I have decided not to renew and to let this blog expire, rather than just let it attrite, as so many sites do. I use to have a lot more time, but now as I have a reasonably secure job (that I really enjoy), a wife, a lovely 2 year old and another on the way, I have no time to wax lyrical on tech any more.

You will be able to find me here:

http://www.rupert-plumridge.co.uk/

and my tech writings / scripts will move over to here: http://github.com/pruperting

So, it leaves to me say thanks to everyone who read and commented, it’s been fun.

After getting so annoyed by trying to get it work, I gave up. However, having a need to encode some stuff with FFmpeg again, I decided to return to the script. In the interveening period, my post on the Ubuntu forums, apart from getting one of the best response ever (see the matrix comment) had also encouraged others to give their solutions.

I thus used a combination of these ideas to create my updated version, which is below. It seems to be slightly more robust and still give the same useful info. I will be using this version for now, so I  hope it is of use for others

YAFF by Rupert Plumridge is licensed under a Creative Commons Attribution-Non-Commercial-Share Alike 2.0 UK: England & Wales License

If you simply install Ubuntu from the mini.iso containing the basic Ubuntu Maverick install, then on reboot, download my script using:

wget http://www.prupert.co.uk/scripts/xbmcmaverickrevo.sh

make it executable, using:

chmod a+x xbmcmaverickrevo.sh

then finally run it using:

sudo ./xbmcmaverickrevo.sh

I find it is best to run it again, once the script reboots the machine, as this fixes a few issues.

It is still a work in progress, as you can’t yet shutdown / restart etc via XBMC since it seems the method to add permissions in Ubuntu 10.10 have changed and they conflict with the setup here, so I am still looking into that. To shutdown / reboot at the moment, I log in via SSH and issue:

sudo shutdown -hP now

Also, it is designed to work just with the Acer Aspire Revo 3610, but there is no reason why it wont work with any other device (apart from maybe the audio configurations).

Enjoy.

Download It Here

Despite the most recent version (0.11.0) of lcd4linux being released in November 2009, Ubuntu Lucid comes with an older version and thus misses out on some new features and drivers, in particular drivers for the cool and cheap Pertelian X2040.

It is however, possible to build the latest version of lcd4linux from source, but you need to do a few things in order to get it to work.

This little guide assumes you have built stuff before on Ubuntu, so have all the necessary build tools.

First, navigate to the folder you want to download and build the source in.

Then, download the latest version via svn:

svn co https://ssl.bulix.org/svn/lcd4linux/trunk lcd4linux

Next, navigate into the newly created source folder using cd lcd4linux and install some dependencies:

sudo apt-get install automake autoconf m4 perl libtool gettext

Then, run ./configure, if you want to include a specific driver, for instance for Pertelian, include it here, like this:
./configure --with-drivers=Pertelian
Now, if you try to make, it wont work, I am not sure why, but there are some issues. To fix these, issue the two following commands:
mkdir m4
sudo ln -sf /usr/share/libtool/config/ltmain.sh .
Now, you can run:
make
sudo make install
And you now have a working and up-to-date lcd4linux. I’d recommend using checkinstall rather than make install, because you can uninstall the package at a later date if you want.

Doing all this allows me to run my Pertelan display on my little thin client that I use as a baby monitor. It now tells me the PIDs of FFmpeg and motion (so that I know that they are working and running), that the Internet connection is UP and the system load, alongside the date and time!

From what I remember when I looked in to it, these prices are pretty good compared to dedicated online backup services (though anyone who needs and can afford the 16TB option needs their head examined!). The only issue is how to get all your files into Google Docs. Well, if your Command Line-Fu is strong, you could use Google CL, but it doesn’t work with a batch of files, unless you write some cunning script. A much easier alternative is to use software from Gladinet, in particular their Gladinet Cloud Desktop tool. This allows you to define various backup tasks, allowing you to sync folders and their subfolders to a folder in Google Docs (or Picasa or a number of other online sources). All your folders and subfolders are backed up to Google Docs, with the folder structure being replicated in Google Docs. You can set these tasks to run every day and it should only update new and changed files. I have been using it for the past month or so and it seems to work fairly well. The only real issue is due to slow upload speeds, but I am not sure whether this is an issue with my broadband provider or a limitation of the Google Docs servers. Either way, the results are that all my files are backed up to Google Docs automatically.