So, my daughter Rose was 18 months old, and I decided it was time for her first synthesizer. Apparently this is not normal, my wife didn’t ever have a first synthesizer (mine was a Yamaha PSS-680), but personally I think everyone should have a cheap and nasty electronic musical instrument at some point in their life! I had a bit of a look round, and you can’t really buy toddler friendly musical instruments that aren’t rubbish. So I am currently trying to build / program one myself, based around an arduino micro-processor module I had hanging around. I had a bit of a requirements/specifications think first and came up with:

1. It should respond quickly without any obvious delay.
2. It should be able to play actual tunes and have an octave of in-tune notes.
3. It should be able to play chords, and/or make horrible noises when you hit lots of notes at once. Kiddy keyboards where it just plays one note at a time are annoying.
4. All the delicate electronic parts should be inside a box away from prying hands.

I made a quick cardboard and tin-foil prototype of an 8 key capacitative sensing keyboard which worked surprisingly well and is jolly clever – each key only requires one wire, which minimises gubbins in the circuit, and there are no exposed moving parts, which should hopefully reduce the amount of toddler damage. I’m thinking of using some kind of metal, something like a bolt through the box for the final version. As a prototype, I’ve made a synth in a Hazer Baba box and a bunch of carriage bolts. I’ve written the synth engine now – features:

1. 8 independent pitch and volume oscillators for nice chord sounds
2. Touch sensing for chord input (up to 8 notes at once)
3. Touch sensing for filter modulation and pitch bend
4. Square, sine, sawtooth, triangle waves
5. Resonant low pass filter
6. All the hard bits written in assembler, so oscillators are fast (14 processor cycles per sample), clipping is handled nicely, and the touch sensing code is extremely accurate (it senses the touch capacitance using an unrolled  loop with 1 input read per cpu cycle). I think that the oscillator main loop is the minimum possible number of instructions per wave sample (see in the code for why I think this).
7. There is time to run a little bit of other stuff on the arduino (there would be tons of time if it didn’t have 8 oscillators constantly running – 1 or 2 oscillators take far less time.) Without the resonant filter it will happily play 16 independent oscillators at once.

Current version of the source code: octosynth-source.

 
Video of Rose playing with the first boxed version of the Octosynth

Video of pitch bend and filter modulation:

Video of a first amplifier test version – I’ve since fixed some electronics bugs that were making the amplifier distort a lot, and it sounds a fair bit better, will do more videos soon. It is very hard to hear on the video as I’m just using a piezo buzzer while I wait for some headphones out plugs / an amplifier chip etc. in the post, but I’ve previously made a couple of quick test videos of the first (very far from toddler proof) setups.

Paul Tennent, Duncan Rowland and a few of us put together this paper (PDF) for ACM Advances in Computer Entertainment 2011 (ACE 2011). The paper is based on a selection of breath based games that we have worked on, using the gas mask breathing interface.

These games are (read the paper for more details obviously):

Serious Sam Aim Control – A modification to the first person shooter Serious Sam which makes aiming of the gun be perturbed by breathing, meaning that to shoot straight, you have to time breathing with shots, or hold your breath.

Hyperventilation Sports – a twist on the early joystick based computer sports games that required frantic joystick waggling to go fast. In Hyperventilation Sports, you have to breathe as quickly as possible, whilst trying not to overdo it and have to stop. This plays on the physical limitations of breathing, and flirts with the danger of hyperventilation (although games are deliberately short enough to avoid any real danger).

PerPing – a two player breathing controlled tennis game – breathe in to move the bat up, breathe out to move it down, hyperventilate to split the ball in two. Miss the ball and your opponent scores a point. PerPing was run for a large audience at Cheltenham Science Festival (read more about Cheltenham).

Sneak-em-up Players have to sneak past guards without alerting them to their presence. If a player is breathing more heavily, the guards can hear them from further away. This creates a game mechanic where people must strategically hold their breath to pass guards.

Tunnel Run is a game for two players. One player’s breathing controls the shape of a tunnel shown on the screen. The second player’s breathing controls the flying of a plane going through the tunnel. The first player must try and make the tunnel as hard to fly through as possible, whilst the second player tries to fly for as long as they can. Players take turns creating the tunnel and flying through it, and are scored based on how long they can fly safely for.

(picture by nottinghamgamecity)

The Broncomatic went to the GameCity Festival, where it was run by Stefan Rennick Egglestone with the help of some other people from the lab, as I was busy doing some other work. 116 riders rode it during the day, and there were pretty decent crowds of people watching from the sound of it.

I did some work before this event on making the control panel more sturdy and resilient, and it held up fine, even with quite a large number of riders, which was a big relief.

We showed the Broncomatic and related bits and pieces to the Discovery Channel for a segment on Daily Planet, which aired on 1st November 2011. Some great close-ups of the Broncomatic control panel construction.

Currently (18th November) the clip is available at the link below:
Discovery Channel Broncomatic Clip.

Want to lay a patio with a random layout?

If you have a load of slabs of different sizes, and want to lay a patio, working out a layout can be fiddly. You need to make sure that you fit in the slabs, taking account of the border shape of the patio, the size of the slabs, how many slabs you have etc.

In addition to that, there are a couple of rules you need to follow, to avoid things that look bad (see the brilliant Paving Expert site for more information).

There are a few programs out there that let you generate a random layout. But at least all the ones I’ve found have two problems:

1) They don’t follow the rules of making a nice looking patio, so you often get cross patterns where four slabs meet, or long unbroken lines along slab edges.
2) They don’t allow you to set limits on how many of each slab you have, which is a pain, as stone merchants often sell slabs in convenient ‘patio packs’, with a fixed set of slabs to cover a certain number of square metres.

So, to do our patio, I hacked together this software, Patio Tile-o-matic. It works with a set of slabs where all edge lengths are multiples of the width of the smallest slab, or roughly so.

For example, my slabs were:
290x290mm
290x600mm
600x600mm
900x600mm

which are all roughly multiples of 300mm.

You can tell the program how many of each slab you have, by pressing ‘M’ to open the materials editor. All measurements are in ‘grid size’, which is a multiple of the smallest slab width – for example with mine, I had
1×1 slabs (290×290)
1×2 slabs (290×600)
2×2 slabs (600×600)
3×2 slabs (900×600)

To use it, you first need to measure the shape of your patio. Then, take the smallest slab width, and create a grid of that size aligned to your patio, for example my patio was 400mm long, so I had 13 300mm grid squares in that direction (with a bit of slack left for pointing).

In the program, you need to edit the patio shape to be the size and shape of your patio, press E to go into the patio editor (or select it from the menu), and click to make the bits of your grid that are not part of the patio black, and the bits that are, grey. Looking at the image above, you can see my patio has a wider section at the top. The simpler and more square you make this, the more likely you will get a decent layout. If a grid square is half patio and half outside, make it grey, you’ll need to do a cut there (you might want to play with the alignment of your grid to the patio shape to reduce the number of cuts)

Once you have done your patio shape, press R to generate a slab layout.

If you can’t generate a layout, check 3 things:
1) Do you have enough slabs to cover the whole space.
2) Try smoothing the edges of the patio to make them less complex – you can always do cuts if you have bits of wall coming into the patio or whatever.
3) Layout happens from the top downwards. There will never be cuts on the top edge – so if you need an edge to be easiest to lay, orient your grid that way round (I put this end next to the house for my patio). It may be easier for it to lay out if the widest edge is at the top.

If you like the layout, press ‘S’ to save it out, this saves it as three files, in the directory that the program lives in – one file is a nice coloured and labelled slab layout diagram (as a png image), the second file is a slab layout in grey, so you can see what it will look like a bit more clearly. The third file is a .slb file, which you can open up again to reload your layout in the software.

You can also add ‘information boxes’, by pressing ‘I’. These let you label elements of the patio layout, such as manhole covers, in order that you can see how many slabs you will need to mess with (for example I generated a load of patio layouts, because I wanted to have big slabs over the manhole cover so as to minimise the number of slabs that needed cutting).

You can alter the layout constraints by pressing ‘C’. Layout constraints stop the layout having long lines of slabs, and also stop it allowing four corners to meet.

Bear in mind:
1) The diagrams produced from this do not take account of differing pointing widths – with slabs like I had, where the small ones have 10mm less width to make pointing neater, this is no problem, with more evenly sized slabs, you may find yourself having to fiddle things slightly on the ground.
2) This software presumes everything is a square grid, due to the nature of real stuff, if you have an space enclosed by walls or something, you will probably find yourself doing cuts even if you generate something that theoretically fits perfectly into the space.
3) It is only as good as your measuring and gridding. Make sure you measure accurately (leaving some space for pointing).
4) For some complex designs, or if you don’t have enough stone, this just won’t work.
5) This is only any good if you want to lay everything to a rectangular grid – if you want lovely curving stuff, you’re on your own.
6) The whole thing is a dirty hack, so don’t expect it to be that polished. It worked for me, and I have a patio thanks to it, but I can’t guarantee much more!

On the plus side, for the job I had to do, it is much better than anything else I could find.

I’m releasing it here, for free, no guarantees or anything. Windows only I’m afraid. If you use it and like it, please do let me know. If you use it and really love it, you’re welcome to paypal a donation to “paypal at joemarshall.org.uk”.

Download it from here.

“Breathe in, and your bat goes up, breath out and your bat goes down. Breath quickly and your bat size increases. Hyperventilate, and the ball splits into two. Miss the ball and your opponent scores a point”

PerPing is a further use of the gas mask breathing interface originally developed for the Breathless breathing controlled swing ride.

We used the breathing interface to make a selection of games, including PerPing, a two player breathing controlled tennis game somewhat inspired by the classic early video game Pong.

We demonstrated PerPing at Cheltenham Science Festival, where several hundred people had a go, and thousands of people watched them play.

PerPing and the other breathing games we designed are also the subject of a paper at ACM Advances in Computer Entertainment 2011.

I presented this paper (PDF) on my running work at CHI 2011.

I did a lot of work studying ‘I Seek the Nerves‘, and used this to suggest some ways of using exercise as part of an interactive system. We suggest that rather than seeing exercise purely as something to encourage people to do with computer systems, we see exercise as something interesting and rich in itself. In particular we argue that doing hard exercise can make people increasingly emotionally involved with the experience which they are taking part in.

We published a paper (PDF) on the gas mask work in alt.chi 2011.

Steve and Brendan presented the paper entirely whilst wearing gas masks, using gas mask microphones to put their voice into the PA system.

I presented a paper that we wrote (PDF) about the Broncomatic breath controlled Bucking Bronco ride at CHI 2011.

For more information on the Broncomatic, see the Rides theme.

Android phones have lovely built in google maps navigation. However, the default ‘car home’ application, and the navigation application are a bit fiddly to enter addresses into compared to a dedicated sat nav system. For example, the onscreen keyboard has letters and numbers separately, with a button to switch between them, and it does not take up much of the screen, in order to allow for content.

You also have to go through a bunch of steps to get to the point where you enter an address – click navigate, select ‘type destination’, then enter it using the small onscreen keyboard.

QuickGo is basically a full screen keyboard for entering addresses or postcodes. Once you enter the postcode in, pressing the search button launches navigation to that address.

Download it from Android Market on your phone.

It is free, but if you like it, feel free to paypal me a donation (to paypal at joemarshall.org.uk)