I must admit that’s a pleasure to work with Flash and to be able to produce multi-screens games (Facebook, iPhone and Air) without redeveloping from scratch.
I’ll be back later with a more detailed article about development and distribution process, so stay tuned!

Thanks to Adobe guys for tips, help and feedback: Aditya Bansod, David Knight, Mark Silverman, Arno Gourdol, Adrian Ludwig, Mike Chambers.
Thanks to testers: Virginie Pain, Alex Aigret, Olivier Costes, François Meunier, Benjamin Guillotte, Pierre-Louis Gonon, Jonas Monnier, Germain Lecourtois, Frédéric MagicFred Smyczynski, Cédric Néhémie, Matthieu Huguet, Julien Faure…
Special thanks to Virginie Pain for bonus items and girl design.

To finish, here’s a teaser that showcases some gameplay, music and sound.

Are you smarter than your friends ?

So I followed these steps to reach my goal:
- Dynamic sound generation: Audio tracks (volume and pan) with insertion of basic dsps.
- Implement time based sound generation: Midi tracks (tempo and steps with mp3 samples assignedon each).
- Create some randomness based on received events: Midi tracks content can be controlled by any event source.

As you can imagine, the possibilities are endless to use this approach in real cases. You could easily imagine to plug a dynamical sound loop with user interactions on a website, generate interactive music loop from game actions to rythm a gameplay, or synchronize any soundscape with screen transitions…

More features and demos to come !

Here’s a basic code example made with a draft implementation of this concept.

var tempo : Number = 120.0;
 
// Create midi patterns defining tempo and steps (quarter notes count) for each
var loopPattern : Pattern = new Pattern( tempo, 32 );
loopPattern.addStep( 0 );
 
var kickPattern : Pattern = new Pattern( tempo/2, 16 );
kickPattern.addSteps( 0, 3, 6, 8, 11, 15 );
 
var snarePattern : Pattern = new Pattern( tempo/2, 16 );
snarePattern.addSteps( 2, 4, 12, 14 );
 
var hhPattern : Pattern = new Pattern( tempo, 16 );
hhPattern.addSteps( 0, 2, 4, 6, 8, 10, 12, 13, 14 );
 
var randomPattern : Pattern = new Pattern( tempo );
randomPattern.addSteps( 1, 3, 5, 7, 9, 11, 13, 15 );
 
// build core audio station and add audio tracks defining volume and pan
_audioStation = new AudioStation( );
_audioStation.addAudioTrack( new Track(new SoundSequencer( Loop1, loopPattern ), 1, 0 ) );
_audioStation.addAudioTrack( new Track( new SoundSequencer( Kick1, kickPattern ), 0.3, 0 ) );
_audioStation.addAudioTrack( new Track( new SoundSequencer( Snare1, snarePattern ), 0.3, 0 ) );
_audioStation.addAudioTrack( new Track( new SoundSequencer( HH3, hhPattern ), 0.3, 0 ) );
 
// create some randomness trigged by keyboard events.
var soundList : Vector.<Class> = new Vector.<Class>();
soundList.push(Kick1, Snare1, HH1, HH2, HH3);
var interactiveSequencer : RandomSoundSequencer = new RandomSoundSequencer( soundList, randomPattern );
textfield.addEventListener( KeyboardEvent.KEY_UP, interactiveSequencer.eventTrigger );
_audioStation.addAudioTrack( new Track( interactiveSequencer, 0.3 ) );

Now, here is a quick demo of the concept.
Drums fills and bells notes arpeggios are trigged by keyboard events.
First, get the focus in the flash animation by clicking with your mouse. The music starts!
Now, you can hit some keys on your keyboard to interact with the music loop in realtime.

Final note: Special thanks to Andre Michelle and his Audio Sandbox for giving to me some light to begin this project.

I tried it this morning and it’s working great. Thanks to Xavier Agnetti for this good piece of software ! I didn’t test at this time how to set custom rules. That’s on my todo list because I think that’s the most interesting part of this project.

I encountered some minor problems to run Ant task as described on the official site. Few imports was missing (commons-lang-2.4, flex-pmd-files, as3-parser-api), some mismatched with the current versions included in the zip (as3-parser and as3-plugin-utils). – Edit: Xavier just fixed the documentation.
Another thing, default rule file was unreachable. I didn’t check the jars to check if the rule file was included. – Edit: See Xavier’s comment below.
I choosed to export it directly from ruleset creator application and target the new file from my Ant task. If you need it, I uploaded it here.

Here’s a snapshot of my Ant task modified to make it working with my project.. Check from line 237 to 258.

Here’s the workaround I found now. Overwrite /usr/lib/flumotion/python/flumotion/common/package.py with the latest committed to branched revision 7960.
Enjoy, everything should be ok now !

So let me introduce the first one.
I needed a multi-users game server. So after some googling to browse OS projects, I decided to build a new one from scratch with Apache Mina 2.0 using my R&D xp.
I choosed to install Ubuntu with Parallels to have an easy workflow with a local test workstation.

At this time, I’ve obtained something basic which works fine. That’s full AMF (using BlazeDS ) and it manages users and rooms.
That’s all based on simplicity. No events, no verbosed implementation or mess with abstract design…

My next step and goals:
- Have something robust.
- Build a graphical chat showcase with text to speech feature.
- Build a game showcase with realtime synchronicity. I read a lot of resources about it, and I’m looking forward to work on this.

But I’m wondering about few things at this time:
- Should I go to open source ?
- Should I mix this implementation with Darkstar project ?
- Should I work on a cheap product ?
Any advice would be welcome !

Here’s a basic echo demo to end this thread and showcase the concept:

import net.tweenpix.pixserver.amf.AMFClient;
 
	public class EchoAMFClient 
	{
		protected var _client : AMFClient;
 
		public function EchoAMFClient() 
		{
			_client = new AMFClient( );
			_client.registerClass( TestVO );
			_client.connect( "10.211.55.4", 9123 );
 
			_client.addEventListener( "EchoMessage",  onEchoMessage );
			_client.sendRequest( 	"EchoMessage", 
								new TestVO( "francis", 38 ), 
								new TestVO( "virginie", 35 ) );
		}
 
		public function onEchoMessage( vo1 : TestVO, vo2 : TestVO ) : void
		{
			// do something
		}
	}

Liens:
- Adobe AIR 1.5.1 Developer and User Release Notes
- Adobe AIR 1.5.1 Documentation
- Download the Adobe AIR SDK
- Adobe AIR 1.5.1 Known Issues and Limitations

Via air team blog

Oui, je suis un inconditionnel de la première heure.
Merci à sh3sha pour l’info !

- Arrivée d’un nouvel SDK pour la lecture des PDFs et eBooks sur mobile.
- Publication d’une solution de packaging pour Flash Lite
- Adobe et Nokia s’associent et débloquent 10 millions de dollars pour encourager le développement d’applications multi-screen.
- Palm rejoint l’Open Screen Project
- Disponibilité du Flash Player 10 pour tous les smartphones dès 2010

Retrouvez tous les détails de ces annonces sur interfacesriches.fr avec comme promis lors de mon précédent billet, un article de fond sur Open Screen Project et la stratégie globale d’Adobe en matière de conquête du marché RIA.

A paraître dans les prochains jours, un article d’investigation qui devrait vous plaire. So stay tuned ! ;)