Here's an interesting experience made on Youtube, to show what happens when a video is uploaded to Youtube, ripped from it and re-uploaded... a thousand times. "Canzona" did this, uploading its first video one year ago, and the last one on may 27nth. He made this as an hommage to Alvin Lucier who made the same experiment, recording his voice and then playing it again, re recording etc...
All thousand videos are available on Canzona's page, but here are ten videos to show the progression of the photocopy effect.
This evening, a friend of mine asked my help on his project with Virtools : he had a detailed model, without the source file, and without any other way to add some occlusion.
As his model had a strong tesselation, I made him a small script to add Ambient Occlusion in the vertex color. After about 15 minutes, he had the rendering he wanted (a little shader helped to have some reflection added to the ambient occlusion). Of course it's not very much, and not very innovative, but anyway it's been a long time since I last posted a Virtools demo, so here it is :). Don't forget that you'll need the 3DVIA Virtools Player to see the demo.
A few weeks earlier, the following image was posted on DogHouse Diaries. Randall Munroe, from xkcd created an online survey and asked internauts the name of random colors. 220 000 people took the test, giving many interesting results (see here).
How men/women see colors
The most interesting thing from this survey is the image underneath, making a map of color names. We see that there is no bijection between the RGB space and human language, either or french, english or whatever. This mean that given a specific RGB value, we can find the corresponding name, but from a single name (pink for instance), hundreds of color values are valid.
Map of color names
This would be an interesting approach to look at, in arts and artificial intelligence. Computers do not have the fuzzy boundaries between red, orange, gold and maroon. They only think in digits, which is too heavy for artistic purpose. One would need to setup a blurry color palette, where it would pick for blue to paint the sky, green for plants, instead of picking #0088FF or #49DE59. Of course these are just thoughts as I browse my RSS, but I guess some of my friends at "Les Algoristes" would find this interesting.
All colors are the friends of their neighbors and the lovers of their opposites.
As you know, in the beginning of april was the 11th Laval Virtual, and this year was my sixth time in a row.
So here are my thoughts and feedbacks on this years edition.
On Laval Virtual itself
On a general scale, I found this year's edition bigger, and richer than previous ones. There was more space
inside for animations and exhibitions, and the award ceremony was very... inspiring. Anyway it was also very
good to meet many enthusiast researchers, students and experts.
Open Space 3D
Every one who hes been to at least one Laval Virtual knows that half of the booth are Powered by Virtools,
and it's rare to see a new software in the place. That's one of the reason why I've been glad to discover Open Space 3D,
which is a Virtools like editing software.
As its name suggests, Open Space 3D is an open source realtime 3D engine, using a system similar to Virtools' Building Blocks.
Created by french engineers, it's still in alpha state, and does not have the maturity of Virtools, but
being open source, it may gather some of Virtools fans and become the Blender of relatime interactions?
The "new" Virtools ?
Mommy Tummy
The nearest feeling to being pregnant
Here's an interesting work from a Kanazawa Institute of Technology, in which man can experience
(some of) the feelings of bearing a child. Using a harness with pockets filled with sensors and water,
the man can understand how tiresome life can be to a pregnant woman.
Using some motors, men can also feel the "virtual baby" moving and growing.
The experience is very strange, cause you really feel something moving, almost inside your belly. And
women who had child already, said that the feeling was much the sam to a real pregnancy.
This project is both very simple yet very poetic. Based on a classic IR head tracking and a wand, you move
in front of a screen, or ideally in a CAVE, and paint in space in front of view. With the head tracking,
the view is adapted to your position, letting you move around your painting.
Along with the particles, symbolise the virtual paint, you have sounds and music being played according to
the strokes being made. The resulting experience is very strange, yet I found myself staying in the cave for
10 minutes without getting tired. A small example is available in video on
Dailymotion
Camera-less Smart Laser Projector
Laser Projector used to highlight veins
This project use a detection system without a camera, using an infrared laser which can detect either ink
on a paper, or veins through the skin, or just black/white contrasts, and then control a tiny mirror which let
a visible red laser to lighten the exact same point. Then both lasers are moved on the targeted surface fastly
to give the impression that there is a standard video projector.
It can also be used for more artistic projects, like in the video below, to create sounds and ride on the hands
of someone. This project has won the Emerging Technology award, and will be presented at SIGGRAPH 2010.
Using a mix of water and starch, the team from Osaka University made a canvas which can change between
solid and liquid state, locally. With a glove enhanced with local tubes, they can pump or inject water
to make these changes around your finger and make you experience different feelings : stickyness, hardness
and roughness.
A state changing substance under your hand
Organic Motion
The Open Stage platform is a quite stable system enabling real time markerless motion tracking
of a person, using 12 infrared cameras. Each camera gets a depth map of a human body moving inside
a cubic space, then all the images are combined to get an accurate 3D representation of the body.
Then an analysis is performed to find all the bones and bind them to a 3D avatar, moving on a screen
accordingly. The latency is very small, and the tracking is accurate even to the motion of the head.
Testing Organic Motion's tracking system
Lexip
3D mouses have been available for quite some time now, yet each year we see a new mouse trying
to find a place on the market. This year, Lexip, a french company, presents its mouse. Using a tilt/roll
control directly on the base of the mouse, plus a joystick on the thumb, enabling 6 DOF liberty.
The use of the joystick together with the tilt/roll is quite strange, and somewhat less instinctive than
other mouses where all the 6 DOF are using the same control (like Logitech's space mouse). But you're not
bound to use it for moving in 6 axes, and use the Joystick or tilt/roll independantly for your application,
as the pro version comes with a SDK to make your own bindings. The basic version coming only with profile
for severall games and DCC softwares (3DS Max, Trackmania among others).
Lexip's 3D Mouse
AR Pool
I already talked about this project earlier here, but here's a feedback after really trying the thing.
One of the main drawback of this project is that the human hand is not as precise as needed
to be able to use the help of the system. Besides, one need to setup by hand the force applied
to the cue, and the angle between the cue and the horizontal plane.
Trying the AR Pool
CRISTAL
The living room of tomorrow
Grand Prix of this year's Laval Virtual, CRISTAL, or Control of Remotely Interfaced Systems
using Touch-based Actions in Living spaces, is an example of working domotic. Using a touch
screen on a cofee table, one can control the lights in the room, browse a library and send
either videos, photo albums or music to the television, or even control a vacuum cleaner.
I got the opportunity to talk with David Arenou, a student behind the project Immersive Rail Shooter, so I decided to try something new and make an interview. Here are 7 questions to know what this project is about.
Hello David, can you first present yourself?
Hi! Well, I am 22 years old. I am in my last year at L'Ecole de Design of
Nantes where I studied interaction design. I am currently doing my final
internship at Alcatel-Lucent Bell Labs France, in the User Experience
Design department.
How would you describe your project, "Immersive Rail Shooter", in one
sentence? Did you work on it alone?
I'll try in one sentence: « IRS » is a video game which uses real space as
playground and which engages player's body in the intense action of a rail
shooter. Let me add a few words: beyond this application, my diploma
project is a reflexion about immersion, gaming, and mixed reality as a way
to immerse. It was a solo project and I dedicated it half of my final year.
What was the main technical challenge you faced while making this project?
From the beginning, my project was fixed on the current video game context
(Wii, Natal, Move) and I meant not to define a long-term concept. So,
during the creative phase, while prospective concepts abounded, the
challenge was to keep in mind this short-term issue. I also really wanted
to prototype my concept in order to make some usability tests. Thus, I had
to anticipate and think about which technologies were accessible to me.
Regarding marker detection, which technology did you use? Any particular
reason for this choice?
I'm used to work with 3DVIA Virtools because it's kind of easy to get to a
convincing result. So I naturally headed toward this software to create my
game. However, I still needed a detection technology to track the player's
movements. I knew that ARToolkit was available on Virtools and after few
tests with big markers, I decided to use this library for my prototype.
Finally, it was the opportunity to catch up with Frantz Lasorne who worked
with ARToolkit for his diploma project « Scope ».
Were there features you wanted to add which weren't implemented for one
reason or another? Are there some you didn't plan and which were included
while experimenting?
I would have loved to add head tracking to my prototype. It could have
been possible. When isolated, this feature worked pretty well actually,
thanks to the head marker orientation. The only problem was that it
definitely took to much computer resource. Once added to the rest of the
application, it slowed down the whole thing.
Otherwise, I didn't have the chance to discover spontaneous features. I
would rather say that I did have the chance to validate almost all
features that I thought, during the prototype phase.
Immersive Rail Shooter concept
What are the possible evolutions of "Immersive Rail Shooter"? What are
your plans for the future (for yourself and/or the project)?
I truly believe that something which looks like « IRS » is already being
developed in video game studios. Of course I'm referring to the Microsoft
Natal project which could have been the perfect device for « IRS ». Let's
wait until next E3 (in June) for announcements from Microsoft and other
publishers, to see what they're going to propose.
As far as I'm concerned, I'm still working on my prototype. I'm going to
participate in the Virtual Fantasy student contest at Laval Virtual 2010
(the International Conference on virtual reality). My orange helmet and I
will be there to make some demos !
Is there something I didn't think of you'd like to talk about?
Come to Laval next week! I'll be glad to show « Immersive Rail Shooter » …
and even let you people try!
Thanks, David, and see you next week. I'll also be at Laval Virtual, probably mostly around the booth H4 with the Paris ACM Siggraph.
A reader is not supposed to be aware that someone's written the story. He's supposed to be completely immersed, submerged in the environment.
Voici un projet qui rentrerait bien dans le cadre des Algoristes, et qui pourtant en est assez loin. Christian Bök un poète (experimental) canadien a pour projet d'écrire un poème... avec de l'ADN.
Utilisant une bactérie extrêmement résistante, il compte changer un partie de son code source (ADN), pour que la chaine de nucléotide forme un poème compréhensible. Bien évidemment il ne va pas se contenter d'écrire un poème sur GATACA, mais simplement définir un alphabet ou chaque lettre est représentée par une succession de trois nucléotides. Ainsi ACT donnerait A, AGT donnerait B, etc...
Mais la difficulté ne s'arrête pas là : la chaine ADN va ensuite créer une séquence d'acides aminés. Après avoir composé son poème sur ordinateur, des laborantins créeront la chaine de nucléotide et l'injecteront dans une bactérie, qui sera alors libre de se multiplier, et survivre à Bök, et peut-être même à l'humanité...
Les entrechats d'Eros mènent aux entrelacs de l'ADN par un sacré culbutage des chromosomes.
Lui : la fille de l'appart en face, j'arrive pas a savoir si elle est bonne ou si elle est vieille. Moi : j'ai envie de dire qu'elle est jeune. Lui : le reflet du pc avec winamp qui tourne me ferait dire aussi que c'est une jeune mais... Moi : t'arrives a reconnaitre winamp et tu vois pas si elle est jeune ou vieille?! Lui : ben j'arrive mieux a reconnaitre winamp que les filles...
Encore une fois, le succès étant au rendez vous, le Paris ACM Siggraph organise une nouvelle soirée "Rencontres Start-up et Recherche" le 1er avril prochain (non ce n'est pas une blague!). Au programme, comme toujours, trois intervenants :
The Bakery
la jeune société, qui développe des logiciels pour la synthèse d'images, présentera sa suite de logiciels Relight, orientés vers l'éclairage en 3D;
Delacave
studio d'animation (dont j'ai déjà parlé ici) viendra présenter son pipeline d'outils temps réels pour la production de film;
HD3D
le projet HD3D a pour but d'aider la collaboration au sein des différentes boites de productions.
Plus d'information sur la soirée et les intervenants, ainsi que le formulaire d'inscription, est disponible en ligne, n'oubliez pas de vous inscrire à l'avance.
Et en plus de cela, un évènement régulier va prendre place tous les deuxième jeudi du mois : Bring Your Own Animation, ou des étudiants, amateurs, juniors, peuvent venir présenter leurs animations à des professionels seniors, et avoir des retours. La deuxième édition aura donc lieu ce soir, au bar Le Plastic, 13 rue Jean Bausire, près de Bastille. L'entrée est libre, et pour ceux qui veulent présenter une animation, il suffit d'avoir une clef USB avec une video lisible dans VLC.
C'est ce que nous sommes tous, des amateurs, on ne vit jamais assez longtemps pour être autre chose.
The Laval Virtual Festival is only one month away, and we start seeing videos of what will be shown there. Here is an example of an Augmented Reality project called ARPool, where augmented reality is used to enhance a pool table.
The system, called Deep Green (in reference to Deep Blue, an AI used in chess game), analyses whatever is on the table, balls and cue, and can compute the physic path of the white ball, should the player hit it without error. Then it displays this path on the table, allowing the player to adapt his game to get a better shot. The video below should explain this better than me.
Now when I first saw this video, it instantly reminded me of an old TV show : Quantum Leap. Geeks of my age should remember this SciFi serie in which Sam Beckett leaps through times in the mind of people to put right what once went wrong. In one of the episodes, he uses a trick from the future to display lines only he can see, in much the same way, except that the line displays the best shot ever, and he only has to follow the guide. Here the video from this episode.
Now this is one of the great thing with Laval Virtual (and its the same with the Emerging Technologies at Siggraph), we can experience ourselves what movie makers, and book authors have made us dream of. So this is one dream come true, maybe by next year someone will build an actual working Tardis... or not.
It is impossible to imagine Goethe or Beethoven being good at billiards or golf.
Unlike most of the people of my generation, I don't really like modern music, and keep to classical music and original soundtracks from movies. Anyway, today I learned that David Cope had built an Artificial Intelligence software capable of composing classical music.
Sample music from Emily Howell
Emmy, from Experiments in Musical Intelligence, was born in 1981, when David Cope started research in computer music. Himself a composer, he wanted a software that could analyse his music, and tell him what should be the next note, or the next measure, or even all the notes to finish the music piece.
Now, after a few years of silence, Cope is back with Emily Howell, his new software whose aim is not too complete his work, but create her own compositions. The way Emily works is by conversing with David Cope, analyzing succession of notes and giving a rate.
Like many Algorists out there, David had to cope with people saying that computer generated music has no soul. But the same man (a chemistry professor) heard the same piece twice, the first time without knowing that it was composed by an artificial intelligence. The first time he heard it, he though this was the best piece of music ever, but the second time, when he knew it was computer generated, he simply said : You know, that’s pretty music, but I could tell absolutely, immediately that it was computer-composed. There’s no heart or soul or depth to the piece.
So I guess the problem is not the art made itself, but just the fact that it was not made by a human being. Anyway, Emily's music is the best computer generated music I've ever heard, with some kind of Philip Glass feeling (I think).
A composer is a guy who goes around forcing his will on unsuspecting air molecules, often with the assistance of unsuspecting musicians.
Some time ago I talked about Net.Art projects which used websites as raw material to create art works. Today, here's a new example of this kind of work, but instead of creating images from the content of the page, Code Organ uses the source code of a website to compose music.
Code Organ converts a web page into music
The basic idea is to find all the characters between A and G, corresponding to the 8 notes, then uses them to build a melody. The analysis also chooses a specific instrument effect as well as a percussion track. This then creates a unique music corresponding to the input web page.
If you wan't to know what my blog sounds like, just go to this page, and try to make your own website music...
Nikki Graziano defines herself as both photographer and mathematician, and i must agree with her. When she take a photo of nature, she then draws a math functions to map (more or less) the shapes in the photo.
Here's an exemple of what she can do, more exemples being available on her website.
Found Functions - Nikki Graziano
As far as the laws of mathematics refer to reality, they are not certain, and as far as they are certain, they do not refer to reality.
