Multi Touch and Houdini - Is it possible?

[rohandalvi]

hi,

Of late , there have a been quite a few videos of multi touch systems in operation being uploaded on Youtube.

Also, companies like autodesk and adobe have demonstrated their softwares on Multi touch systems.

( photoshop on Multi touch ) very cool stuff

(Autodesk Design on Jeff Han's Perceptive Pixel Multi-Touch)

Personally I am a big fan od Multi Touch. I think it is the way of the future, as far as day to day computing tasks are concerned.

It just got me thinking, I can visualize a normal 3d software on a multi touch system, especially modeling and navigation tasks.

But could a nodal system like Houdini be adapted to a multi touch interface? Because once the node network becomes too complex selection and connection could become a major issue.

So, 2 questions:

1. Do you think Multi touch and 3d softwares is a good idea?

2. Do you think a node based software like Houdini could be adapted to a Multi Touch environment?

with regards

Rohan Dalvi

[Jason]

Well, he have a couple of years before the tech is at a consumer price point - but yeah, two-handed input is always quite interesting.

I remember a while back when 3D shutter-glasses became available, along with SpaceBalls and force-feedback styli and such, I thought we'd all be totally absorbed into a stereo, haptic virtual world. It turned out that people would rather just use a mouse, lean back and put their feet on the desk. Working for 10 hours in an immersive world like that would just be too exhausting

Multitouch might be a nice balance between immersive interaction and practical workspaces. I guess the first few people with big R&D depts are going to beat out the best UI conventions and promptly slap IP copyrights on them effectively scattering the market -- like god-damned Apple and their "pinch-zoom" gesture, sigh.

[pclaes]

Interesting topic, I actually wrote an essay (website) on this very recently - if you want I can put it online, but it's a bit big because of all the videos. Here's the gist of it anyways:

If you haven't seen these videos yet, they are quite interesting:

Jeff Han's demo in 2006 of a multi touch screen - since he's also interested in computer graphics he shows of a few demo applications more related to effects ( 03:30 ):

(he also demos it on TED - www.ted.com )

And these two by Johny Lee in 2008 where he uses the wii remote to create a multi touch interface and head tracking, very cheaply ($50 he claims).

( 05:40 )

http://www.ted.com/index.php/talks/johnny_...mote_hacks.html

This shows off the whiteboard a bit more:

his site:

http://johnnylee.net/projects/wii/

1) -> yes, but the interface will play a crucial role. In a way the Hotbox in maya is trying to do this a bit already, by defining zones (north,east,south,west) and quick drag gestures. I also think that since the I-phone a whole new wave of touchscreen gadgets is coming. Once the masses are used to it I don't think people will go back, instead they will demand the same kind of ease of use - but without loss of accuracy - at work. I think the thousands of apps that are flooding the app store are a great way of prototyping all kinds of gestures and interfaces. The easiest and best ones will start forming new standards.

First see how well the new nvidia 3d glasses work with 3d softwares. Games are starting to support them, Maya 2009 might already support them - not sure. That will also help make a difference in helping people let go of their familiar interfacing hardware.

2) -> Not fully, there is too much "scientific data" or expressions that you need to operate houdini well, which lends itself less to multitouch. That said, I still would prefer to type on a flat surface that shows images of keys rather than a physical keyboard. (As Jeff Han mentions in his Ted talk you would be able to set the keyboard to the perfect size for your hands.) I also would prefer one "tabletop surface" while still having a screen in front of me (don't want to look down all the time to avoid straining my neck), this would give you one flat surface. Part of that surface you could turn into a keyboard. To the side you could have a "touchpad" area for using your finger as a mouse, getting rid of a physical mouse altogether.

I think it's much more straightforward for an application like z-brush or mudbox to support multitouch. Still I believe in term of previz and laying down timing for an effect this could save time. I find when I'm explaining a certain effect to someone I tend to use my hands a lot. Even if those movements would only be transferred in a 2d space at the moment, it would allow for showing off different motions fast. Ideally you want to capture those raw movements in 3d, but than you start dealing with 3d motioncapture ( as in: http://www.motion-capture-system.com/shapehands2.php ).

The next step I think will still be full immersion - two cool technologies in particular:

and a few more if you want to know a bit more about 3d ui:

http://people.cs.vt.edu/~bowman/3dui.org/CHI_2008.html

I am curious as well how this will evolve and affect the 3d industry. In any case I don't think anything will drastically change in the next 5 years. And I think Microsofts 2019 vision is very optimistic (

), still it's good to keep an open mind and have people like Johnny Lee who will experiment and push our way of interfacing at an affordable cost.

[edward]

but without loss of accuracy

I think this is a huge problem with touch technology and 3D apps. I made this toy app on my iphone so that one could box select some text. It took lots of practice before I could even box select something simple, you have to take into account the fact that your finger covers up what you're trying to select. I had a novice user try it and it was impossible for them either without the hours of practice I had with it.

[rohandalvi]

wow pclaes, those last two Japanese videos were amazing.

I have seen Jeff's videos, thats what got me thinking. Because in one of his videos there is a small part where someone is working on a nodal network.. But the screen is huge.

http://www.youtube.com/watch?v=9zGDNFpOMcA ( watch it at 0:50, he works on something like nodes)

But as you said, it's all a wait and see.

It would be nice though if Sidefx actually demoed something like autodesk and adobe have done. Just wondering

with regards

Rohan

by the way, any idea what OS Jeff's multi touch stuuf works on, or is it custom?

Edited April 3, 2009 by rohandalvi

[goldleaf]

Interestingly, Pixar posted a paper on this subject recently:

http://graphics.pixar.com/library/MultiTargetUserStudyA/

[Jason]

This is interesting too:

http://video.stumbleupon.com/#p=fx6uximsny

[RonensArt]

Reactable!! - this is Amazing!

This looks like a Node based procedural sound FX.

[RonensArt]

basic demo -

Edited April 13, 2009 by RonensArt

[rohandalvi]

man that's just amazing stuff