FIXED: Movie In TOP to check server for images

RFE: If possible please add the ability of the Movie In TOP to check a server on the web
for any new images. This could be really interesting when using something like FLICKR’s
API. What if Touch users could enter a search term like FERRET and specify
how many random images to grab.

Another way to use this would be for networked performances.

If we had a live jam session and could somehow all access(read/write) into a directory
on the web somewhere that everyone patched into via Touch.

Is this already doable but undocumented?

thanks

oooo… Greg, can we have a Ferret TOP, pretty please?

Selina

Maybe it could be a new TOP that acts like Movie IN but with ftp-like access and tables support
for searching/sorting/etc.

Eventually, if it becomes popular I bet users will want the same thing for movie clips.
And even after that I’m sure people will start saying : “hey it would be cool to output the result
of the processing out to a live QT stream on the web”.

So, this is a big chunk of requests and I leave up to you Touch architects
to consider it all and figure out if it’s useful and how far to take it.

seems like it would be useful for various applications
from networked art to pulling scientific visualisation data off of servers.

so that might extend itself into a table reader that interfaces with a common table
format that already exists out there on the web. plugging touch in to large public data servers.

The only thing that makes sense for a Movie In TOP to do is to add support for paths like http:// or other notations to open a file from a known remote location.

The current Movie In TOP’s Reload parameter can be scripted to attempt re-opening whenever you want.

Tapping into a video stream (like a TV station) is different - there is no length or end of the movie, so that would have to be a separate new TOP I would think.

Searching and re-opening folders looking for new files is too hi-level for any TOP. That’s the domain of scripting, where the script obtains lists of files, puts them in a DAT, makes a Movie In TOP or component for each, and cooks it all. Aside from the searching, TouchDesigner can do that all now I believe.

All these new features are on our radar. For now, maybe someone can try running external python scripts and feeding tables into TouchDesigner.

Yup- that makes sense. Can I constrain the scope of this RFE to:

please add support to the Movie IN Top for filepaths with http:// in them?

all scripting should clearly be done in DAT as you suggest. they rock!

Piclens is a Web Browser (Firefox) plugin
that lets me do an google search on images and then plugs the results into a virtual 3d wall
of images that I can quickly scroll through. It works quite nicely and has a simple and elegant
navigation interface. the pictures on their website doesn’t show you the latest 3D picture
wall (they just added this feature in their most recent release/beta).

it gives a good idea of some possibilities and new ways of searching image data and seeing
the results in a far more compelling way than just a 2d page.

addons.mozilla.org/en-US/firefox/addon/5579

Getting a little off topic now, but your firefox extension reminded my of this other media browsing gem I found recently.

Photosynth can take collections of images from flickr or google and find spatial relationship between them. The related images are analyzed to determine how close they are in the real 3D environment they were taken from. The image set can be viewed in a couple of ways. There is the 2D wall with the image of interest in the middle, and the images surrounding it placed according to how closely they match the original. Then there is the more interesting 3D view which places all the images in the 3D space based on their content. It can even figure out the approximate location of the camera taking the shot and show you that location as well.

What I find interesting is that by matching images spatially like this, you can network huge amounts of information through images tagged with meta-data. Say someone uploads pictures of all the buildings in downtown LA, and tags the images with the names of the buildings (Bank of America, HSBC, etc). Now if you add your own picture of the entire LA skyline to the collection, you can then retrieve the names of all the buildings in your image from the spatially related meta-data. Cool!

Try out photosynth here: [url=http://labs.live.com/photosynth/]http://labs.live.com/photosynth/[/url]
Watch a presentation of the technology here: [url=http://www.ted.com/index.php/talks/view/id/129]http://www.ted.com/index.php/talks/view/id/129[/url]

DONE! You guys are so fast. Thanks a bunch Derivative for the new feature - MOVIE IN TOP as of today’s build now supports fetching images from an http:// path. Very useful!

Thanks seed(),

And your search is here, in a raw format. Put this in a Web DAT:

images.google.com/images?q=dogs

and you will get embedded in the html the URLs of a bunch of images of dogs.

Parsing the html may be tricky, byt the raw stuff is there.