Powered by Twitter Tools.

Powered by Twitter Tools.

@alpaykasal Hah, good, not just me then Will check out seesmic, thanks for the heads-up… in reply to alpaykasal #

@vj_anomolee Just checked out a few of your iSensorii vids – strange and wonderful in reply to vj_anomolee #

@awhillas Poor Apache, you ask too much… in reply to awhillas #

Powered by Twitter Tools.

FrameRunner in action with After Effects CS3

I’ve only just got around to testing FrameRunner in After Effects CS3!

It works fine but thought I might clear things up a little, optimise the code, and add in a requested feature: super sampling.

Some FreeFrame plugins can be a bit blocky, so the super sampling feature scales up the image buffer to smooth things out.

Using super sampling exponentially increases the processing time, but the results look nicer!

FrameRunner Version 1.1 is available from http://www.frame-runner.com

Haven’t had much of a chance to do work on PatchBox over the past few days apart from Sunday where I implemented native support for the fugStreaming library.? I wasn’t getting quite the performance I was after streaming WinAmp across a memory connection so I added an Adaptive Buffer system that allows PatchBox to bypass the usual receive processing and read data directly from the shared memory mapping.

Very fast indeed!

With AVS running at 40fps at 356×256 I get full frame rate speeds with less than 0.75 CPU usage in PatchBox.? Above are some examples where I’m feeding the output through a FreeFrame Gaussian blur, then blending it with a video layer and alpha mask on the GPU.

Also put in the basic code for MIDI support but haven’t actually mapped anything yet.

And fixed video loading for AVI files that need row padding.? Now I’m less limited to 320×240 I’ve been playing about with different formats.? Played around with some 720×576 video that worked fine but for some reason started playing back too fast!? Will investigate…

It’s not done, but I’ve made a start. Am developing in Eclipse and compiling with MinGW GCC. Hopefully this will make everything as cross platform as possible.

Am trying to set everything down as clearly and understandably as possible to enable anyone else thinking about developing FreeFrame effects to use it as a good starting point.

Not sure it’ll be ready very soon (am busy now till Saturday) but I’ll get there.

All suggestions welcome.

There’s nothing quite like jumping in to a new event, in a new
venue, with a software program that’s never been tested publicly
before, and doing an hour of completely improvised visuals in front of
a room full of strangers (many of whom where film makers) to really get
the concentration focussed!

Technically it performed pretty
well. Lots of video clips, FreeFrame effects, LUA scripts, all being
fed into a GLSL shader. I kept vertical sync on, just for safety, and
it seemed quite happy purring along at 30fps @ 1024×768.

It did crash the once but fortunately just as I was fading out anyway!

What
rapidly became apparent was the need to be able to get rapid access to
key features. I had to spend too long dragging and dropping, switching
between property pages, and entering values via the keyboard.

It’ll
be a priority to get MIDI support implemented as soon as possible,
which will help alleviate the value control situation dramatically.

However I was pleased on the whole. There’s obviously a couple of stability issues that need ironing out ASAP, and a fresh set of entries on the to-do list, but altogether a pretty successful first live test!

And they’ve asked me back to do next month, too: Sunday, February 17th.

Sadly didn’t manage to get any filming done this weekend. Think I’ll prepare some clips tonight and try and film the thing in action on Wednesday (am in London tomorrow afternoon/evening for a meeting about this year’s InTransit festival).

In preparation for its first outing on Sunday, I’ve been putting PatchBox through its paces, fixing bugs, and adding the final bits of functionality.

While the main key feature of PB is that it can draw and project on clearly defined areas of architecture, the gig will (ironically) be on a 4:3 screen, as they’re also using my projector to show the films.

Above are some screen shots of an experiment with overlaying several screens to create an interesting composite.

There are four layers of WinAmp AVS output (only one copy of WinAmp running, feeding the output through the fugVideoStreaming plugins, into PatchBox, then applying that to the four layers), and a single video layer on top.

Each layer has its alpha set to the brightness of its pixels, giving the rather pleasing effect.

As part of achieving this I had to hunt down and kill the bug in the WinAmp AVS plugin that was stopping it reconnecting. Will post an update on the wiki shortly…

Spain

Santa Ana, California, United States

Melbourne, Victoria, Australia

Vilnius, Vilniaus Apskritis, Lithuania

Jakarta, Jakarta Raya (djakarta Raya), Indonesia

Paris, Ile-de-france, France

Beijing, China

Duisburg, Nordrhein-westfalen,Germany

Who are you people??

It’s quite closely modelled on Resolume: three layers of video with various mixing effects, but given the speed and flexibility of using shaders I added a few goodies:

• Three texture layers, but each layer has an A/B mixer – so effectively six layers
• Add, add signed, subtract, modulate, and soft luma key mixing between each layer
• Supports alpha on all layers
• A forth alpha masking layer (also with an A/B mixer, so eight layers)
• Oh, you can set the colours of the corners of the background screen to get simple gradients. A ninth layer?
• All processed on the GPU

That’s a lot to play with even before applying FreeFrame effects and scripting!

What really made my brain almost collapse was realising that the shader can be applied to multiple virtual screens, each completely independent of the other. Nine layers per screen. Heh…