KINORAW Free Libre software and hardware for cinema

We want to introduce you project Kinoraw, in wich we research and document open source software and hardware in the audiovisual context. For this we count with an open hardware camera, the Elphel 353.

We have focussed the first phase of the project in making more accesible the use of open cameras in educational context, in the way arduino was. For this we don’t need to develop new software but to better document the knowledge from elphel and apertus community and give it a didactical shape. We also plan to document simple use of custom made scripts for non programmers users, showing the floss possibilites to everyone.

For the tutorials we are planning to use blender as video editting software, because of its powerful node compositor and the python scripting. We want to automate some process with addons in order to import video and setup proxies. But as we need to learn a lot about the camera and it’s possibilities and the blender development is so fast (and with mango in the middle), we don’t know yet about specific functionality we will need. That’s why we delayed, perhaps too long, the publication of this notice…

As mentioned before, the first phase of the project covers the didactical aproach to the camera and the software needed (imagemagick, gstreamer, pfstools, blender…) and the development of the scripts will come after that.

In the next phase we’ll be looking for help in the blender community to implement JP4 format (the raw fileformat from elphel). We want this to be decoded natively by blender (nodes and vse). This will be a killer feature to work with the camera, as we will save a lot of conversion time in the actual workflow. For now, only gstreamer, mplayer and an elphel command line app are able to read jp4. For this we plan to spend some money from the crowdfunding paying a developer to do this.

But for the moment we stay with the first phase. Learn and document different workflows with educational purposes.

We are running a crowdfunding campaign in Goteo (an spanish crowdfunding platform in wich commons are the key) with the propossal to release the contents of our research in the form of cc manuals and video tutorials. Our plan is to fund the purchase of some key components we need to work with the camera (a tablet and a beagleboard to develop a viewfinder, some arduino stuff to work timelapses, an ssd storage) and also some money to help us ensure the stability of the project in the coming months. If we reach the first phase of the crowdfunding, we spend part of the money paying the development of the file format issue.

In the first steps, we have been looking for institutional support and we are close to obtain it, even in the last days of the crowdfunding campaign. But some of our first contacted partners have failed us at the last moment, so we rely on a rescue at the last minute. No worries, this is cinema :slight_smile: We have till 10th May to reach the target amount.

Documentation, manuals and videotutorials will be in english, spanish and maybe some other languajes (we have a few volunteers to help with this)

If you want to help us with little (or big) donations, here is the info:

Here is the web project (for now it’s only in spanish, due to technical problems with our drupal server, but will be translated to english as soon as we get some funds…):

Here is the info about jp4 format, maybe someone can check it and tell us how difficult is this to implement.

http://wiki.elphel.com/index.php?title=JP4#JP46_image_decoding_in_MATLAB

Have you tried contacting news sites like OMG! Ubuntu?

Carlos, JP4 is really easy to decode - it is compressed as JPEG just with pixels re-ordered to increase the efficiency for raw Bayer mosaic.
Here is the basics : http://community.elphel.com/jp4/jp4demo.php
The current code we (at Elphel) use ourselves is in Java - http://elphel.git.sourceforge.net/git/gitweb.cgi?p=elphel/ImageJ-Elphel;a=blob;f=JP46_Reader_camera.java . This ImageJ plugin include some other related stuff, so decoding itself is much smaller. Additional advantage of this code is that it has the most current MakerNote data decoding - it is needed to automatically un-apply the in-camera settings (gamma-correction, analog gains) - the result is linear-scale floating point TIFF image.

Andrey

thanks Andrey for the info, and lot of thanks for your support!
Now we have reached the minimum amount and the project will survive!
we only need some blender developer willing to help with this…
in a few weeks i’ll have spare time again and will start to investigate the better way to have this feature into blender.

If there’s a (gplv2 compatible) C lib to read/write JP4 it would be easy to add support to blender.

Carlos, we only returned you the money we received for the camera and are really happy that it helped you to pass the threshold. Yours is a really interesting project and we wish you good luck with it.
I never worked with Blender myself, but my kids will love to have such integration. And as we now have serious plans in 3d (preparing for the SIGGRAPH this year) there will likely be more mutually interesting projects in this area.

Andrey