Developing an opensource camera platform

[Adrian Onsen]

Has anyone ever considered starting an opensource (community driven) effort to develop an open source camera system?

I'm referring to a system that first of all runs on an OS that's not closed source, a system that is lens type agnostic (with adapters could use Canon/Nikon/Sony/etc lenses) and is modular in that different parts can be upgraded individually (kind of how a PC is modular in what PCIe card, or CPU, or soundcard, or videocard can be used)

I've given this a bit of thought on this lately and have a lot of ideas, though due to lack of deep electronics and programming knowledge, (have lots of mechanical experience) I've stagnated at conceptual/dreaming phase.

There are a lot of clever people on this group and I would be willing to set something up if anyone else is interested in such a project.

Any thoughts?

adrian

[Wesley Clouden]

Sure thing it makes great plan

It's all about putting our resources together and creating opesoure arm then we can move Ffwd 

IPhone 3Gs dualboot  android / 5.1.2ios beta

ウェスリー D クラウデン

Wesley D Clouden

CIO/ Media Content Producer

Nappiee Corp Inc. // Mogulus LLC. // Twitter LLC

Qik.com , foursquare.com, InArtMedia.com

InYoFaceFilm.com, mogulus.com, Twitter.com

Skype //  Aol // Msn // Yahoo ID: Nappiee007

Wesley....@gmail.com

Dir: NYC 24/7: +1646 248-6680

***********************************

--
http://magiclantern.wikia.com/
 
To post to this group, send email to ml-d...@googlegroups.com
To unsubscribe from this group, send email to ml-devel+u...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/ml-devel?hl=en

[mohan manu]

Why not try on existing on going Open source project Apertus?

As in Apertus.org :
The goal of the Apertus project is to create a powerful, free (in terms of liberty) and open cinema camera that we as filmmakers love to use. The idea of using an Elphel camera for this particular purpose was born in 2006, found many followers over the years and ultimately resulted in this community driven project entitled "Apertus" (word definition) and this website.

[Morgan Look]

The Apertus project looks really interesting.  The current camera is based on a very small sensor, but they are about to launch a kickstarter campaign to develop a large sensor front end.

Morgan.

[Jim Burnes]

I imagine basing it on Linux and using the real-time subsystem would do the trick.   The problem is that it requires not just a large sensor, but all of the supporting ASICs (application specific integrated circuits).  On the Canon this would be the Digic supporting circuitry.  The real trick to making it economically viable is to find a sensor and circuit manufacturer that mass produces them but doesn't have a vested interest already in, say, the indie film market.   Good luck with that.

Or you could take one of the mass produced camera's and completely reverse engineer its sensor and control system (sort of what ML is, but more hard core).

The thing is, producing a truly high quality sensor is a very significant engineering feat and requires an incredible investment in R&D.

But assuming a company somewhere is doing such a thing, it would be great if you could find a quality HD sensor based on CCD technology instead of CMOS.   That way you could avoid the rolling shutter / jello effect.  Even the Red One had rolling shutter / CMOS artifacts.  This leads me to believe that HD CMOS sensors are crazy expensive.

But it would certainly be worth looking into.   It would be an incredible challenge to develop such a camera and make it mostly portable.  Keeping it to within twice the size of a T2i would be a miracle, but a worthy challege.   If you could just make the head end the grabs the sensor stream and sends it out an HDMI port, you could let BlackMagic / Atomos worry about the recording.

jim

[mohan manu]

Jim,
It has already produced nearly production ready camera with Aptina sensor using elphel toolkit. It already recorded raw bayer data and needed an computer to do high quality de-beyaring. The apertus community produced software tools to do de-bayer to Cinema DNG format. It is CS mount and i think it could mount Super 16 lenses. Only the thing is it has to be connected to a laptop as viewfinder cum recorder.. It is a great project still evolving. They now doing crowd funding for creating much more superior sensor and good workflow for converting raw data and all.
Link on more info about Apertus project:

http://www.youtube.com/watch?feature=player_embedded&v=mvI7rJ_AZys

They have done 3D recording too.. With 2 sensor on one board!!

http://www.youtube.com/watch?v=popTxxJ4zpo

But assuming a company somewhere is doing such a thing, it would be great if you could find a quality HD sensor based on CCD technology instead of CMOS.   That way you could avoid the rolling shutter / jello effect.  Even the Red One had rolling shutter / CMOS artifacts. 

CMOSIS CMV4000 is an Global Shutter sensor, which can do 2k at 180fps. But its high quality industrial purpose sensor.

[arm.indy]

See also
http://graphics.stanford.edu/projects/camera-2.0/

Indy

> >> idea of using an Elphel <http://www.elphel.com/> camera for this

> >> particular purpose was born in 2006, found many followers over the years
> >> and ultimately resulted in this community driven project entitled "Apertus"

> >> (word definition <http://en.wiktionary.org/wiki/apertus>) and this
> >> website.
>
> >> On Sat, Apr 14, 2012 at 7:25 AM, Wesley Clouden <wesley.clou...@gmail.com

> >> > wrote:
>
> >>> Sure thing it makes great plan
> >>> It's all about putting our resources together and creating opesoure arm
> >>> then we can move Ffwd
>
> >>> IPhone 3Gs dualboot android / 5.1.2ios beta
> >>> ウェスリー D クラウデン
> >>> Wesley D Clouden
> >>> CIO/ Media Content Producer
>
> >>> Nappiee Corp Inc. // Mogulus LLC. // Twitter LLC
>

> >>> Qik.com , foursquare.com, <http://InArtMedia.com/>InArtMedia.com
> >>> InYoFaceFilm.com, mogulus.com, <http://Twitter.com/>Twitter.com

>
> >>> Skype // Aol // Msn // Yahoo ID: Nappiee007
>

> >>> Wesley.Clou...@gmail.com

> >>> Dir: NYC 24/7: +1646 248-6680
> >>> ***********************************
>
> >>> On Apr 13, 2012, at 9:48 PM, Adrian Onsen <aon...@gmail.com> wrote:
>
> >>> Has anyone ever considered starting an opensource (community driven)
> >>> effort to develop an open source camera system?
> >>> I'm referring to a system that first of all runs on an OS that's not
> >>> closed source, a system that is lens type agnostic (with adapters could use
> >>> Canon/Nikon/Sony/etc lenses) and is modular in that different parts can be
> >>> upgraded individually (kind of how a PC is modular in what PCIe card, or
> >>> CPU, or soundcard, or videocard can be used)
>
> >>> I've given this a bit of thought on this lately and have a lot of ideas,
> >>> though due to lack of deep electronics and programming knowledge, (have
> >>> lots of mechanical experience) I've stagnated at conceptual/dreaming phase.
>
> >>> There are a lot of clever people on this group and I would be willing to
> >>> set something up if anyone else is interested in such a project.
> >>> Any thoughts?
>
> >>> adrian
>
> >>> --

> >>> <http://magiclantern.wikia.com/>http://magiclantern.wikia.com/

[mohan manu]

Hi Indi,
Camera 2.0 is an University project funded by Adobe Systems, Kodak, Hewlett-Packard, The Walt Disney Company, Intel, Texas Instruments, Google, NVIDIA, and Sony. So not Open source. Only Software for Nokia N900 is Open source. So, out of discussion.

[Jim Burnes]

Thanks for the update Mohan.  Looks like they've done some fantastic work.  I spent the morning looking at sensors you can purchase from the industrial and photograpic industry.  You can quickly get overwhelmed.   For example, Foveon sells the X3, which doesn't need Bayer processing, but it can only sample at 4fps. -- etc.

I'll check out their site.