Bolex D16: RAW, 16/Super 16 sensor size, costs less than MkIII

Fahnon said:
I'd love to see the day when you can build a camera like a PC. Buy your own sensor in the size you prefer, mount that works with your lenses, body that has the ergonomics you like etc. You could build a camera for specific projects or your filmmaking style.

As long as it wasn't prohibitively expensive that would be the dream...
Click to expand...

This is pretty much where Otis Grapsas is headed with his Drama camera project. He is packaging a GIG-E camera head with his software, your choice of several sensors, installs an OLPF, and calibrates the camera. You can record DNG raw to a laptop computer over a 1gig Ethernet link. All off the shelf hardware. It would be pretty easy to build a powerful micro ITX I7 PC with onboard SATA SSDS into a reasonable camera sized body with one or more HDMI outputs from the processor onboard graphics for EVF's or monitors. Cheap USB touch LCD for menu controls.
 
Postmaster said:
Hmmm....

we have a ton of knowledge on this board here. When we would put all our brains and some money together, combined with some tried and trusty off-the-shelf parts and maybe some custom parts, I bet we could come up with something great. Just think how those communities, that developed this quad copters worked out.

The word "modular camera" could get a whole new meaning.

I mean, think of Dan Hudgins, Vialy, (a pity we lost Otis with his sensor calibration knowledge), some guys of our sound department, Richard Andrewski with his Asia manufacturing know-how, that Russian CNC wizard(forgot his name) andwhatnot combined.

just my ct2

Frank
Click to expand...

You betcha Frank!

I know that if we all wanted to on DVXuser to do this and were to collect all our knowledge- put our resources together into a camera design + system, it would rival this bolex ten times over. Not because they are not doing a good job. Cause there are way to many people here with a good grasp on this stuff, and quite a talented bunch at that in so many other aspects.

Pappas
http://pappasarts.wordpress.com/
http://twitter.com/pappasarts
 
Last edited:
Are you guys being serious? A collaborative camera... interesting concept. But it may be harder than everyone's making it out to be.

To be honest I've never really been in love with that Kodak sensor. But if a collaborative effort here put together a RAW camera with that Sony sensor, well, I'll just have to kill myself!!

We could get A1ex to help with the firmware design and/or programming. I know we also have many other programmers here that could chip in on the software side.

And I'd say we'd be able to raise a significant portion of funds just through the membership base on these forums.

Kind of reminds me of the Apertus cinema camera project. How would this differ (besides the Sony sensor)? Also, they've taken a long time so far and haven't "finished" yet.
 
Last edited:
You guys... it can't be done. It's just the way things are. You cannot change anything or do anything unexpected, or out of the proven tried and true. Ability is nothing.

 
The apertus project is already a good start I think. My only gripe is the sensor they want to use in there next release. It seems super poweful but I have only 1 gripe and it is the DR. Same as with the Kodak CCD in this bolex camera. Look at those ugly ugly highlights. Look at the hat of the little girl it is just clipping and I mean ugly colored clipping in a flat controlled environment. Imagine this in an outdoor environment. Many will have a bad surprise if these shots are representative of the final image quality. It is do digital and just the contrary of the old bolex type of very washed out film big dr film.
 
this seams interesting...
company who makes VAF-5D2 oplf filter for canon video...
http://www.mosaicengineering.com/capabilities/optical-anti-aliasing.html
lens olpf filter with no blurring (full resolution) ...
just additional software correction ...
no need to install in front of sensor (not that easy) ...
buy some cheap industrial camera (gx 2300 is too expensive becuse its very advanced camera )
put olpf lens filter and use Otis Grapsas or Dan Hudgins software to recode dng on laptop ...
what u get is cheap version of si2k mini...
 
@Danielvilliers
i think that is not sensor problem (its good sensor better than one in si2k)...
magenta tint and highlight clipping is bad software raw conversion ...
its a same problem ikonoskop had 3 years ago ...
 
mini-ITX and i7 are way too big and power hungry for what they would be delivering
I was thinking of an android platform with a fast ARM processor with a good graphics side (maybe a Tegra 3 T33 or A9+SGX545)
It also has big advantages for making menus, inputting settings, etc (standard OS, 800x480 touchscreen)
 
Simon Shasha said:
I was just wondering why the makers of the D16 didn't opt for sensor with a bit more grunt, like the CMOSIS CMV12000:

http://wiki.elphel.com/index.php?title=Sensors_table#Hi_speed_sensors

Perhaps they can use it in a D35 4K version down the road if the D16 project is successful?
Click to expand...

I think they wanted to keep a CCD sensor to give it some characteristics DSLRs don't have. Waving around a DBolex with a CMOS would be some crazy jello (unless that sensor has such a quick read out... I am not tech savy to know that)
 
As a quick comment on the whole 4:4:4 debate, I just wanted to point out that these are ratios.

To get 4:4:4, start with 4:2:2 and throw away half the luma data.
 
There are other issues with the CMV12000:

1. Price. It costs about three times what the Kodak KAI-04050 costs, and they'd probably never be able to buy it in the volumes (thousands) necessary to significantly reduce the unit cost.

2. Size. It's APS-C, whereas the Digital Bolex folks wanted to go with something closer to the 16-mm frame size. Elle Schneider did, however, say they might be interested in creating an S35 product in the future.

3. Unknown image characteristics. The KAI-04050 has fairly well-known characteristics, as it and similar Kodak sensors have been used for a while in other cameras. The Prosilica GX 2300 uses a 14-bit ADC, as (I presume) will the Digital Bolex, which outputs 12-bit CinemaDNG raw. The CMV12000's built in ADC can operate at 8, 10, or 12 bits. Would the 2-bit difference in conversion definition matter when the result would be the same 12-bit output? I don't know. Also unknown is the CMV12000's effective dynamic range. The Kodak is rated at 64 dB, while the CMV12000 is rated at 60 dB, although supposedly that can be extended to 90 dB "through several HDR (high dynamic range) operation modes" (here).

4. Data demands. This isn't restricted to the CMV12000 but is true of 4K (and higher resolutions) in general. Because a 4K frame has four times the data of a 2K frame, an uncompressed format like CinemaDNG would be impractical. At 2K CinemaDNG is 60-84 MB/sec (2.5-3.5 MB/frame at 24 fps) and at 4K is, well, 240-336 MB/sec. The lightly (though visually losslessly) compression CineForm RAW at 4K, by contrast, is typically about 30-50 MB/sec. But adding onboard CineForm processing would add to the complexity and cost of the camera.
 
@iamacamera
same sensor as gx2300 ...
CMV12000 dr its not great about 10 stops...
but it offers 2x2 pixel binning so it can shoot in 2K + better DR and in 12 bit mode does 180 fps
 
You must log in or register to reply here.
Back
Top