That is a nasty workflow for sure, but at least it works.

The new software we're working on will implement the new streamlined workflow, but until it's tested and debugged you might need to use the StreamPix workflow. Erg .. that's the software that costs $1500 isn't it?

it's around that price, yes

HEy it DOES have realtime preview of framerate WHILE I capture..this is a very good thing! and it's at 1280x720 full res

Software update
 

I discovered last night that my 22 fps limit wasn't because of the hard drive or disk-writing thread. The frame generation code was not running at full speed for some reason. I am now getting 25-27 out to disk frames pretty consistently. See my updates here if you're interested:
http://www.obscuracam.com/wiki/wiki/...ional%20status

That's good! Despire the nasty workflow, it sounds like you should stick with that while we develop and test our new software. I am sure you'd prefer to save the $1500, but Rob Lohman is right -- this is not going to happen overnight.

<<<-- Originally posted by Rob Lohman :
>David: is 10 or 12 bits always left justified? Is there a reason for it being left instead of right? -->>>

The issue is the editing or compositing software. Pixels stored on disk can be packed anyway you like (compressed or uncompressed, YUV or RGB etc), yet once the application sees the data is most have the format and white level it is expecting. For most 16bit compositors the white level is 65535, for After Effects the white level is 32768 (not 32767 -- so still using 16bit but with lots of headroom.) So the data must be left shifted is it less than the target applications bit depth. Note: 10 bit data for After Effects only needs to be shift by 5.

new windows media clip coming soon, RAW no white balance no color work 8bit

Steve, again why do I need to set gain at 3db for the whites in the image to be true white not grey?

David: thanks... weird that this would be different for different
applications. That's gonna be an issue in the second phase of
the capture process. But the good news is that it is easy to do
anything in that stage/phase.

Steve: Thanks for the PC/104 note.

All:

Personally I was not planning on having the OS on a fixed drive.
I've opted (to Rob) to go with Windows XP Embedded. It is
designed to be small, fast, configurable and able to boot of a
read-only device. So this could fit into flash memory. I'm not sure
how large it is fully.

It allows you to select which component / drivers you want to
include etc. so it should have a pretty small foot/memory print.
I have it lying around here (due to my work) and was hoping to
take a look at it lateron. Our attention is first on getting it to work
before working out on how to make it all small and portable.

The only problem with WXP Embedded is that you can't buy it
in the store. I'm not sure if you can buy this from any company
or Microsoft self as a consumer. An OEM who builds these camera's
should be able to buy those licenses, though.

We'll see where this goes!

Rob: ofcourse you can use my list and other details we talked
about through e-mail on your wiki page. Do put my name next
to it so we know later who said what and why...

Obin:
I'm guessing a bit since the analog stages are internal to the sensor. It sounds like the output of the amplifier without any gain is not large enough to reach the full scale input range for the A/D. The are a bunch more registers internal to the Micron chip that are not documented in our manual - that is only for the basic information that most people will use. We do allow access to all the registers (5F-64 provide black level calibration). XCAP lets you write serial strings to the camera if you want to experiment sending commands. This is a very helpful tool for a programmer to manually try out register commands. They may have done this on purpose to allow the auto black calibration some headroom in shifting the black reference before saturation occurs at the high end.

before after gama work 12bit file 1/2 resolution/size from photoshop jpeg

http://www.dv3productions.com/test_i...itexposure.jpg

still encoding wmv HD!

Hey Obin,
Can you post two consecutive raw bayer pix for us to see the noise on? 10 bits?
Thanks a bunch!

it's going to take 45min to upload but the address is:

www.dv3productions.com/Video Clips/RAW-8-bit.wmv

this is UNTOUCHED footage no white balance no color work jsut converted from 12bit to 8bit and compressed

ouch! the Bayer filter in Streampix hurts my eyes! it looks like a VERY basic filter

guys what is this:

1 Connector for LVDS module (Optional)

the ITX board has that...is that fast? do cameralink cards work on that?

LVDS is the older camera interface, before cameralink.
RS422.
My Atmel 8M's use it.
Is that option expensive?

16bit tiff file

http://www.dv3productions.com/test_i...-24fps1835.tif

pre-Bayer filter black and white image:

http://www.dv3productions.com/test_images/pre-bayer.tif

Wow ... there is a lot of the "zipper effect" left over in that one. (But PaintShopPro is all I have installed on this laptop, so perhaps it's not displaying it right.)

<<<-- Originally posted by Rob Scott : Wow ... there is a lot of the "zipper effect" left over in that one. (But PaintShopPro is all I have installed on this laptop, so perhaps it's not displaying it right.) -->>>

Like I said Bayer plugin is bad in Streampix