I tried your settings using dcrawms. The resulting images never matched the quality of FlyCap's YUV422 output. They were fairly splotchy and some stippling remained.

However, today I discovered something very promising. Using your version of FlyCap, I selected RAW16 instead of RAW12 and chose DNG export. When I opened up the resulting DNG in Photoshop, the top 75% was garbled, but the bottom 25% looked perfect - very similar to the YUV422, no stippling at all. I've attached a graphic that compares YUV422 exported to TIFF, RAW12 exported to DNG, and RAW16 exported to DNG.

Do you think my stippling issue could be related to bit depth interpretation?

Did you try the dcrawms fix I mentioned ?

Can you also try to set Gain to 0db and capture dng ? Turns out that gain is only modifying the digital gain which will definitely introduce artifacts.

> The A/D converter provides a PxGA gain stage (white balance/preamp) and VGA gain stage. The main VGA gain stage is available to the user, and is variable up to 24 dB in steps of 0.046 dB.

Here's a dropbox link for a zip that includes a .raw12 file and a .DNG created from that file in DANcinec: https://www.dropbox.com/s/m50y2mxsrcn818x/Stippling%20Sample.zip

I will use this site as a reference, and print out DNGs with various matrices. Then I will pass them to you - so you can process them. http://www.brucelindbloom.com/index.html?Eqn_RGB_XYZ_Matrix.html

Once we can figure out what works, I will incorporate them into DNG headers.

I did a couple tests to evaluate whether it's FPN.

First, I recorded 70 frames with the lens cap on and the gain set high, then I averaged those frames in Photoshop. The result is an almost pure black frame - no obvious FPN. If I boost the exposure of the average frame by +8 in Photoshop, I'm able to see some faint vertical lines and a couple hot pixels, but nothing severe. I tried subtracting the average dark frame from a normal frame shot with the same settings, but it didn't help at all with the stippling.

Second, I recorded 400 frames of panning across a room, both in and out of focus. After converting these frames to a minimally compressed video file, I'm unable to see any FPN. The noise seems utterly random.

Out of curiosity, I also shot some sunny outdoor footage with the gain at minimum. Regardless of the low gain, I get the same stippling. It's more or less just as bad as footage shot at high gain in low light.

Have you tried dark frame subtraction and bad pixel removal first ? Using one of the astrophysics software. Like Pixelfixer or Deep Sky Stacker

is that stippling or sensor noise? do you see it with dng generated with dancinec?

can you post a screenshot of the stippling?

Try this release build for DNG - https://dl.dropboxusercontent.com/u/9906333/FlyCap2.exe Also, make sure you are using streaming mode not buffered. This one will generate the correct DNG header for any resolution you are using.

Are you using the Intel USB3 drivers or the Point Gray ones ? Can you open DriverControlGui and confirm that the Point Gray drivers are being used ?

If you connect a USB stick to your laptop, can you see the USB stick ?

With the PGR drivers, you will only be able to use the USB3 ports for a PGR USB3 camera. You would need to manually swap them using the GUI to use any other USB3 devices.

Also, the build is the debug build - without optimizations.