4:4:4 10bit single CMOS HD project

[Obin Olson]

here is the dv image we shot lastnight for a scene:

www.dv3productions.com/test_images/hot_background.jpg

the windows in the back of that shot about 30min before that was shot looked very hot and they smear all the way across the shot

[Juan M. M. Fiebelkorn]

Obin,
So?
Do you have a frame showing the smearing?
Are you running the sensor at 48 fps?
What happens with the smear if running at 24 fps?

You mean that white portion of the blue window that goes white??

[Ben Syverson]

I've had absolutely no problem with smearing. Sometimes when I run the shutter at certain speeds, I see some... I would call it "throbbing." It's like a slow pulsing of the image values, and it travel downward. But it's easy to pick a different shutter speed, and that usually takes care of it.

To answer the question about dynamic range, I see nowhere near 1000:1. It seems to be higher contrast than DV. I've played with the image for an hour or two trying to increase the range, but I think it's a sensor issue. If you can control the lighting, you can get good effects, but it's really really easy to overexpose.

- ben

[Juan M. M. Fiebelkorn]

As every video camera, isn't it :)

[Richard Mellor]

Hi everyone I found a link to viper shooting techniques.

some of this may be of help.


http://freespace.virgin.net/shaw.clan/dpviper.html ]





http://www.digitalpraxis.net/


http://freespace.virgin.net/shaw.cla...reen_rev8.html

[Jason Rodriguez]

Ben,

If you saying that it's high contrast because you can't get a well-exposed image, isn't that because you're shooting linear images?

If the camera is outputting a signal with a gamma of 1.0, then I'm assuming that's going to be very dark looking. You'll have to add gamma correction to that after the fact, like a gamma of 1.7 or 1.8 to fit your display.

For instance, I added a gamma correction to the shot that you sent, and I could see way into the shadows, probably a bit too much. But that's just the nature of a linear image, again if the camera is outputting an image with a gamma of 1.0, then you'll never see a "bright" image like you'd see with DV without it clipping severly in the highlights, or you have no contrast in the scene so that everything can fit in the last couple bits.

I've seen this plenty of times before when I've worked with Linear files off of my Canon D60. They are very, very dark, (esp. because they're padded zeros in the last four bits), and they have no gamma correction, those must be added in Photoshop via a color profile. So initially you have a very bad looking, practically unwatchable image, and then it springs to beautiful life with a good deal of contrast range after the gamma correction and color balance is added.

So again, if you're outputting linear images (which it seems like you are), then things are going to be very dark, but that's not an error on the camera's part, or a "there's so much info in the shadows". The only reason there's so much info in the shadows is because the gamma of your monitor is not matched to the gamma of the image so the monitor's gamma makes the image look like it's all in the shadows, but the bits are there, they just need gamma correction.

[Richard Mellor]

link to viper techniques

http://freespace.virgin.net/shaw.cla...reen_rev8.html

http://freespace.virgin.net/shaw.clan/dphd.html

[Jason Rodriguez]

Hey Ben,

Here's a color-corrected version of your shot with gamma correction and a little saturation bump. The only problem I see with it is that there is a good amount of noise and banding in those shadow areas, actually it would be quite interesting to see this stuff in motion.

http://home.mindspring.com/~jrod/head_shoulders_B.jpg

Tell me what you think, does your camera have potential now?

BTW, it would be interesting to see just how dark those shadows were, but you don't have a light meter for me to tell what the dynamic range of that scene is. If that's a table lamp lighting you though, and there's no other lights in the room, then you could easily be looking at 8-9 stops in that room.

Curious to see how your outdoor stuff came out.

Also can you see on the monitor how the camera is exposing, i.e. clipping the highlights? Also are the demosaiced images you posted with your plug-in or Sumix's? It looks as though it's with the ones from Sumix.

[Juan M. M. Fiebelkorn]

Jason, your link doesn't work.You cannot put direct links to files at free hostings and expect them to work.
Just link to a page which contains the image or a link to it and that will work.
Tahnk you.

[Ben Syverson]

Jason, the camera has hardware gamma, contrast and brightness controls. If you want to affect the gamma, it's best to do it on the hardware, because it happens in 10bit instead of 8bit.

The thing is, since it is linear, gamma acts strange -- gamma is really only useful when you have a 0-1.0 range. If you have a larger range than 0-1 (like this camera does before it downsamples to 8), increasing the gamma actually darkens your highlights and brightens the lowest black. So you have to increase contrast and decrease brightness. But that's not too difficult to do. It just makes everything more fiddly.

The problem is really the color contrast -- it's almost non-existant. And as you see, if you increase the saturation, you're asking for color noise. The problem with the noise is that it doesn't move -- those patterns are static. It's not signal noise, it's sensor noise.

You can see exactly how the camera is exposing -- they provide a live histogram for you to use as you adjust all the parameters. And as I mentioned before, the de-Bayering is via my plug-in, not their stuff. I was able to get far sharper and smoother images with my software than with their built-in bilinear or laplacian.

I can't stress to you enough how far this is from 8-9 stops of information. The light on me was a 150w bulb about 18" from my face. There's a perfectly bright overhead light behind me (3 60w bulbs) but it looks like blackness.

And I do have a light meter. :) Never assume.

- ben

[Juan M. M. Fiebelkorn]

So, can't you add to your plug in a noise pattern remover?

What Gain values are you using??

[Ben Syverson]

I don't remember what the gain was on the night shots -- I think there was a little.

The reason I don't add a noise pattern removal system to my plug-in is that there's already one built-in to the camera software. However, it can introduce artifacts too. I'll play with it.

[Juan M. M. Fiebelkorn]

I don't remember well who, but I think I read somewhere here someone was using a gain of 5.0 ....

If you could get a correct balance among Gain, Gamma and contrast, may be you would get better images, but I don't know...

[Jason Rodriguez]

Okay, the link is fixed:

http://home.mindspring.com/~jrod/head_shoulders_B.jpg

The image looks a little softer because I applied Ben's plugin (again), and also added Grain removal in Combustion.

Color-Correction was done with Color Finesse.

So Ben, you're saying that the banding I'm seeing won't move at all? It's a fixed band? It seems as though there would be some noise there.

I'm curious to see what would happen if you were back-lit, so that your face wouldn't be the hottest part of the scene, and you didn't clip the rim on the backside of your head.

[Juan M. M. Fiebelkorn]

Still has a green cast.

To everybody,
Don't be too afraid of color noise, it is easily removed without affecting overall image quality.
What is really important is to keep low noise on Luma..(or Green in this case)