Manual Mode, AGC off, 1080p30, 1/60 shutter, -3db, ND, 5.6 Fstop. Shot in factory preset and a vivid preset with same results. There is noise in the blue sky and on the blue part of my house. Shot it with my other A1 as well. Link is below. Any suggestions?


TEST on Vimeo

What noise? you must be pixel pipping.

I see noise

Hi Chuck

There is a setting inside the menus called "SKY DETAIL" and the manual describes it as "With the sky detail function, blue sky areas in the picture will be detected and the detail level in these areas will be decreased to obtain a softer picture." It's under the "Menu > Camera Setup > Sky Detail".

I know it's there but have never used it myself so I can't tell you if that's the solution to your problem but it might be worth to try it. If you do, it would be nice if you could share your experience from it.


Regards,

/Bo

The sky detail did nothing. I reset my camera, same problem. I'm more concerned with the dancing grain on my house than in the sky. I shot a red, green, and blue book indoors and noticed noise in the green and blue, but not the red. I also notice when i white balance outdoors it comes out slightly green. My camera is 3.5 years old and has been through a lot of conditions. Could it be something internal? Help.

Is this raw footage from the camcorder if not, what conversion or processing has it undergone?

I do not see much in the way of traditional noise, but I do see two other effects/artifacts

One is caused by the gop size as the full image updates every 15 or so frames, causing apparent jumps in portions of the image as the update happens. The other is a sort of quantization effect possibly caused by the interaction of 8-bit video on large flat areas that are fairly dark coupled with the gop size and a very high contrast image.

You may be able to mitigate it by playing a bit with exposure and gamma.

Chuck, I second the no noise vote. I'm glad I don't see it. What a world where you're bothered by that picture.
Bill

[QUOTE=Don Palomaki;1542406]Is this raw footage from the camcorder if not, what conversion or processing has it undergone?

I do not see much in the way of traditional noise, but I do see two other effects/artifacts

One is caused by the gop size as the full image updates every 15 or so frames, causing apparent jumps in portions of the image as the update happens. The other is a sort of quantization effect possibly caused by the interaction of 8-bit video on large flat areas that are fairly dark coupled with the gop size and a very high contrast image.


Don, are you implying that the noise/artifacts are the result of HDV format recorded onto tape? If that's the case, how would you deal with that?

I see mosquito noise. Long GOP MPEG is my guess as well.

It isn't bad but when the OP said it was there, my eye went to it immediately.

ADDENDUM: it's kind of like film grain in a movie theatre - some people see it, some people block it out naturally as they sort of expect it and then ignore it.

For those of you not seeing any noise, please click on the clip so you can look at it in full screen at vimeo. Because the noise is plain as day and makes the video unusable. Its hard to see if you're just using the window given to see it. And this is what is happening in the camera, not the conversion. The conversion is pretty clean. Thanks

Looks also like a fair bit of chromic aberration. Very hard to judge when it is uploaded to vimeo, I don't reckon it looks too bad though... what was the process from capture to upload on vimeo?

"Because the noise is plain as day and makes the video unusable"

Unusable to you but usable for many, I've seen video with 10 times more noise than that in broadcast and in movies that makes millions, if I am you I would stop worrying about noise and go make some money with it.(-:

Don, are you implying that the noise/artifacts are the result of HDV format recorded onto tape? If that's the case, how would you deal with that?

The MPEG format and/or processing are the likely sources of the noise I see. The sort of banding in the sky caused by where there is a jump in the shade (a digital step) and the goop artifacts.

What we do not yet know is the path (e.g., processing) this video took to get from the camcorder to the post.

I've heard the average human eye can see see about 128 steps (7-bit video) between black and white, but 8-bit (256 steps) looks continuous, no steps/bands. Video signal standards do not provide a lot of data space for the blue component (Blue accounted for about 11% of the NTSC Analog luminance, so subjects that are blue may suffer.)

With out further information, not much I can suggest beyond using a higher bit rate and being vary careful about exposure settings (including knee, toe, and gamma) to minimize the chance that the sky and siding are driven to where banding is apparent. Applying some filters (e.g., noise reduction or blur) in post might help.

I watched it full screen trying to see it. I have to say, I've used event video with big grain so I just don't care that much about using a magnifying glass to find noise. Are you doing a film out or something where you would want this to be pixel perfect? Otherwise I'm at a loss. Ignorance truly is bliss.
Bill

Regardless if the video is usable or not, the problem is how do you avoid noise in an image in the first place? If you are in a studio with great monitor you can tweak your camera, but if you are in the field with nothing but the tiny flip-out monitor on the A1s, you won't know until after the fact. If the noise is the product of compression, then you really have no chance. So, I guess it's just trial and error.

What are the equipment and workflow priorities, then, to ensure the best picture with the A1s or any camera in the field? A good camera-mount monitor? bypass the HDV tape recording? NLE import/export settings? custom presets?

I tried the Neat Video de-noise plug-in. It works great if you don't mind waiting 15 hours to render one hour of video - unless you have a top-of-the-line computer with xeon processors, then the wait may only be 11 hours.

Having done all I could do, then i would consider a bit of imperfection an artistic thing.

Looking at sky footage I shot with almost the same settings with the same camera, I'd say the noise is normal. It looks a bit worse (blocking) after all the compression. Noise is just a fact of video life, and nobody notices. OK - maybe a couple of people. If you want to get it past them, put a little movement in the scene. Then it is completely invisible.

If you do have a huge still monochrome object in the frame, like an unmoving sky, and your shot lazes upon it for a very long time, with no other interesting things going on in the frame, then the sky detail setting on the camera is made for this. Frankly, though, your camera is ok.

Switch Noise Reduction 2 (NR2) to High and see how you go.

Cheers
Shelton.

Looking at sky footage I shot with almost the same settings with the same camera, I'd say the noise is normal. It looks a bit worse (blocking) after all the compression. Noise is just a fact of video life, and nobody notices. OK - maybe a couple of people. If you want to get it past them, put a little movement in the scene. Then it is completely invisible.

If you do have a huge still monochrome object in the frame, like an unmoving sky, and your shot lazes upon it for a very long time, with no other interesting things going on in the frame, then the sky detail setting on the camera is made for this. Frankly, though, your camera is ok.

I agree.

There is some other ways:
- sharpness set to 0 (-1)
- corring set to 0 (2)
- HDF set to M (from "sharp" L), it is a very good way how decrease a visible noise in the whole picture.
- DHV set to 0

you can probably set NR2 to low, but I don't like this way, becouse this setting remove detail from the whole picture. Try use the sky detail setting rather.

Use ND filtr and F4-4,5, for good sharpness of the glass.

Noise reduction settings in the camera cause a trailing effect in the image.

Noise reduction settings in the camera cause a trailing effect in the image.

NR1 cause a trailing effect, NR2 and corring not.

Correct - NR 2 does NOT cause a trailing effect.

Cheers
Shelton.