Are thehere dead hot or stuff pixel in this vieo footage?

[Stears555]

HI!

Here is the 50 Megabyte little sample file, which was recorded in a dark room (by this why I can check the pixels): http://dictaphone.atw.hu/00069.MTS

Concentrate around the white pixels the bottom left and bottom right corners. The pixels are not always appear, they appear disappear rhythmically alternating (appearing and disappearing on the same place from frame to frame) on the same position. watch the video with frame step)

One of my friend said , this is not a real dead pixel, because dead pixels are stable on the video. But what is the reality?


Important Notice, the video is a so called progressive segmented frame video (progressive material in interlaced format), so for good result, please turn of the deinterlacer/automatic deinterlacer in your player!

What do you think about it?

Thank you for your reply!

[Stears555]

The camera gain was turned off. ( 0db. )

[jagabo]

I think that's normal CCD noise. Crank up the gain (in software) and you'll see flickering white spots all over the frame.

[LMotlow]

I think that's normal CCD noise. Crank up the gain (in software) and you'll see flickering white spots all over the frame.

Same here.

Important Notice, the video is a so called progressive segmented frame video (progressive material in interlaced format)

?
MediaInfo says it's interlaced.
Either way, same conclusion. White dots are everywhere.

[Stears555]

I think that's normal CCD noise. Crank up the gain (in software) and you'll see flickering white spots all over the frame.

Same here.

Important Notice, the video is a so called progressive segmented frame video (progressive material in interlaced format)

?
MediaInfo says it's interlaced.
Either way, same conclusion. White dots are everywhere.

But my problem is the following: The white points appear and disappear always exactly on the same place. (after three frames later. ) Do you think that, If they are real dead pixel, they would appear constantly instead of flashing rhythmically some frames later?

[Stears555]

I think that's normal CCD noise. Crank up the gain (in software) and you'll see flickering white spots all over the frame.


[Attachment 27622 - Click to enlarge]

But my problem is the following: The white points appear and disappear always exactly on the same place. (after three frames later. ) Do you think that, If they are real dead pixel, they would appear constantly instead of flashing rhythmically some frames later?

[jagabo]

But my problem is the following: The white points appear and disappear always exactly on the same place. (after three frames later. )

Some do, some don't. Some pixels are noisier than others. The ones that are especially noisy flicker more.

[Stears555]

But my problem is the following: The white points appear and disappear always exactly on the same place. (after three frames later. )

Some do, some don't. Some pixels are noisier than others. The ones that are especially noisy flicker more.

Than, how can we make difference between the white noise-points and the real dead pixels in the case of a video? What is exactly the difference? Are the real dead pixels permanent white/black instead of vibrating?

[jagabo]

Are the real dead pixels permanent white/black instead of vibrating?

Yes. Though, I suppose it depends on your definition of "dead". A pixel that only flickers black and white may not be "dead" but it's malfunctioning.

If you're looking to get your camera replaced under warranty, I doubt the manufacturer would consider this a problem.

[DB83]

Well I would define a dead pixel as permanent black since there is no hex value attached. White has a value so, technically, it is 'not dead'

Surely there must be a better way to check a camera's pixels. This method seems half-baked.

[Stears555]

Those pixels vibrating between white and the correct colors (tested with 0db gain).

[LMotlow]

What's the purpose of recording black video? Are you just trying to create a length of black lead-in or lead-out video segments? You can do that very easily in Avisynth or any number of other software apps without "recording" a dark scene.

All digital recording devices such as cameras have a residual noise level. Even in a fairly dim room, there should be enough signal information to make the residual noise level invisible. But if your problem is that white dots show up even when recording "normal" or daylight scenes, then your camera has a problem. Do the white dots show up in everything you record?

[jagabo]

What's the purpose of recording black video?

To test for stuck pixels, obviously. Then you record white video to test for pixels stuck at black.

[Cornucopia]

Take a minute long clip with the lens cap on (and NO gain). That'll give you your "dark noise" or "dark current reference. Running a median filter on all that, and then a tri-level threshold filter would exaggerate the differences to where you should see BLACK for dead pixels, GRAY for random noise & regular pixels, and WHITE for faulty HOT pixels. I guess there might be the possibility of faulty gray pixels, but you probably don't need to worry about those because unless they were vast in quantity you wouldn't notice them (too little contrast with surround ing pixels).

Scott

[Stears555]

Here is an other sample file, can you recognize the white spots on the video (with the help of a software?) http://diktafon.atw.hu/00110.MTS

[Stears555]

bump

[jagabo]

I can occasionally see the light dots in dark parts of the picture.