Chroma Noise - Not Banding

Why does everyone call Chroma Noise - Banding?

This is real banding:



Did the industry change the name of Chroma Noise? or is this just another example of internet forum mis-education.

But the Quotes (below) are what we used to call "Chroma Noise", but everyone calls it Banding on internet forums.

Am I missing something?

Member said:

willyhill said:

SPy said:

Wanted to compare shadow noise between 1DsII (so called old tech), and 5d2 due to all the commotion on these forums.

All settings are same, at 200ISO, 100% crops, under household tungsten lighting. Both photo had exposure boosted by 1 stop and processed from RAW in DPP. (Just noticed I need to set the clock in 1DsII hence time difference).

5d2...





and 1DsII...





I'm liking the 1DsII !

--
--
SPy

Click to expand...

Click to expand...

Click to expand...

Member said:

--

Click to expand...

My interpretation of bandings is lines of different colour and what you showed is posterization.....chroma noise being speckles of noise of different colour.....

--
Slainte (cheers)

You cannot depend on your eyes when your imagination is out of focus. :- Samuel Langhorne Clemens

You are just right.

The Low-ISO Banding is artifact of a pattern in the noise of the sensor, analog amplification and the analog to digital conversion. To some extend it is caused due to the 14-bits we now have.

Cheers

Thomas

Your sample shows posterization , which is due to limited bit depth.

If you convert an image to any color space that has a luminance channel and two color channels, chroma noise is quasi-random noise in the two color channels (as opposed to luminance noise).

Banding is straight horizontal or straight vertical noise variations in the image. The samples that I've seen are most pronounced at very high ISO.

Regards, Bill

Stephen Christie said:

But the Quotes (below) are what we used to call "Chroma Noise", but everyone calls it Banding on internet forums.

Click to expand...

When your noise appears as bands in the image, what do you call it? Banding has been a term for horizontal or vertical lines of noise since the first CCD sensors.

All it means is that there are visible bands of something in the image. Your example shows bands caused by posterization. There can also be bands caused by regular pattern noise that is constant from frame to frame. Another kind of banding forms from electronic interference from cell phones or other electromagnetic devices. Banding can be caused by an interference between a regular pattern in the subject and on the sensor, causing moire.

The banding recently being discussed regarding the 5D Mark II is different from all of these, but it is banding nevertheless. I don't know if anyone has identified the particular source of the problem, hence, it makes sense to use a generic term like banding to describe it.

Since the bands move around from frame to frame, or dance around in video mode, I'm inclined to think the problem is caused by electronic interference of some sort, probably from the camera's own processing. I recall the 10D had a problem, for example, when an image was being processed through the A/D step while AI Servo was engaged. The electronics being used to drive the lens focus interfered with the electronics reading the image from the sensor. The result was banding in the image just taken. Bursts of images shot with AI Servo showed the banding. Single shots where the shutter was released after taking the shot did not. I suspect something similar is happening with the 5D Mark II.
--

P.S. I wouldn't call it chroma noise, which, normally, is random in nature, and thus not arranged in bands.
--

Right, and the 2nd set of images I posted show random non patterned noise, which everyone calls bandnig for some reason.
Am I blind I dont see any bands those 2nd images.

The first image I posted is an exageration of what banding might look like, having a visable line that stands apart in color/shade from the other colors in the image, not dithered properly.

Stephen Christie said:

This is real banding:



Did the industry change the name of Chroma Noise? or is this just another example of internet forum mis-education.

But the Quotes (below) are what we used to call "Chroma Noise", but everyone calls it Banding on internet forums.

Am I missing something?

Stephen Christie said:

willyhill said:

SPy said:

Wanted to compare shadow noise between 1DsII (so called old tech), and 5d2 due to all the commotion on these forums.

All settings are same, at 200ISO, 100% crops, under household tungsten lighting. Both photo had exposure boosted by 1 stop and processed from RAW in DPP. (Just noticed I need to set the clock in 1DsII hence time difference).

5d2...





and 1DsII...





I'm liking the 1DsII !

--
--
SPy

Click to expand...

Click to expand...

Click to expand...

Stephen Christie said:

--

Click to expand...

Click to expand...

--
-Steve

Stephen Christie said:

Right, and the 2nd set of images I posted show random non patterned noise, which everyone calls bandnig for some reason.
Am I blind I dont see any bands those 2nd images.

Click to expand...

Apparently, yes. There is clear horizontal and vertical banding in both the 5D II and 1DS II images, but it is more pronounced in the 5D II image, which I think was the point of posting them. If you can't see the banding, perhaps your monitor needs calibration or the lights in your viewing room are too bright.
--

Well, I'll give up then, since I'm blind.

No matter what the ambient light in my room I cant see streaks or lines of bands on the images, My monitor is calibrated, and I have even tried viewing on my laptop and 2nd computer, that have different screens that are also calibrated using a new Spyder 3 Elite.

Victor Engel said:

Stephen Christie said:

Right, and the 2nd set of images I posted show random non patterned noise, which everyone calls bandnig for some reason.
Am I blind I dont see any bands those 2nd images.

Click to expand...

Apparently, yes. There is clear horizontal and vertical banding in both the 5D II and 1DS II images, but it is more pronounced in the 5D II image, which I think was the point of posting them. If you can't see the banding, perhaps your monitor needs calibration or the lights in your viewing room are too bright.
--

Click to expand...

--
-Steve

Stephen Christie said:

Right, and the 2nd set of images I posted show random non patterned noise, which everyone calls bandnig for some reason.

Click to expand...

That 5D2 photo shows banding , which is noise in a grid-like pattern - not random noise. The 1Ds2 photo has random noise - not in a grid-like pattern.

RedFox88 said:

Stephen Christie said:

Right, and the 2nd set of images I posted show random non patterned noise, which everyone calls bandnig for some reason.

Click to expand...

That 5D2 photo shows banding , which is noise in a grid-like pattern - not random noise. The 1Ds2 photo has random noise - not in a grid-like pattern.

Click to expand...

What he calls banding, THAT is actually banding. When you start to see bands of colour gradations.

First with the Nikon D200, the term "banding" was misused, describing odd striped patterns near bright objects at certain ISO settings.

Then later, people also began to use the term "banding" for patterns in noise. That people started to call that "banding" does not mean it is the correct use of the term.

Banding in (digital) photos/images still means bands of colour gradations.

1st of all, yes, the real meaning of banding in images is colour bands in gradations, like you showed as example.

Lately, people use that term (rather annoyingly) for MANY different artifacts that in some way or another can be considered as some kind of "band".

The "banding" in regards of noise gets used when you see a repeating horizontal or vertical "pattern" in the noise grid, as if the image is printed with printer with a misaligned head. You start to see stripes in the image.

While it visually does NOT look like actual banding, the term still gets used with it.

It would be better to call it vertical or horizontal noise pattern banding, to make a distinction with actual banding.

Stephen Christie said:

No matter what the ambient light in my room I cant see streaks or lines of bands on the images, My monitor is calibrated, and I have even tried viewing on my laptop and 2nd computer, that have different screens that are also calibrated using a new Spyder 3 Elite.

Victor Engel said:

Stephen Christie said:

Right, and the 2nd set of images I posted show random non patterned noise, which everyone calls bandnig for some reason.
Am I blind I dont see any bands those 2nd images.

Click to expand...

Apparently, yes. There is clear horizontal and vertical banding in both the 5D II and 1DS II images, but it is more pronounced in the 5D II image, which I think was the point of posting them. If you can't see the banding, perhaps your monitor needs calibration or the lights in your viewing room are too bright.
--

Click to expand...

--
-Steve

Click to expand...

Ok, I get it now, noise grids, and not gradient bands.

I see that it is a very faint GRIG like pattern to some of the noise, and I noticed it before, but there are also larger splotches of reddish color that are clouding the obviousness of the pattern to my eyes.

Thanks for helping me understand this phenom, and thanks for confirming that many people do actually misuse it.
Now I know what to look for (noise grids) I can make an educated conclusions.

Steve

brightcolours said:

1st of all, yes, the real meaning of banding in images is color bands in gradations, like you showed as example.

Lately, people use that term (rather annoyingly) for MANY different artifacts that in some way or another can be considered as some kind of "band".

The "banding" in regards of noise gets used when you see a repeating horizontal or vertical "pattern" in the noise grid, as if the image is printed with printer with a misaligned head. You start to see stripes in the image.

While it visually does NOT look like actual banding, the term still gets used with it.

It would be better to call it vertical or horizontal noise pattern banding, to make a distinction with actual banding.

Stephen Christie said:

No matter what the ambient light in my room I can’t see streaks or lines of bands on the images, My monitor is calibrated, and I have even tried viewing on my laptop and 2nd computer, that have different screens that are also calibrated using a new Spyder 3 Elite.

Victor Engel said:

Stephen Christie said:

Right, and the 2nd set of images I posted show random non-patterned noise, which everyone calls banding for some reason.
Am I blind I don’t see any bands those 2nd images.

Click to expand...

Apparently, yes. There is clear horizontal and vertical banding in both the 5D II and 1DS II images, but it is more pronounced in the 5D II image, which I think was the point of posting them. If you can't see the banding, perhaps your monitor needs calibration or the lights in your viewing room are too bright.
--

Click to expand...

--
-Steve

Click to expand...

Click to expand...

brightcolours said:

RedFox88 said:

Stephen Christie said:

Right, and the 2nd set of images I posted show random non-patterned noise, which everyone calls banding for some reason.

Click to expand...

That 5D2 photo shows banding , which is noise in a grid-like pattern - not random noise. The 1Ds2 photo has random noise - not in a grid-like pattern.

Click to expand...

What he calls banding, THAT is actually banding. When you start to see bands of color gradations.

First with the Nikon D200, the term "banding" was misused, describing odd striped patterns near bright objects at certain ISO settings.

Then later, people also began to use the term "banding" for patterns in noise. That people started to call that "banding" does not mean it is the correct use of the term.

Banding in (digital) photos/images still means bands of color gradations.

Click to expand...

brightcolours said:

Then later, people also began to use the term "banding" for patterns in noise. That people started to call that "banding" does not mean it is the correct use of the term.

Click to expand...

What then is the correct term to describe bands of noise in images?

brightcolours said:

It would be better to call it vertical or horizontal noise pattern banding, to make a distinction with actual banding.

Click to expand...

The problem with that is that it is not pattern banding like we normally think of it, where the pattern is consistent, like on a printer head, to continue your analogy. Furthermore, when an accurate, technical term like posterization is available to describe that phenomenon, I think it's a bad idea to try to use a generic term like banding in its place. The term posterization has been around for a very long time and works just fine.
--

Scott Larson said:

What then is the correct term to describe bands of noise in images?

Click to expand...

Stripings, apparently?

--

technically what you are showing is called posterization, although it has often, commonly, been called banding as well.

the banding mentioned say in the 5D2 threads isn't best called chroma noise at all,

it's a patterned type of noise with a lot of spatial coherence along lines- often called pattern banding or just banding which seems fine enough to me

Stephen Christie said:

This is real banding:



Did the industry change the name of Chroma Noise? or is this just another example of internet forum mis-education.

But the Quotes (below) are what we used to call "Chroma Noise", but everyone calls it Banding on internet forums.

Am I missing something?

Stephen Christie said:

willyhill said:

SPy said:

Wanted to compare shadow noise between 1DsII (so called old tech), and 5d2 due to all the commotion on these forums.

All settings are same, at 200ISO, 100% crops, under household tungsten lighting. Both photo had exposure boosted by 1 stop and processed from RAW in DPP. (Just noticed I need to set the clock in 1DsII hence time difference).

5d2...





and 1DsII...





I'm liking the 1DsII !

--
--
SPy

Click to expand...

Click to expand...

Click to expand...

Stephen Christie said:

--

Click to expand...

Click to expand...

brightcolours said:

RedFox88 said:

Stephen Christie said:

Right, and the 2nd set of images I posted show random non patterned noise, which everyone calls bandnig for some reason.

Click to expand...

That 5D2 photo shows banding , which is noise in a grid-like pattern - not random noise. The 1Ds2 photo has random noise - not in a grid-like pattern.

Click to expand...

What he calls banding, THAT is actually banding. When you start to see bands of colour gradations.

First with the Nikon D200, the term "banding" was misused, describing odd striped patterns near bright objects at certain ISO settings.

Then later, people also began to use the term "banding" for patterns in noise. That people started to call that "banding" does not mean it is the correct use of the term.

Banding in (digital) photos/images still means bands of colour gradations.

Click to expand...

actually people called that stuff posterization and then later some started casually referring to that as banding

i don't see what the big fuss is either way

Stephen Christie said:

Why does everyone call Chroma Noise - Banding?

This is real banding:

Click to expand...

Actually, that is "contour" banding.

Stephen Christie said:

Did the industry change the name of Chroma Noise? or is this just another example of internet forum mis-education.

Click to expand...

It's not just chroma in effect. It is patterned, in lines or bands.

Stephen Christie said:

But the Quotes (below) are what we used to call "Chroma Noise", but everyone calls it Banding on internet forums.

Am I missing something?

Click to expand...

The phenomenon could have had a better name, like line noise or 1-dimensional noise, but really, contour banding should never drop the "contour" modifier.

--
John