Velocity Reviews - Computer Hardware Reviews

Velocity Reviews > Newsgroups > Computing > Digital Photography > luminance noise and chroma noise: What is it?

Reply
Thread Tools

luminance noise and chroma noise: What is it?

 
 
Marc Wossner
Guest
Posts: n/a
 
      07-03-2007

Hi ng,

Trying to understand noise in digital photography I read about
"luminance noise" and "chroma noise" without getting what those
concepts really mean. - Are they the peculiarities of some more
general noise phenomena (like dark current, quantization or photon
noise) in those two information channels or are they in fact noise
phenomena on its own and if so, what causes them?

Thanks for your input!
Marc Wossner

 
Reply With Quote
 
 
 
 
John Sheehy
Guest
Posts: n/a
 
      07-03-2007
Marc Wossner <(E-Mail Removed)> wrote in news:1183502269.277259.69850
@k79g2000hse.googlegroups.com:

> Trying to understand noise in digital photography I read about
> "luminance noise" and "chroma noise" without getting what those
> concepts really mean. - Are they the peculiarities of some more
> general noise phenomena (like dark current, quantization or photon
> noise) in those two information channels or are they in fact noise
> phenomena on its own and if so, what causes them?


They're just ways of interpreting RAW data. Most digital cameras use a
sensor covered with little red, green, and blue filters; 1/4 covered with
red, 1/2 covered with green, and 1/4 covered with blue. How chromatic the
noise is, in the final image, depends on how the RAW converter interprets
the pixel-level data; it's a matter of style. Generally, manufacturers
with lots of noise in their cameras are more prone to use an interpretation
that filters out the noise in the virtual chroma channel, leaving the noise
to appear to only affect luminance.

--

<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<>
John P Sheehy <(E-Mail Removed)>
><<> <>>< <>>< ><<> <>>< ><<> ><<> <>><

 
Reply With Quote
 
 
 
 
acl
Guest
Posts: n/a
 
      07-04-2007
On Jul 4, 3:50 am, John Sheehy <(E-Mail Removed)> wrote:
> Marc Wossner <(E-Mail Removed)> wrote in news:1183502269.277259.69850
> @k79g2000hse.googlegroups.com:
>
> > Trying to understand noise in digital photography I read about
> > "luminance noise" and "chroma noise" without getting what those
> > concepts really mean. - Are they the peculiarities of some more
> > general noise phenomena (like dark current, quantization or photon
> > noise) in those two information channels or are they in fact noise
> > phenomena on its own and if so, what causes them?

>
> They're just ways of interpreting RAW data. Most digital cameras use a
> sensor covered with little red, green, and blue filters; 1/4 covered with
> red, 1/2 covered with green, and 1/4 covered with blue. How chromatic the
> noise is, in the final image, depends on how the RAW converter interprets
> the pixel-level data; it's a matter of style. Generally, manufacturers
> with lots of noise in their cameras are more prone to use an interpretation
> that filters out the noise in the virtual chroma channel, leaving the noise
> to appear to only affect luminance.
>


even if all 3 channels are sampled at each location, you'll still get
chroma noise. This is because noise is uncorrelated between the R, G
and B "sensors". But what you say is indeed true.

So, to answer the OP's question: in a bayer-sensored camera, each
pixel actually samples one of R, G or B colours, and the actual colour
is interpolated using neighbouring pixels. Since there is noise, and
this noise is not correlated between pixels (ie if the fluctuation is
"up" in "this" pixel, it says nothing about the neighbouring pixel,
where it could be "down"), we get fluctuations in colour, in addition
to fluctuations in intensity of received light. What it ends up as
depends on the raw conversion.

This is probably extremely confusing as stated! If you (ie the OP)
don't understand, do ask, it's probably the fault of a bad explanation
rather than you.

 
Reply With Quote
 
Marc Wossner
Guest
Posts: n/a
 
      07-04-2007
<big snip>

> This is probably extremely confusing as stated! If you (ie the OP)
> don't understand, do ask, it's probably the fault of a bad explanation
> rather than you.


No, both of you gave very good explanations and I do understand now
where the two types of noise come from.
Thanks!

Marc Wossner

 
Reply With Quote
 
Gautam Majumdar
Guest
Posts: n/a
 
      07-05-2007
On Tue, 03 Jul 2007 23:37:49 +0100, Marc Wossner wrote:
> Trying to understand noise in digital photography I read about
> "luminance noise" and "chroma noise" without getting what those concepts
> really mean. - Are they the peculiarities of some more general noise
> phenomena (like dark current, quantization or photon noise) in those two
> information channels or are they in fact noise phenomena on its own and
> if so, what causes them?
>

My understanding of these two noises is rather simplistic. The luminance
noise is when pixels show different brightness but are of the same colour.
An example is that of a grainy looking sky in an high ISO image. All the
pixels are blue but with different luminosity. Chroma noise on the other
hand is when pixels show different coulours when they all should be of the
same colour. One example is that of an underexposed image brightened up
during pp. Quite often both types of noises are present simultaneously.


--
gautam
 
Reply With Quote
 
Don Stauffer in Minnesota
Guest
Posts: n/a
 
      07-05-2007
On Jul 5, 1:20 am, Gautam Majumdar <(E-Mail Removed)> wrote:
> On Tue, 03 Jul 2007 23:37:49 +0100, Marc Wossner wrote:
> > Trying to understand noise in digital photography I read about
> > "luminance noise" and "chroma noise" without getting what those concepts
> > really mean. - Are they the peculiarities of some more general noise
> > phenomena (like dark current, quantization or photon noise) in those two
> > information channels or are they in fact noise phenomena on its own and
> > if so, what causes them?

>
> My understanding of these two noises is rather simplistic. The luminance
> noise is when pixels show different brightness but are of the same colour.
> An example is that of a grainy looking sky in an high ISO image. All the
> pixels are blue but with different luminosity. Chroma noise on the other
> hand is when pixels show different coulours when they all should be of the
> same colour. One example is that of an underexposed image brightened up
> during pp. Quite often both types of noises are present simultaneously.
>
> --
> gautam


I would say that luminance noise is a fundamental noise. The noise in
the pixels themselves is essentially a luminence value, because the
pixels are responding to the real or apparent fluctuations in the
arrival of photons.

In a focal plane with a Bayer or similar array of color filters,
chroma noise is an apparent noise due to the fluctuations in pixel
response altering the ratio between various colors. The CCD itself
knows nothing about "color".

 
Reply With Quote
 
acl
Guest
Posts: n/a
 
      07-05-2007
On Jul 5, 3:03 pm, Don Stauffer in Minnesota <(E-Mail Removed)>
wrote:

>
> I would say that luminance noise is a fundamental noise. The noise in
> the pixels themselves is essentially a luminence value, because the
> pixels are responding to the real or apparent fluctuations in the
> arrival of photons.
>
> In a focal plane with a Bayer or similar array of color filters,
> chroma noise is an apparent noise due to the fluctuations in pixel
> response altering the ratio between various colors. The CCD itself
> knows nothing about "color".


Even if there is no bayer sensor, you'll get chrominance noise. If you
have 1 "sensor" for each colour at each pixel, the random variations
in intensity will not be correlated between them, thus you'll get
colour variations on top of the luminance variations. The source of
both would be, as you say, randomness in the luminance at each
photosite/sensor (but since each measures number of photons within a
single passband, this ends up contributing to the luminance noise
too).

If, instead, you could directly measure the wavelength of each photon,
you'd still get both luminance and chrominance noise: luminance for
the same reasons as now (eg shot noise, read noise contributing to
variations in the number of photoelectrons etc; that is, variation in
the number of photons detected), and you'd certainly get noise in the
measurement of the wavelength (since you're essentially measuring
energy when you measure wavelength). I have no idea of the magnitude
of this noise in the wavelength measurement; it would depend on how
you measure it, for example. Maybe it could be made inconsequential,
maybe not; I haven't thought about it and can't be bothered to find
out what has already been done in this area to get an idea. Anyway, in
this case, chrominance and luminance noise would have different
fundamental causes.

Of course, what you say is true for actual existing cameras

 
Reply With Quote
 
 
 
Reply

Thread Tools

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are On
Pingbacks are On
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Recommended monitor luminance levels? Kulvinder Singh Matharu Digital Photography 39 03-13-2008 06:43 PM
Chroma Key Background Image? Monk Digital Photography 4 04-10-2005 07:04 PM
Illumination Needed On Luminance! secheese Digital Photography 9 01-27-2005 10:12 PM
Does Sony DVP-NC875V/B have the chroma bug?? Dan Williams DVD Video 0 11-23-2004 05:56 AM
Unsharp mask on luminance channel in lab mode? no Digital Photography 3 06-17-2004 02:30 AM



Advertisments