What's a dead pixel?

Archived from groups: rec.photo.digital (More info?)

"Newbie" <chromallly@yahoo.com> wrote in message
news:1114219705.159159.174970@z14g2000cwz.googlegroups.com...
> What's a dead pixel? What else do we test for on a newly bought
> digicam?

A dead pixel is simply a photosite on the sensor that doesn't generate any
data, and therefore shows up as a black spot...most noticable within smooth,
bright portions of an image. Most cameras nowadays have built-in algorythms
that map these out...so that you never know they're there.

The other thing to check for is the problem of "stuck pixels," which show up
as conssatntly bright...often of a particular color of the spectrum in all
images. The easiest way to check for this is to take an exposure with the
lens cap on the with a shot of a half second or so. Longer exposures will
usually show hot pixels of some sort on nearly ALL digital cameras, unless
they have a facility for mapping them out automatically.

Archived from groups: rec.photo.digital (More info?)

Not exactly. You forgot to distinguish between stuck pixel and hot pixel.

See http://www.nikon.com.sg/TechSupp/K [...] Random.htm for more.

Gregor

"Mark²" <mjmorgan(lowest even number here)@cox..net> wrote in message
news:7Ihae.3087$Zi.1597@fed1read04...
>
> "Newbie" <chromallly@yahoo.com> wrote in message
> news:1114219705.159159.174970@z14g2000cwz.googlegroups.com...
>> What's a dead pixel? What else do we test for on a newly bought
>> digicam?
>
> A dead pixel is simply a photosite on the sensor that doesn't generate any
> data, and therefore shows up as a black spot...most noticable within
> smooth, bright portions of an image. Most cameras nowadays have built-in
> algorythms that map these out...so that you never know they're there.
>
> The other thing to check for is the problem of "stuck pixels," which show
> up as conssatntly bright...often of a particular color of the spectrum in
> all images. The easiest way to check for this is to take an exposure with
> the lens cap on the with a shot of a half second or so. Longer exposures
> will usually show hot pixels of some sort on nearly ALL digital cameras,
> unless they have a facility for mapping them out automatically.
>
>
>

Archived from groups: rec.photo.digital (More info?)

Mark² wrote:

>
> "Newbie" <chromallly@yahoo.com> wrote in message
> news:1114219705.159159.174970@z14g2000cwz.googlegroups.com...
>> What's a dead pixel? What else do we test for on a newly bought
>> digicam?
>
> A dead pixel is simply a photosite on the sensor that doesn't generate any
> data, and therefore shows up as a black spot...most noticable within
> smooth,
> bright portions of an image. Most cameras nowadays have built-in
> algorythms that map these out...so that you never know they're there.
>

By "built in", you mean most have to be sent in for pixel mapping as very
few have their own mapping system onboard.
--

Stacey

Archived from groups: rec.photo.digital (More info?)

Mark² wrote:

>
> "Stacey" <fotocord@yahoo.com> wrote in message
>>
>> By "built in", you mean most have to be sent in for pixel mapping as very
>> few have their own mapping system onboard.
>
> My understanding is that many have this built in as we speak, but it's not
> widely publicized--since manufacturers have nothing to gain by telling
> folks that tiny portions of their images are not really generated by the
> shot scene...

> I will be happy to shift this understanding if shown otherwise.

I've read canon dSLR owners have to send theirs in for "pixel mapping" so
don't think most do this automatically. I know on my E300 is has a menu
item for the camera to do this and it takes a while for it to run though
this procedure. I don't see how they could do this "on the sly" so people
wouldn't know? That and most people do consider this a positive feature for
a camera to have so it doesn't seem likely they would hide that their
camera has this feature.

--

Stacey

Archived from groups: rec.photo.digital (More info?)

Stacey <fotocord@yahoo.com> wrote:
: Mark? wrote:

: >
: > "Stacey" <fotocord@yahoo.com> wrote in message
: >>
: >> By "built in", you mean most have to be sent in for pixel mapping as
: >> very few have their own mapping system onboard.
: >
: > My understanding is that many have this built in as we speak, but
: > it's not widely publicized--since manufacturers have nothing to gain
: > by telling folks that tiny portions of their images are not really
: > generated by the shot scene...

: > I will be happy to shift this understanding if shown otherwise.

: I've read canon dSLR owners have to send theirs in for "pixel mapping" so
: don't think most do this automatically. I know on my E300 is has a menu
: item for the camera to do this and it takes a while for it to run though
: this procedure. I don't see how they could do this "on the sly" so people
: wouldn't know? That and most people do consider this a positive feature for
: a camera to have so it doesn't seem likely they would hide that their
: camera has this feature.

Maybe the term is being shared by more than one procedure, or some camera
makers have multiple levels of "mapping". But to clarify the need for
mapping (as I understand it), EVERY CCD device has a few pixels that are
not responding correctly (dead, always on, weak, etc). Manufacturing any
semiconductor device involves a certain percentage of failure of
individual components on the chip. The more complex, or more densly packed
the parts the more likely that a failure will impact the circuit
adversely. If all CCDs were rejected if even a single pixel failed, the
cost of the chips would be astronomical! So there is an "acceptable
percentage" of failures for each chip manufacturer and proposed use.

Now, why map? When a chip is about to be installed in a camera a test of
each pixel is run and rated for performance. Then a custom "compensation
function" is generated and burned into the computer in the camera to
minimize the effects. On the cheapest cameras this is probably where the
whole thing stops. With slightly more expensive cameras there may be some
kind of automatic monitoring of ongoing function of the camera which will
further adjust compensation over time. Then for some high end cameras
(which could demand much higher quality) a periodic factory mapping, under
lab conditions (which has to be much more precise than an on the fly
test) and recompensation would be offered.

The compensation may take many forms. It could be as simple as copying one
of the adjacent pixels, or an averaging of the surrounding pixels, all the
way up to a complex formula calculating the most likely value of the
missing pixel. Unless you are taking lots of images that require
contrasting details as small as a single pixel, most "compensated" pixels
will never be noticed by MOST users. (notice I am specifically saying
"most" as there will always be those here and elsewhere that will not find
ANY variation "acceptable".)

Such a compensation is not uncommon in computers. Have you ever looked at
the "details" while doing a HD defrag? Some of the "locked" sectors are
simply sections of the disk that are not functioning properly and they are
locked out of future use. A majority of these "bad sectors" are found and
locked at the time of manufacture. And over time, wear may cause the
computer to lock out more sectors as they become damaged (more prevalent
in computers that are moved around or jostled a lot). So the same process
in CCDs is not suprizing and not normally spelled out in literature. It
would be like making a "major selling point" on a new brand of car depend
on the fact that it has "a device for lifting the car for replacement of a
flat tire" as standard.

So to sumarize this (overly) long posting, some form of pixel mapping is
likely to be done at the time of manufacture and/or be done on an ongoing
basis within the camera software. But a "lab quality" mapping done at the
manufacturer may be offered for those who require a higher level of
precision. The former would be "built in" and the latter would be "an
optional service".

Does this help?

Randy

==========
Randy Berbaum
Champaign, IL

Archived from groups: rec.photo.digital (More info?)

In article <d4d289$vfn$1@wildfire.prairienet.org>,
Randy Berbaum <rberbaum@bluestem.prairienet.org> wrote:
>Such a compensation is not uncommon in computers.

>A majority of these "bad sectors" are found and
>locked at the time of manufacture. And over time, wear may cause the
>computer to lock out more sectors as they become damaged (more prevalent
>in computers that are moved around or jostled a lot). So the same process
>in CCDs is not suprizing and not normally spelled out in literature.

The difference is that serious disk manufacturers document how to access
bad sector lists (for example the SCSI mode pages). A bad sector does
not affect the performance of the disk much if it is allocated on
the same track as it was located originally.

On the other hand, bad pixels have (potentially) much more impact than
bad sectors, but camera manufactures treat their professional customers
as simple consumers who are allowed to play with their toys and don't
need any technical documentation.


--
That was it. Done. The faulty Monk was turned out into the desert where it
could believe what it liked, including the idea that it had been hard done
by. It was allowed to keep its horse, since horses were so cheap to make.
-- Douglas Adams in Dirk Gently's Holistic Detective Agency

Archived from groups: rec.photo.digital (More info?)

Stacey wrote:
> Mark² wrote:
>
>
>>"Stacey" <fotocord@yahoo.com> wrote in message
>>
>>>By "built in", you mean most have to be sent in for pixel mapping as very
>>>few have their own mapping system onboard.
>>
>>My understanding is that many have this built in as we speak, but it's not
>>widely publicized--since manufacturers have nothing to gain by telling
>>folks that tiny portions of their images are not really generated by the
>>shot scene...
>
>
>>I will be happy to shift this understanding if shown otherwise.
>
>
> I've read canon dSLR owners have to send theirs in for "pixel mapping" so
> don't think most do this automatically. I know on my E300 is has a menu
> item for the camera to do this and it takes a while for it to run though
> this procedure. I don't see how they could do this "on the sly" so people
> wouldn't know? That and most people do consider this a positive feature for
> a camera to have so it doesn't seem likely they would hide that their
> camera has this feature.
>

It may be that the procedure is done at the factory, and stored in
non-volatile memory.
I doubt that even the most particular among us would be able to find 4
or 5 dead pixels among 4 to 8 million on the sensor since the firmware
would disguise them by averaging the pixels adjacent to them. This is
somewhat analogous to the mapping of bad sectors on hard drives,
although with hard drives, this process doesn't result in loss of space.



--
Ron Hunter rphunter@charter.net

Archived from groups: rec.photo.digital (More info?)

In message <Dhnae.3131$Zi.1498@fed1read04>,
"Mark²" <mjmorgan(lowest even number here)@cox..net> wrote:

>As I say...I may be wrong here.

The Canon 10D and 20D have an internal list of bad pixels, as tested at
the factory. The RAW data interpolates these pixels (and I assume that
they are interpolated from the closest pixels with the same color
filter). There is no facility for the user to update this list, but
using DCRAW you can give it a list of newly bad pixels, and it is
trivial to write an app that does the same to an intermediate .DNG file
if you want; if you convert to an uncompressed .DNG, the data after the
header is straight-forward 12-bit RAW data which you can overwrite
before submitting it to ACR or RSE, or any converter that supports .DNG.
--

<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<>
John P Sheehy <JPS@no.komm>
><<> <>>< <>>< ><<> <>>< ><<> ><<> <>><

Archived from groups: rec.photo.digital (More info?)

In message <d4df7l$hsb$03$1@news.t-online.com>,
Michael Schnell <mschnell_at_bschnell_dot_de@aol.com> wrote:

>That brings up the question how missing pixels are handled with RAW
>files. Does the camera provide a missing pixel list within the file or
>does it try (much less appropriately) to replace the missing pixels by
>means of the neighbors of the same color in the Bayer scheme ?

It couldn't possibly use the neighbors with different color pixels; that
would cause pixels that stand out, defeating the purpose. The only time
the levels in the three color channels are roughly the same is when a
subject is cyan or cyan-green, with most CFAs.
--

<>>< ><<> ><<> <>>< ><<> <>>< <>>< ><<>
John P Sheehy <JPS@no.komm>
><<> <>>< <>>< ><<> <>>< ><<> ><<> <>><