Velocity Reviews - Computer Hardware Reviews

Velocity Reviews > Newsgroups > Computing > Digital Photography > Dslr Noise

Reply
Thread Tools

Dslr Noise

 
 
johnJ
Guest
Posts: n/a
 
      06-07-2004
> Man, I've only been here two days or so and he already tops any troll in
> any of the other groups I subscribe to. I didn't think anyone could top
> Usenet Kook of the Current and Last Century: Mike Vandeman, Phd, of
> alt.mountain-bike fame.


Welcome to our hell...
 
Reply With Quote
 
 
 
 
David Kilpatrick
Guest
Posts: n/a
 
      06-07-2004


Randall Ainsworth wrote:

> In article <(E-Mail Removed)>, Jim Townsend
> <(E-Mail Removed)> wrote:
>
>
>>Phil Wheeler wrote:
>>
>>
>>
>>>I can see no diff between ISO 100 and ISO 400, none.

>>
>>I find the same with my 10D.. With images of ISO 400 and
>>under, I have to check the EXIF data to see what it was shot at.
>>You can't tell by looking. ISO 800 begins to show noticeable
>>noise.
>>
>>However.. When it comes to processing, it doesn't take
>>much contrast and sharpness enhancement to bring out the
>>grain. You have much more latitude with ISO 100 and 200.

>
>
> There's a difference between 100 and 400 on the 10D, but you really
> have to get some magnification to see it.


Most noise assessments are made viewing the image file at 100 per cent -
pixel for pixel. There's no benefit in enlarging pixels on the screen,
and at anything less, you can't judge noise (especially at 66.6 and 33.3
per cent views in Photoshop, where both noise and sharpening may appear
exaggerated while actual detail is reduced).

So what is 'some magnification', as opposed to the normal way of
examining the image?

David

 
Reply With Quote
 
 
 
 
bagal
Guest
Posts: n/a
 
      06-07-2004

"David Kilpatrick" <(E-Mail Removed)> wrote in message
news:ca1c3g$qbj$(E-Mail Removed)...
>

<delete key used extensively>
>
> So what is 'some magnification', as opposed to the normal way of
> examining the image?
>
> David
>


I think I may be able to explain David - with apologies to those who may
have read my earlier postings.

The question may be about JPEG and it is important to recall what JPEGs hope
to do, when the solution evolved what "upgrades" are required and so on.

JPEGs started to become very popular in the early 1990s. Say what?
Yep the early 1990s In computer terms that is a very long time ago and
quite a few topics in computer terms have last quite as long without some
revision or upgrade. JPEG2k is some attempt at a revision but it would
appear not to have been taken up.

It really shows how robust JPEG are to last so long.

Time frame
The early 1990s is pre-Pentium processor and probably pre-486 (when JPEG
started) and probably 286-386 CPUs were the predomoinant or most popular at
the time.

Memory - hmm I think £100 per MB at the time rapidly falling to £10 per MB

Plus - the aim behind JPEG is to provide a downloaded image file that is not
to gross in size or download times using 9,600 or 14,400 bps (honest m8 -
broadband was not even talked about) and provide an accurate image at (and
this is the important bit) the default setting.

If an image is 1024 by 768 and at 400% zoom shows artefacts this that and
the other the reply should be: well what do you expect? It is a JPEG
designed for display at 1024 by 768

It is good being aware of the limits of JPEGs but equally important to know
what they were originally designed to do

I hope this helps

das B


 
Reply With Quote
 
bagal
Guest
Posts: n/a
 
      06-07-2004
Hello Georgette

This may be going slightly off-topic - I know I shouldn't really do so
however I wanted to seek a view on the Foveon P&S and I wonder if you may
help?

The points are
1 - image burnout and associated dynamic range
2 - image analysis on non-pro monitors using JPEG images as the basis for
comparative tests
3 - temperature effects on sensors in general
4 - revamp of nomenclature

1 - Does the Foveon sensor have increased dynamic range (I undesratnd this
will make it less likley to experience highlight burnout)?

2 - It seems a shame that most image comparisons are based on JPEGs on
non-color tuned monitors. Will RAW output feature in the P&S using Foveon
sensors?

3 - Does the Foveon sensor output change with operating temperature (ambient
and/or local to the sensor)

4 - Does nomenclature, definitions and "words" used in digital cameras need
to be changed? For example, is it better to define sensel as a sensor
element, pixel output as a picture element position based on data from a
number of sensels, and so on ..

Ta

das B
>
> http://www.dpreview.com/reviews/sigmasd10/page13.asp
>
> Less noise from 10.3M sensors @ ISO 800 than 6M sensors @ ISO 100
> represents a real revolution in digital photography, annd it is all
> due to the Sigma having 25% higher sensor pitch even with its 170%
> higher sensor count.



 
Reply With Quote
 
Lionel
Guest
Posts: n/a
 
      06-07-2004
Kibo informs me that "bagal" <(E-Mail Removed)> stated that:

>Hello Georgette
>
>This may be going slightly off-topic - I know I shouldn't really do so
>however I wanted to seek a view on the Foveon P&S and I wonder if you may
>help?
>
>The points are
>1 - image burnout


This has much more to do with the user than the camera.

> and associated dynamic range


This is similar for most DSLRs. That said, you'll generally get better
dynamic range on any DSLR by shooting RAW, rather than JPEG, as most
image sensors have better dynamic range than the 8 bit limit in the JPEG
format.

>2 - image analysis on non-pro monitors using JPEG images as the basis for
>comparative tests


Bad idea. You need to at least calibrate your monitor. I do mine about
every fortnight.

>3 - temperature effects on sensors in general


Hot sensors tend to be noisier. In general, you'll get less noise in
cold environments & more noise in hot ones.

>4 - revamp of nomenclature
>
>1 - Does the Foveon sensor have increased dynamic range (I undesratnd this
>will make it less likley to experience highlight burnout)?


Again, this is something the user needs to get right. Some cameras are
better at rescuing burnt highlights than others, but you should be
exposing correctly to prevent them in the first place. That said, I've
found that I can usually rescue 1-2 stops of burnt highlights from a 10D
RAW file.

>2 - It seems a shame that most image comparisons are based on JPEGs on
>non-color tuned monitors.


I only process/compare on a calibrated monitor. (A La Cie Electron Blue
22, calibrated a couple of times a month.)

> Will RAW output feature in the P&S using Foveon
>sensors?


I don't know, but I doubt it, considering the market that they intend it
for.

>3 - Does the Foveon sensor output change with operating temperature (ambient
>and/or local to the sensor)


Um? If you're referring to noise, then 'yes' - same as any other silicon
image sensor.

>4 - Does nomenclature, definitions and "words" used in digital cameras need
>to be changed? For example, is it better to define sensel as a sensor
>element, pixel output as a picture element position based on data from a
>number of sensels, and so on ..


I doubt it, but marketing bunnies always like to screw around with
standard terminology to make their product sound better, so anything
that'll help nail the buggers down to the truth is a bonus, IMO.

--
W
. | ,. w , "Some people are alive only because
\|/ \|/ it is illegal to kill them." Perna condita delenda est
---^----^---------------------------------------------------------------
 
Reply With Quote
 
PTRAVEL
Guest
Posts: n/a
 
      06-07-2004

"Georgette Preddy" <(E-Mail Removed)> wrote in message
news:(E-Mail Removed) om...
> "PTRAVEL" <(E-Mail Removed)> wrote in message

news:<(E-Mail Removed)>...
> > "Georgette Preddy" <(E-Mail Removed)> wrote in message
> > news:(E-Mail Removed) om...
> >
> > <snip>
> >
> > As you can see,
> > > the Canon is not competitive on noise at ISO 100, in fact, the 1 layer
> > > 6M sensor Canon CMOS shows more noise at ISO 100 than the 3-layer
> > > 10.3M sensor Foveon CMOS at ISO 800...
> > >
> > > http://www.dpreview.com/reviews/sigmasd10/page13.asp

> >
> > Once again, Preddy doesn't actually read the articles that he cites.

>
> The noise samples speak for themselves, Canon is tremendously noisey
> at ISO 100 compared to the Sigma--just look at the unchangable factual
> samples.


No, Preddy -- the article speaks for itself, and the article concluded this:

What the grey patch doesn't show
However, it's not all rosy for the SD10, what the crops above don't show is
what happens to the color patches as ISO increases. The EOS 10D manages to
maintain both the hue and saturation of the distinct red, green and blue
patches all the way through to ISO 1600, at ISO 3200 all the patches are
affected by the high noise level, noticably the blue patch is turning
purple. The SD10 gradually loses both saturation and hue accuracy as
sensitivity is increased, up to ISO 400 things appear to be acceptable, at
ISO 800 it's borderline with the red patch taking quite a lot of yellow /
green, at ISO 1600 noise levels in the color patches is quite strong.





Why do you always lie, George?



>
> > If he
> > had, he would have read this:
> > "However, it's not all rosy for the SD10

>
> That was so Canon didn't put him out of business. Same reason Phil
> said he wanted to do a color resolution test but was unable to publish
> one--Canon would've aggressively pulled their dollars from his site
> forever, ending his business. He always throws a bone to
> Canon/Nikon/Fuji when he shows how a competitor eats their lunch.



 
Reply With Quote
 
David Kilpatrick
Guest
Posts: n/a
 
      06-07-2004


bagal wrote:

> Hello Georgette
>
> This may be going slightly off-topic - I know I shouldn't really do so
> however I wanted to seek a view on the Foveon P&S and I wonder if you may
> help?
>
> The points are
> 1 - image burnout and associated dynamic range
> 2 - image analysis on non-pro monitors using JPEG images as the basis for
> comparative tests
> 3 - temperature effects on sensors in general
> 4 - revamp of nomenclature
>
> 1 - Does the Foveon sensor have increased dynamic range (I undesratnd this
> will make it less likley to experience highlight burnout)?


I can only give one answer - when the SD10 review screen shows a
burned-out highlight (flagged up red), it does not mean you have lost
this detail. The raw file seems to have a vast amount of leeway, even
compared to the Kodak Pro/n file, and if you don't use the 'auto'
adjustments of Sigma software (which tend to push you towards small
areas of burn out), the histogram usually has plenty of space to the
right. You have to over expose by two stops to wreck the raw file and
end up with real burnout problems.

David


 
Reply With Quote
 
Randall Ainsworth
Guest
Posts: n/a
 
      06-07-2004
> The Simga SD9/SD10 are the only two DSLRs built that can read the date
> on a penny with a 1 yard wide FOV, the Canon 10D can't even see that a
> date field is present let alone read the digits.


OK, you allegedly take pictures of test patterns and pennies at a
distance. You must lead a full life.
 
Reply With Quote
 
tekfull@yahoo.com
Guest
Posts: n/a
 
      06-07-2004
show those great pics you sell Preddy


On 6 Jun 2004 21:23:55 -0700, http://www.velocityreviews.com/forums/(E-Mail Removed) (Georgette
Preddy) wrote:

>"PTRAVEL" <(E-Mail Removed)> wrote in message news:<(E-Mail Removed)>...
>> "Georgette Preddy" <(E-Mail Removed)> wrote in message
>> news:(E-Mail Removed) om...
>>
>> <snip>
>>
>> As you can see,
>> > the Canon is not competitive on noise at ISO 100, in fact, the 1 layer
>> > 6M sensor Canon CMOS shows more noise at ISO 100 than the 3-layer
>> > 10.3M sensor Foveon CMOS at ISO 800...
>> >
>> > http://www.dpreview.com/reviews/sigmasd10/page13.asp

>>
>> Once again, Preddy doesn't actually read the articles that he cites.

>
>The noise samples speak for themselves, Canon is tremendously noisey
>at ISO 100 compared to the Sigma--just look at the unchangable factual
>samples.
>
>> If he
>> had, he would have read this:
>> "However, it's not all rosy for the SD10

>
>That was so Canon didn't put him out of business. Same reason Phil
>said he wanted to do a color resolution test but was unable to publish
>one--Canon would've aggressively pulled their dollars from his site
>forever, ending his business. He always throws a bone to
>Canon/Nikon/Fuji when he shows how a competitor eats their lunch.


 
Reply With Quote
 
Bill Funk
Guest
Posts: n/a
 
      06-07-2004
On 6 Jun 2004 21:23:55 -0700, (E-Mail Removed) (Georgette
Preddy) wrote:

>"PTRAVEL" <(E-Mail Removed)> wrote in message news:<(E-Mail Removed)>...
>> "Georgette Preddy" <(E-Mail Removed)> wrote in message
>> news:(E-Mail Removed) om...
>>
>> <snip>
>>
>> As you can see,
>> > the Canon is not competitive on noise at ISO 100, in fact, the 1 layer
>> > 6M sensor Canon CMOS shows more noise at ISO 100 than the 3-layer
>> > 10.3M sensor Foveon CMOS at ISO 800...
>> >
>> > http://www.dpreview.com/reviews/sigmasd10/page13.asp

>>
>> Once again, Preddy doesn't actually read the articles that he cites.

>
>The noise samples speak for themselves, Canon is tremendously noisey
>at ISO 100 compared to the Sigma--just look at the unchangable factual
>samples.
>
>> If he
>> had, he would have read this:
>> "However, it's not all rosy for the SD10

>
>That was so Canon didn't put him out of business. Same reason Phil
>said he wanted to do a color resolution test but was unable to publish
>one--Canon would've aggressively pulled their dollars from his site
>forever, ending his business. He always throws a bone to
>Canon/Nikon/Fuji when he shows how a competitor eats their lunch.


Ah, yes, the "There's a vast conspiracy" school of thought.
It's a real shame that this is the only way you can come to terms with
reviews that don't favor the Sigmas.

--
Bill Funk
replace "g" with "a"
 
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
ISO noise vs. long exposure noise Cynicor Digital Photography 5 08-27-2005 06:14 PM
Non-noise words are incorrectly recognised as noise words. Peter Strĝiman ASP .Net 1 08-23-2005 01:26 PM
Noise about noise... Stacey Digital Photography 3 02-18-2005 05:57 AM
Noise Ninja custom noise print- worth the effort for stacked photo?? Jason Sommers Digital Photography 4 01-19-2005 06:54 AM
Canon 1Ds Mark II - Noise? What noise? Brian C. Baird Digital Photography 9 09-21-2004 09:54 PM



Advertisments