Can you upload some photos taken?

Note that phone default to 3.5M so if you want higher quality, you need to
manually up it to 5M

Comment 2 Andre Klapper 2010-01-18 12:52:11 UTC

"poor" is always highly subjective.

Comment 3 Neil MacLeod 2010-01-18 13:07:01 UTC

The N900vsN85[1] picture looks like it might be suffering from glare off the
chrome trim surrounding the flash and lens.

See thread on t.m.o[2] for more discussion on this with similar photos.

The only solution is to cover up or remove the chrome trim. Nokia may want to
consider a design change for future production.

Hopefully Mugen Power won't make the same mistake with their forthcoming
extended battery which will be supplied with a new back[3] as it is physically
larger than the current battery.

1.
http://lh4.ggpht.com/_lfOpSNw6yAA/S1Oz2A_tivE/AAAAAAAABgM/4gs6RG8W7YU/s160-c/NokiaN900VsN85.jpg
2. http://talk.maemo.org/showthread.php?t=36675
3.
http://www.nokiausers.net/General/Mugen-Power-Extended-Batteries-for-the-N900-are-Coming.html

(In reply to comment #2)
> "poor" is always highly subjective.
> 
If one reads the submission completely, one will find there is EMPIRICAL
EVIDENCE provided so as to move the description of "poor" from subjective to
objective.  There is an URL attached to the submission where you can find
photos taken at the same time of the same subject in the same place with the
same settings (default automatic @ 5MP) with the N900 vs a N85 8MP.
Here is the URL once again:
http://picasaweb.google.com/drew.freyman/NokiaN900VsN85?feat=directlink

(In reply to comment #1)
> Can you upload some photos taken?
> 
> Note that phone default to 3.5M so if you want higher quality, you need to
> manually up it to 5M
> 

You will note in my submission that I indicate the following:
1) both cameras were set at 5MP resolution and all other settings default
2) there is a URL where one can see a comparison pics side-by-side.  Once
again, here is that URL: 
http://picasaweb.google.com/drew.freyman/NokiaN900VsN85?feat=directlink

(In reply to comment #1)
> Can you upload some photos taken?
> 
> Note that phone default to 3.5M so if you want higher quality, you need to
> manually up it to 5M
> 

At the top of the bug submission I note the following:
5.0/(2.2009.51-1)
default camera settings with 5MP resolution.

This means all camera settings were the default settings with the exception of
the 5MP resolution setting (for both cameras).

(In reply to comment #3)
> The N900vsN85[1] picture looks like it might be suffering from glare off the
> chrome trim surrounding the flash and lens.
> 
> See thread on t.m.o[2] for more discussion on this with similar photos.
> 
> The only solution is to cover up or remove the chrome trim. Nokia may want to
> consider a design change for future production.

> 2. http://talk.maemo.org/showthread.php?t=36675

Thanks for looking at the comparison pix I put on Picasa! 

Christexaport's assessment that the flash reflection off the chrome trim is
contributing significantly to the problem is spot on.

However, I do believe there is more than that going on.  The symptom is
exhibited best by the two photos taken with the wall nearest the left field of
the photos. In the N900 photo the reflection of the flash has an inordinate
negative effect on the photo.  Compared to the same N85 photo, in which the
flash reflection off the wall also has an negative effect but less so, the
degradation of photo quality is not proportional.  That is, those photos seem
to indicate that the N85 is able to deal with the reflection of the flash off
the wall surface much better than the N900.

I believe, even if the silver trim were removed, we would find that the N900 is
less able to deal with light reflection situations like the one.

Comment 8 Andre Klapper 2010-01-19 21:09:32 UTC

(In reply to comment #3)
> The only solution is to cover up or remove the chrome trim. Nokia may want to
> consider a design change for future production.

*If* you think that this can be improved by hardware changes then please file a
ticket in http://maemo.org/community/brainstorm/ as bugs.maemo.org is for
handling software errors only.

Comment 9 Andre Klapper 2010-01-22 21:22:53 UTC

Okay, let's see if the Nokians can do something here software-wise...

<a
href="http://maemo-freak.com/index.php/reviews/devices/1374-nokia-n97-mini-vs-n900-camera-showdown-">http://maemo-freak.com/index.php/reviews/devices/1374-nokia-n97-mini-vs-n900-camera-showdown-</a>

It would be very premature to write this bug off as a hardware problem.  The
above link compares performance of the N900 camera with the N97 mini camera,
which is especially enlightening since the camera hardware between the two
phones is pretty much identical (for all intents and purposes).  Thus, any
major differences can be attributable to software.

Firmware updates for Nokia's Symbian phones have resulted in very impressive
camera improvements.  These are clearly the result of SOFTWARE fixes (again,
see the link above for a great side-by-side comparison).

The marked difference between pictures taken at a distance minimizes the
likelihood of the chrome trim playing a major role in poor N900 photos (again,
see the link above).  With the open-sourcing of Symbian, it would be very
worthwhile to compare the image processing techniques between the two platforms
to try and isolate why the N900 seems to suffer compared to other Nokia phones
with similar/identical camera hardware.

"minimizes the likelihood of the chrome trim playing a major role in poor N900
photos"

See the talk.maemo.org thread, however, where there are also side-by-side,
before and after comparisons (either blacking the trim, or the side of the
slider, or both). That seems to count as evidence the other way.

(In reply to comment #11)
> "minimizes the likelihood of the chrome trim playing a major role in poor N900
> photos"
> 
> See the talk.maemo.org thread, however, where there are also side-by-side,
> before and after comparisons (either blacking the trim, or the side of the
> slider, or both). That seems to count as evidence the other way.
> 

The talk.maemo.org thread is - unfortunately - a huge distraction.  99% of the
people there have NO idea what they're talking about.  Yes, the initial example
is interesting - but compare comment #26
(http://talk.maemo.org/showpost.php?p=423940&postcount=26) with comment #40
(http://talk.maemo.org/showpost.php?p=426424&postcount=40).  

Quote: "All conditions exactly the same as yesterday. No reflection from the
flash! This has got me scratching my head..."

As that demonstrates, it doesn't make much sense to blame hardware alone.  I
wouldn't be surprised if a minor tweak to the flash/shutter timing had a huge
impact on any supposed "reflection" (assuming it's a problem at all).  Another
possible explanation is flash intensity variances (e.g. the initial flash is
brighter than subsequent flashes - due to LED fatigue - so maybe the flash just
needs to be toned down.)

As a separate point of reference, my 12mp Kodak point-and-shoot has a shiny
metal housing around the entire lens and it suffers no flash reflection at any
distance - hence why I find it extremely premature to blame the N900's chrome
trim.

Another relevant quote from the same thread:

"I took some photos with and without the cover flash on and did not notice any
hazy effect so I do not think this is a problem affecting all n900 phones."

This has been my experience as well.  My wife and I both have N900s.  Mine
exhibits the corner glare that others in that thread mention (which is why I
looked up this bug in the first place).  My wife's phone does not.  It's
debatable whether something like minute variances in LED placement could cause
that, but ultimately it's somewhat irrelevant since even IF the chrome trim is
a problem, the right software algorithms could still ameliorate it.

To clarify, my main point is this:

Clearly Nokia has found a way to improve camera quality via software updates in
their other N-series phones.  Based on all the comparisons listed here, the
general consensus is that these other N-series phones outperform the N900.  The
most efficient way to handle this bug would be to see what software
improvements were made to the other N-series cameras (if possible), then
discuss why those same changes could/couldn't be applied here.

No need to reinvent the wheel.

Hello.  In my original post regarding this subject I included a pointer to
comparison photos, side by side, of the N900 and the N85 8MP.  Another user had
suggested that it was the N900's metal trim around the lens area causing the
problem.  He/she also commented that covering the trim improved the photo
quality.  I tried this myself and also got a dramatically improved result. 
However, as I continued trying the N900 camera out I found other situations
when the N900 camera's quality was better than my N900.  It seemed so random...
 For the first time I had a good look at my N85 8MP hardware and noticed it has
the exact same chrome trim around the photo lens area as my N900.  So, given
the random nature of good vs. bad quality photos on the N900 and given that the
N900 and the N85 8MP share the same chrome trim, I also now conclude that the
photo quality issue on the N900 is not because of the chrome trim.

I took photos from both my N900 and N97 and the N900's quality is extremely
poor

Comment 16 Andre Klapper 2010-02-21 20:59:37 UTC

(In reply to comment #15)
> I took photos from both my N900 and N97 and the N900's quality is extremely
> poor

Please avoid subjective repetitive (see comment 10) statements, as they don't
add any additional information and just create bugmail. Thanks a lot! :)

Comment 17 Andre Klapper 2010-02-23 22:31:56 UTC

Forwarding Nokia's internal comment and WONTFIX decision:

"The amount of smudge on the lens seems to affect this heavily and the haze is
barely noticeable when the lens is clean. The lux level trigger is already at
50 LUX. Also the problem is worse in darker conditions anyway where most of the
light is from the flash instead of the surrounding light so reducing the lux
level would only help in conditions that aren't that bad anyway.
No fix is available that would make sense."

How can it be claimed that 'smudge on the lens' 'affect[s] this heavily' if,
afaik, no one on this issue has been asked to inspect their lens for same ? 
Ya know, do some informal polling ?
FWIW I just checked mine (the one that was made better by magic marker on the
side of the lens cover) and no smudge.

What is this 'LUX' parameter ?

(In reply to comment #18)
> How can it be claimed that 'smudge on the lens' 'affect[s] this heavily' if,
> afaik, no one on this issue has been asked to inspect their lens for same ? 
I also think someone on Nokia's side should comment about all that tmo flash
voodoo. Especially considering that there were reports of photos being better
when the back cover is removed:
http://talk.maemo.org/showpost.php?p=494372&postcount=50
Dirty lens doesn't explain that.

> What is this 'LUX' parameter ?
http://en.wikipedia.org/wiki/Lux

(In reply to comment #17)
> Forwarding Nokia's internal comment and WONTFIX decision:
> 
> "The amount of smudge on the lens seems to affect this heavily and the haze is
> barely noticeable when the lens is clean. The lux level trigger is already at
> 50 LUX. Also the problem is worse in darker conditions anyway where most of the
> light is from the flash instead of the surrounding light so reducing the lux
> level would only help in conditions that aren't that bad anyway.
> No fix is available that would make sense."
> 

Thanks for the update, Andre.  Unfortunately, the response you've posted seems
to only address a narrow part of the problem at hand (supposed "smudging", be
it from silver trim or other causes) when the core problem is really a larger
issue of overall poor camera performance under a number of lighting conditions.

I realize that this is a difficult problem to quantify.  Would we find more
success by opening a new bug along the lines of "The N900 camera performs
poorly compared to identical 5MP camera hardware/flash in other N-series
phones"...?  For me, that seems to be the most quantifiable description of this
bug (especially with all the links in this report providing clear evidence).

Also: Drew - is your other phone a 5MP N85 or an 8MP N86?  I have a 5MP N85
that has near-identical camera hardware to the N900, and I'd be happy to
provide another batch of side-by-side comparisons if they will help. 
Unfortunately, I worry that the 8MP CCD in an N86 is too different to provide a
clear example of what's wrong with the N900 implementation....

I've compared the 5MP camera on my old Sony Ericsson k850i to the n900....the
n900 is indeed poor.

tannerhelland@hotmail.com asked if my photo comparison was N900 vs. N85 5MP or
N86 8MP.  I apologize for my messiness in not noting that clearly.  My
comparison was of a N900 (set at 5MP quality) vs. N86 8MP (set at 5MP quality).
 Here are the photos for your reference:
http://picasaweb.google.com/drew.freyman/NokiaN900VsN868MP?feat=directlink

Given other camera models with chrome surrounding the camera lens area can take
good quality photos I think we can rule out the chrome trim as the root cause
of the N900's problems.  Taking the back off, covering the chrome with magic
marker, etc. have produced better results under certain circumstances. 
However, as this is neither consistent on the N900 nor reproducible on other
models with chrome trim around the lens area, I am inclined to conclude that
the chrome trim hardware design is not responsible for the photo quality issue.

I have started to believe the problem is related to the way the N900 handles
light under certain circumstances.  I can't be sure what those are.  But, in my
case I have noticed the white haze is more likely to appear in photos where
golds and browns predominate in the tone.  In photos where blues and blacks
predominate I have not experienced the haze problem.

Comment 23 Andre Klapper 2010-10-26 00:32:21 UTC

(In reply to comment #19)
> I also think someone on Nokia's side should comment about all that tmo flash
> voodoo. Especially considering that there were reports of photos being better
> when the back cover is removed:

I posted this exact comment three times internally without a response.
Feel free to draw your own conclusions.