Any thoughts /news on Foveon sensors?
Edmund
As far as I understand things, this foveon sensor
looks the best technology to me.
Why is it hardly anybody uses it?
Are there made to big investments in other systems
and are "they" therefore choosing a lesser system?
Darrell Larose
"Edmund" <nom...@hotmail.com> wrote in message
news:dqap99$oc4$1...@azure.qinip.net...
end. There has been no new developments for several years. Foveon is stuck
at 3.4 megapixels.
Mark˛
> As far as I understand things, this foveon sensor
> looks the best technology to me.
> Why is it hardly anybody uses it?
Because its only incarnation came in the for of the extremely poor Sigma
body.
-Don't know if its the fault of Sigma or Foveon...but the skin renditions
were absolutely horrid, and many other texture renditions were just plain
awful. The lack of a filter over the sensor often leads to problematic fine
line renditions... The list goes on.
Here's dpreview's list of cons:
-No in-camera JPEG, storage implications of shooting RAW
-Very disappointing long exposure / night performance
-Gradual hue shift at higher sensitivities, softer images from ISO 200
upwards
-Heavier and bulkier than some of the competition (although never feels too
big)
-Not as feature rich as other digital SLR's
-Sigma QC should ensure no dust on sensor from factory (we had none, others
have had)
-Continuous shooting rate limited to 2 frames per second at Hi resolution
-No onboard PC Sync flash socket (req. optional adapter)
-The need to have Sigma SA mount lenses
> Are there made to big investments in other systems
> and are "they" therefore choosing a lesser system?
??
Try that question again.
rafe b
number here)@cox..net> wrote:
>Here's dpreview's list of cons:
>
>-No in-camera JPEG, storage implications of shooting RAW
>-Very disappointing long exposure / night performance
>-Gradual hue shift at higher sensitivities, softer images from ISO 200
>upwards
>-Heavier and bulkier than some of the competition (although never feels too
>big)
>-Not as feature rich as other digital SLR's
>-Sigma QC should ensure no dust on sensor from factory (we had none, others
>have had)
>-Continuous shooting rate limited to 2 frames per second at Hi resolution
>-No onboard PC Sync flash socket (req. optional adapter)
>-The need to have Sigma SA mount lenses
The OP asked about Foveon sensors, not Sigma
cameras.
In theory, at least, the two should be distinct, though in
current reality, it appears Foveon=Sigma.
rafe b
www.terrapinphoto.com
Roger N. Clark (change username to rnclark)
> Edmund wrote:
>
>>As far as I understand things, this foveon sensor
>>looks the best technology to me.
>>Why is it hardly anybody uses it?
>
>
> Because its only incarnation came in the for of the extremely poor Sigma
> body.
> -Don't know if its the fault of Sigma or Foveon...but the skin renditions
> were absolutely horrid, and many other texture renditions were just plain
> awful. The lack of a filter over the sensor often leads to problematic fine
> line renditions...
The foveon sensor uses a subtractive color scheme. This makes it
a problem in getting good color balance. One might think
that ink and paper uses subtractive color (CMYK) but that
is absorption of light in the inks doing the subtraction,
not light adding as in a camera sensor. The three color
"filters" response curves in the foveon sensors do not match
that of the eye and have large light leaks at wavelengths covered
by other filters, all making it difficult to get good
let alone accurate color. (I have a link somewhere to the
response curves, but can seem find it right now--try a google
search).
Roger
Gormless
> looks the best technology to me.
> Why is it hardly anybody uses it?
It's not traditional digital technology. There is a small and rapidly
becoming disillusioned group of fanatic diehard adherents of Sigma's two
Foveon cameras, but the absence of any new models for far too long has left
them stuck at a very last-century, mediocre, noisy and rather yellowish
(though that last is down to Sigma, not Foveon) 3.4 Mp. Though its
adherents will swear differently, I've never once seen or heard of a
Foveon/Sigma in professional use. The only other camera which now uses
Foveon is a cheap Polaroid digital, which seems to have had limited success.
While strictly in theory it IS good technology, Foveon's failure to deliver
and Sigma's failure to make best use of it has rather left it floundering
and out of its depth in the modern digital climate.
Apart from a lot of speculation, there is no indication that Foveon are
working on anything new, and I suspect that our Sigma-loving friends, if and
when they ever get a new model, might find it doesn't contain a Foveon
sensor.
Helen
Peter A. Stavrakoglou
news:43c8eb04$0$30416$6d36...@taz.nntpserver.com...
Not true. Although the choice for a reseller using WWL for the Polariod
camera using a Foveon was not a good one, that alone shows that there has
been development in the last two years. That camera could produce some good
images. The Sigma SD10 is a little over two years old and it's 3.4 MP
sensor resolves as well as any 6MP Bayer, sometimes even better.
Peter A. Stavrakoglou
At this time, only Sigma and one Polaroid P&S (if you can find one) are
using these sensors in a consumer camera. There will be a new Sigma SLR
soon, patience will pay off in waiting for it.
Compared to Bayer technology, the Foveon is still in its infancy and yet it
produces very sharp photos with good color rendition in a Sigma DSLR. As
development continues, it can only improve like Bayer technology has.
Despite their being many more years of development down the road, the
current Foveon sensor can hold its own. There needs to be improvement in
noise handling and perhaps in-camera jpeg in the Sigma DSLRs.
Peter A. Stavrakoglou
news:526yf.10707$V.7118@fed1read04...
> Edmund wrote:
>> As far as I understand things, this foveon sensor
>> looks the best technology to me.
>> Why is it hardly anybody uses it?
>
> Because its only incarnation came in the for of the extremely poor Sigma
> body.
> -Don't know if its the fault of Sigma or Foveon...but the skin renditions
> were absolutely horrid, and many other texture renditions were just plain
> awful. The lack of a filter over the sensor often leads to problematic
> fine line renditions... The list goes on.
Saying that skin renditions are horrid is an overstatement. Depending on
the exposure, skintones tend to be yellow but horrid they are not. Easily
correctable but some take issue with havng to do that. The Sigma camera
themselves are not "extremely poor", that is only your opinion. It's a
well-built body with vey good ergonomics and in the right hands produces
outstanding images. Search pbase and you'll see. Those of us who use them
know how good it really is, t seems that statements like yours come from
people who never used the camera but I am not making a judgment on your
situation. Perhaps you've used one, I don't know, but I have for two years
and in my totally amateurish hands it has taken some great photos. The
level of sharpness is surprising at times.
Scott W
bring something to the table that was not otherwise available. With
its 3.4 MP it came very close to matching a 6 MP camera, but it did not
even manage to do this. If they could come out with a 8 MP version and
control the noise better they might get somewhere.
In truth the loss from the Bayer pattern is fairly small, say between
25 to 30% of the pixel count. This assumes you are shooting raw not
just using the in camera jpeg photos.
Scott
Gormless
"Peter A. Stavrakoglou" <nto...@optonline.net> wrote in message
news:rQ8yf.842$IM2...@fe12.lga...
> it's 3.4 MP
> sensor resolves as well as any 6MP Bayer, sometimes even better.
>
I don't understand how that is a technical possibility.
Helen
Malcolm Stewart
news:9eadnTMs1an...@pipex.net...
Bayer needs 4 sites per pixel - Foveon does it with depth, 3 on top of one,
so to speak..
--
M Stewart
Milton Keynes, UK
http://www.megalith.freeserve.co.uk/oddimage.htm
Skip M
and Minolta 5D. It has been left behind by such cameras as the Oly E-500
and Canon RebelXT/350D, and seriously dusted off by cameras like the Nikon
D200 (at the price point the Sigma SD-10 was introduced) and the admittedly
much more expensive Canon 5D.
--
Skip Middleton
http://www.shadowcatcherimagery.com
Randall Ainsworth
<sim...@gormless.com> wrote:
> I don't understand how that is a technical possibility.
> Helen
It's not. The Sigma is a 3.42MP camera.
Mark˛
> "Mark˛" <mjmorgan(lowest even number here)@cox..net> wrote in message
> news:526yf.10707$V.7118@fed1read04...
>> Edmund wrote:
>>> As far as I understand things, this foveon sensor
>>> looks the best technology to me.
>>> Why is it hardly anybody uses it?
>>
>> Because its only incarnation came in the for of the extremely poor
>> Sigma body.
>> -Don't know if its the fault of Sigma or Foveon...but the skin
>> renditions were absolutely horrid, and many other texture renditions
>> were just plain awful. The lack of a filter over the sensor often
>> leads to problematic fine line renditions... The list goes on.
>
> Saying that skin renditions are horrid is an overstatement. Depending on
> the exposure, skintones tend to be yellow but horrid
> they are not. Easily correctable but some take issue with havng to
> do that.
Disagree.
It often went beyong mere color.
Texture was often the culprit in addition to color, which is pretty
impossible to recreate in PS.
>The Sigma camera themselves are not "extremely poor", that
> is only your opinion.
Right. My opinion.
Mark˛
Do you have images posted anywhere?
You seem to indicate better results than I typically saw from
Sigma;--particularly skin tones and textures.
If you've got the goods, I'm open to reinterpretation of my opinion...though
I have yet to see skin renditions that consistently un-do my perception of
this.
Alan
the image is interpolated.
Scott W
> Scott, how can you say only a 25 to 30% loss? With a CCD fully 66% of
> the image is interpolated.
Here is a test I did
http://www.pbase.com/konascott/image/54437327
The top photo is a straight out of the 20D (well from raw), The middle
photo is about as sharp as be had for the number of pixels, it is down
sampled by a large amount. The difference between these two images
then is the loss from the Bayer pattern. To test how much loss there
is simple down sample and back up, when you find the amount of down
sampling need to degrade the middle image to match the top you pretty
much know what the loss is. Down sampling by 85% and back up come very
close to what the camera produces, square this and you end up with a
loss of 28%. But some of this loss is most likely due to the lens as
well as the Bayer pattern so the real loss from the Bayer pattern will
be somethin gless then 28%
Scott
Gormless
"Randall Ainsworth" <r...@nospam.techline.com> wrote in message
news:140120061152511739%r...@nospam.techline.com...
I think Peter A S must be a Sigma adherent.
If someone believes something hard enough if it becomes true for them.
David Dyer-Bennet
> As far as I understand things, this foveon sensor
> looks the best technology to me.
> Why is it hardly anybody uses it?
Because it's a LOT better in theory than it turns out to be in
practice. At least so far.
--
David Dyer-Bennet, <mailto:dd...@dd-b.net>, <http://www.dd-b.net/dd-b/>
RKBA: <http://noguns-nomoney.com/> <http://www.dd-b.net/carry/>
Pics: <http://dd-b.lighthunters.net/> <http://www.dd-b.net/dd-b/SnapshotAlbum/>
Dragaera/Steven Brust: <http://dragaera.info/>
Randall Ainsworth
<sim...@gormless.com> wrote:
> I think Peter A S must be a Sigma adherent.
> If someone believes something hard enough if it becomes true for them.
He's a hardcore Sigma owner. I've tried to show him the truth, but
these Sigma people adhere to the company's PR and refuse to see the
limitations of the severely flawed hardware.
Peter A. Stavrakoglou
I'm a Sigma user and if you were then you would see it's true, you don't
have to "try" to believe anything you only have to see for yourself. Read
the reviews of the camera on www.dpreview.com and other sites, they'll all
tell you that the Foveon resolves as well or better than a 6MP Bayer. Look
at the photos, you can't argue with the physical evidence. The Foveon can
resolve detail that a Bayer sensor cannot.
Peter A. Stavrakoglou
If you perused the Sigma forum on dpreview you'd "meet" a few pros who use a
Sigma DSLR, a couple of whom earn big bucks from poster-sized prints.
Peter A. Stavrakoglou
LOL! You are too funny. I'm hardly hardcore, if I was then I wouldn't have
added a Nikon to my camera bag.
Mark˛
Highlight on the word "few."
:)
Mark˛
Why did you feel the need to add a Nikon?
Peter A. Stavrakoglou
You are more than welcome to peruse through my images, but I'll warn you
that I don't even rank as an amateur but I nonetheless enjpy taking
pictures. I'll be the first to say that the SD9 and SD10 can produce yellow
skintones, it's a problem but it's easily corrected. It doesn't happen all
the time but it happens a lot.
Here's a few links:
Not the sharpest photo but when you only have a cheap zoom...
http://ntotrr.smugmug.com/photos/12645878-M.jpg
Here's my niece:
http://ntotrr.smugmug.com/photos/13539136-M.jpg
And her brother:
http://ntotrr.smugmug.com/photos/27233203-M.jpg
The skintones are quite accurate, the first one took no adjustment at all,
the other two took a minor tweak.
Here's one that shows the detail this 3.4 MP camera can capture:
http://ntotrr.smugmug.com/photos/31953396-M.jpg
Peter A. Stavrakoglou
news:X1byf.11492$sA3.9427@fed1read02...
> "Peter A. Stavrakoglou" <nto...@optonline.net> wrote in message
> news:lV8yf.843$IM2...@fe12.lga...
>> "Edmund" <nom...@hotmail.com> wrote in message
>> news:dqap99$oc4$1...@azure.qinip.net...
>>>
>>> As far as I understand things, this foveon sensor
>>> looks the best technology to me.
>>> Why is it hardly anybody uses it?
>>> Are there made to big investments in other systems
>>> and are "they" therefore choosing a lesser system?
>>
>> At this time, only Sigma and one Polaroid P&S (if you can find one) are
>> using these sensors in a consumer camera. There will be a new Sigma SLR
>> soon, patience will pay off in waiting for it.
>>
>> Compared to Bayer technology, the Foveon is still in its infancy and yet
>> it produces very sharp photos with good color rendition in a Sigma DSLR.
>> As development continues, it can only improve like Bayer technology has.
>> Despite their being many more years of development down the road, the
>> current Foveon sensor can hold its own. There needs to be improvement in
>> noise handling and perhaps in-camera jpeg in the Sigma DSLRs.
>>
> It only holds its own with the bottom of the line cameras like the Nikon
> D50 and Minolta 5D. It has been left behind by such cameras as the Oly
> E-500
Don't leave out the D70 and 7D
> and Canon RebelXT/350D, and seriously dusted off by cameras like the Nikon
> D200 (at the price point the Sigma SD-10 was introduced) and the
> admittedly much more expensive Canon 5D.
C'mon Skip, comparing current DSLR prices with one that is over two years
old. The SD10 is much less than a D200, half the price or less, if you can
find one. It's been discontinued since there's a new one on the way.
Peter A. Stavrakoglou
> Peter A. Stavrakoglou wrote:
>> "Mark˛" <mjmorgan(lowest even number here)@cox..net> wrote in message
>> news:526yf.10707$V.7118@fed1read04...
>>> Edmund wrote:
>>>> As far as I understand things, this foveon sensor
>>>> looks the best technology to me.
>>>> Why is it hardly anybody uses it?
>>>
>>> Because its only incarnation came in the for of the extremely poor
>>> Sigma body.
>>> -Don't know if its the fault of Sigma or Foveon...but the skin
>>> renditions were absolutely horrid, and many other texture renditions
>>> were just plain awful. The lack of a filter over the sensor often
>>> leads to problematic fine line renditions... The list goes on.
>>
>> Saying that skin renditions are horrid is an overstatement. Depending on
>> the exposure, skintones tend to be yellow but horrid
>> they are not. Easily correctable but some take issue with havng to
>> do that.
>
> Disagree.
> It often went beyong mere color.
> Texture was often the culprit in addition to color, which is pretty
> impossible to recreate in PS.
Perhaps you've grown accustomed to the textures smoothed out by AA filters.
I find the skin textures from the Foveon to be quite accurate, too accurate
for those who want a smoother look.
Peter A. Stavrakoglou
news:1137254791.0...@z14g2000cwz.googlegroups.com...
The noise issue is the biggest one for them. That's why I bought a D50, for
low-light situations. Not that I needed a second body - I really need to
sharpen my skills and no camera can improve on them - but as long as someone
was willing to get me one I couldn't refuse.
SMS
> As far as I understand things, this foveon sensor
> looks the best technology to me.
> Why is it hardly anybody uses it?
> Are there made to big investments in other systems
> and are "they" therefore choosing a lesser system?
Silicon color separation never worked as well as hoped, and the noise
levels were very high. This required tremendous amounts of
post-processing, outside the camera, to obtain decent images.
Mark˛
You've controlled the yellow, though it still has that odd texture and
transition appearance that's hard to describe. These kid photos exhibit the
kind of wax-like skin texture appearance that I've observed in the past on
Sigma images. It's dificult to put into words, but it's something I've seen
with particular tendency in smoother-skinned folk...especially young
children photos. I'm thinking this has something to do with the lack of
filter over the sensor, and resulting tendency to deal with certain textures
a bit strangely.
> Here's one that shows the detail this 3.4 MP camera can capture:
> http://ntotrr.smugmug.com/photos/31953396-M.jpg
I'm sure you've got detail in there, though at this small web-page size,
it's impossible for viewers to judge.
Mark˛
> Here's a few links:
>
> Not the sharpest photo but when you only have a cheap zoom...
> http://ntotrr.smugmug.com/photos/12645878-M.jpg
>
> Here's my niece:
> http://ntotrr.smugmug.com/photos/13539136-M.jpg
>
> And her brother:
> http://ntotrr.smugmug.com/photos/27233203-M.jpg
>
> The skintones are quite accurate, the first one took no adjustment at
> all, the other two took a minor tweak.
>
> Here's one that shows the detail this 3.4 MP camera can capture:
> http://ntotrr.smugmug.com/photos/31953396-M.jpg
I re-posted your image, along with a copy with my notes on it.
Click here:
http://www.pbase.com/markuson/image/54837338/original
and then click the "previous" word in upper-right of screen, which will
switch directly to the notated version.
If you ONLY look at the marked image, it's more difficult to see what I'm
talking about...but if you toggle between the two images (marked and
un-marked), it will help you see where I'm looking.
BTW--If, for some reason, you don't like that I re-posted this to enable
comment...let me know, and I'll remove it ASAP.
(Cute kid, BTW).
:)
-Mark
Peter A. Stavrakoglou
Low noise. I'm confident that the next Sigma DSLR will also be clean at
high ISO (it has to be or it's doomed) but there are extenuating
circumstances for not waiting and getting a Nikon D50 now. My wife wanted
to take ballroom dancing lessons. Over my dead body was my first response
but then I came to my senses and realized that it is for her that I'll do
anything so I agreed, after she said she'd get me a new camera :) So I
struck while the iron was hot not willing to blow the opportunity. The D50
is a nice camera, small but not too small. I didn't like the feel of the
Rebel XT and the 20D is a bit too high in proce for me right now. I added a
Tokina 28-70mm AT-X PRO lens. It's built like a tank and I like the results
thus far. I've only had it for a month. This combo doesn't seem to produce
yellow skintones http://ntotrr.smugmug.com/photos/50519191-M.jpg (she hates
it when I take her picture). I'm very impressed with the low-noise of this
camera at high ISO, Nikon has done a great job with it.
Gormless
> >
> > It's not traditional digital technology. There is a small and rapidly
> > becoming disillusioned group of fanatic diehard adherents of Sigma's two
> > Foveon cameras, but the absence of any new models for far too long has
> > left
> > them stuck at a very last-century, mediocre, noisy and rather yellowish
> > (though that last is down to Sigma, not Foveon) 3.4 Mp. Though its
> > adherents will swear differently, I've never once seen or heard of a
> > Foveon/Sigma in professional use. The only other camera which now uses
> > Foveon is a cheap Polaroid digital, which seems to have had limited
> > success.
> > While strictly in theory it IS good technology, Foveon's failure to
> > deliver
> > and Sigma's failure to make best use of it has rather left it
floundering
> > and out of its depth in the modern digital climate.
> >
> > Apart from a lot of speculation, there is no indication that Foveon are
> > working on anything new, and I suspect that our Sigma-loving friends, if
> > and
> > when they ever get a new model, might find it doesn't contain a Foveon
> > sensor.
> >
> > Helen
>
> If you perused the Sigma forum on dpreview you'd "meet" a few pros who use
a
> Sigma DSLR, a couple of whom earn big bucks from poster-sized prints.
>
I note that this was the only statement of mine with which you took issue.
Again, and for the record, during over 7 years of varied professional
photography in both the UK and the USA I am yet to see or hear of a pro
using Sigma cameras. Lenses yes, I'll grant you, but not a camera. As I
apparently correctly predicted above - "though its adherents will swear
differently".
Gormless
>
> I'm a Sigma user and if you were then you would see it's true, you don't
> have to "try" to believe anything you only have to see for yourself. Read
> the reviews of the camera on www.dpreview.com and other sites, they'll all
> tell you that the Foveon resolves as well or better than a 6MP Bayer.
Look
> at the photos, you can't argue with the physical evidence. The Foveon can
> resolve detail that a Bayer sensor cannot.
>
Yeah right, I'm certainly going to trade in my 1DS Mk IIs and 5Ds for one of
those silly things.
Peter A. Stavrakoglou
Ooops, that one is not large enough to do it justice, you are right. I
don't have the full-size photo posted. That was a tough one to get, given
that I shot it at ISO 100 at 1/45. I'm surprised that there isn't more
camera-shake visible.
Peter A. Stavrakoglou
He's cute - his name is Mark. I don't mind at all that you've reposted the
pic. Actually, I've done the same thing a couple of times when a new Sigma
user posted an image in dpreview forum asking about what to do with yellow
skintones :) I'll have to investigate what you've pointed out. Perhaps the
lighting can cause this, I don't rightly know.
Mark˛
I've noticed many times, and fairly consistently, that facial textures and
color transitions where the angle of the face curves away from the camera
position...that these odd tones/textures appear. The pic I marked on
exhibited only a minor example of this, but I've seen it a lot in Foveon
shots.
Bart van der Wolf
SNIP
> The Sigma SD10 is a little over two years old and it's 3.4 MP sensor
> resolves as well as any 6MP Bayer, sometimes even better.
I'm sorry for having to correct that, but the apparent resolution
similarity stems from the use/absence of an anti-aliasing filter. It
has little to do with the sensor itself, other than how tolerable
aliasing artifacts are on both sensor technologies.
Bart
Bart van der Wolf
SNIP
> The Foveon can resolve detail that a Bayer sensor cannot.
That "detail" is called "aliasing artifacts" and they have little to
do with actual detail. They are larger than life aberrations caused by
the simple inability of any(!) discrete sampling device (like a photo
sensitive sensor element) to reliably reproduce detail that's smaller
than 1 pixel, unless an anti-aliasing filter is used in front of the
sensor.
Bart
Rich
wrote:
>
>"Peter A. Stavrakoglou" <nto...@optonline.net> wrote in message
>news:rQ8yf.842$IM2...@fe12.lga...
>> it's 3.4 MP
>> sensor resolves as well as any 6MP Bayer, sometimes even better.
>>
>
>I don't understand how that is a technical possibility.
>Helen
>
So what? Most of the reviewers have said the same thing and provided
comparison images to illustrate it. Foveon, IMO, is just short of
tricolour in terms of potential quality. It remains to be seen if
it's going to be economical, or "politically" acceptable for any major
mfg. to keep going with the technology.
-Rich
Peter A. Stavrakoglou
Bill Funk
wrote:
>
>"Peter A. Stavrakoglou" <nto...@optonline.net> wrote in message
>news:rQ8yf.842$IM2...@fe12.lga...
>> it's 3.4 MP
>> sensor resolves as well as any 6MP Bayer, sometimes even better.
>>
>
>I don't understand how that is a technical possibility.
>Helen
>
Technically, of course, it's not.
But the Sigma SDx cameras don't use an AA filter, so sharpness right
out of the sensor is higher, but the lack of the AA filter also
produces "fake" sharpness in many types of subjects.
For example, look at photos of a building with a shingle roof taken
with a Bayer-filtered 'normal' (CCD or CMOS) sensor, and a SDx camera.
The roofs look very different; the shingles on the Sigma photo will
change size as the eye travels up to the roof's top, while the
shingles from the 'normal' camera will more fade into an amorphous
blob (this depends on several factors, like distance and angle, of
course). Neither handles this well, but the Sigma offers a sharp image
that is false. The lack of an Anti-Aliasing filter causes this.
So, the Foven-sensor image "looks" sharper, but the sharpness displays
fake detail that isn't in the original subject.
--
Bill Funk
Replace "g" with "a"
funktionality.blogspot.com
Mark˛
No. I think he's just noting the rarity of Sigma bodies within professional
circles.
Bill Funk
<nto...@optonline.net> wrote:
>> Texture was often the culprit in addition to color, which is pretty
>> impossible to recreate in PS.
>
>Perhaps you've grown accustomed to the textures smoothed out by AA filters.
>I find the skin textures from the Foveon to be quite accurate, too accurate
>for those who want a smoother look.
I have often wondered why people will post portraits by the Sigma SDx
cameras as examples of the cameras' prowess.
The very sharpness that is exhibited by those cameras works against
them in portrait work. The portraits look strange, IOW. I'm convinced
this is a result of the lack of an AA filter.
Bill Funk
<nto...@optonline.net> wrote:
>> and Canon RebelXT/350D, and seriously dusted off by cameras like the Nikon
>> D200 (at the price point the Sigma SD-10 was introduced) and the
>> admittedly much more expensive Canon 5D.
>
>C'mon Skip, comparing current DSLR prices with one that is over two years
>old. The SD10 is much less than a D200, half the price or less, if you can
>find one. It's been discontinued since there's a new one on the way.
>
You said this before; can you point to any online rumors/reports of an
upcoming SD camera using Foveon technology?
Does this mean a new Foveon sensor?
With (hopefully) an AA filter? :-)
Mark˛
They are posted because the portrait is what speaks to the yellow skin
issue, and the strange skin texture issue.
I think the lack of AA is definitely related to the odd textures, but I
think the yellowing and odd color transitions have more to do with the
depth-based color receptors in the sensor.
Gormless
> > Again, and for the record, during over 7 years of varied professional
> > photography in both the UK and the USA I am yet to see or hear of a pro
> > using Sigma cameras. Lenses yes, I'll grant you, but not a camera. As
I
> > apparently correctly predicted above - "though its adherents will swear
> > differently".
>
> So I'm lying?
>
Without suggesting you're a liar - I'm sure you're not - my own personal and
extensive experience indicates that your perception of the truth in this
particular matter is not equal to mine.
Helen
Edmund
wrote:
> Alan wrote:
>
> > Scott, how can you say only a 25 to 30% loss? With a CCD fully 66% of
> > the image is interpolated.
>
> Here is a test I did
> http://www.pbase.com/konascott/image/54437327
> The top photo is a straight out of the 20D (well from raw), The middle
> photo is about as sharp as be had for the number of pixels, it is down
> sampled by a large amount. The difference between these two images
> then is the loss from the Bayer pattern. To test how much loss there
> is simple down sample and back up, when you find the amount of down
> sampling need to degrade the middle image to match the top you pretty
> much know what the loss is. Down sampling by 85% and back up come very
> close to what the camera produces, square this and you end up with a
> loss of 28%. But some of this loss is most likely due to the lens as
> well as the Bayer pattern so the real loss from the Bayer pattern will
> be somethin gless then 28%
>
> Scott
I don't know much about this technic and by no means I am an
expert but I think this way you can "prove" the bayer interpolation
has no loss at all. You simply have to take a picture of an total
uniform colored screen with no detail on it.
ch...@go.com
>the image is interpolated.
Scott's tests are very interesting, and rather hard to argue with.
Alan, are you aware of exactly what sort of information is
interpolated, and how it is done? In simple terms my understanding is
that monochrome information from every sensor is used, so there is no
interpolation of the *monochromatic* detail 'underneath' the
interpolated colour data. That would explain why there is not the loss
that you are suggesting, and the real figure is much closer to Scott's.
Largely because much of the detail we see is *not* strictly in one
colour channel and not in another...
Do you have conflicting references or links?
Edmund
> On Sat, 14 Jan 2006 17:44:24 -0000, "Gormless" <sim...@gormless.com>
> wrote:
>
> >
> >"Peter A. Stavrakoglou" <nto...@optonline.net> wrote in message
> >news:rQ8yf.842$IM2...@fe12.lga...
> >> it's 3.4 MP
> >> sensor resolves as well as any 6MP Bayer, sometimes even better.
> >>
> >
> >I don't understand how that is a technical possibility.
> >Helen
> >
> Technically, of course, it's not.
Well that is open to discussion, correct me if I am wrong.
to start with : "whats is a name"
Any "bayer" 6MP camera has only 6MP sensors total
to be divided for all colors.
For each layer the bayer interpolation comes in to
guess the missing pixels. That works remarkably well,
especially in gradually changing colors.
It goes wrong in sharp contrasts between black and white.
The so called Foveon sensor has 3 layers of 3.4 MP
So each sensor (from the 3.4MP ) gives a real existing pixel.
Bill Funk
That's what I think, too, but others tell me it's the fault of the
camera, not the sensor. They say the processing of the image off the
sensor is what does it.
To me it's six of one...
Edmund
wrote:
Like to see your comment on this site.
http://www.foveon.com/article.php?a=70
David J. Littleboy
> Bill Funk wrote:
>> "Gormless" <sim...@gormless.com> wrote:
>> >
>> >"Peter A. Stavrakoglou" <nto...@optonline.net> wrote:
>> >> it's 3.4 MP
>> >> sensor resolves as well as any 6MP Bayer, sometimes even better.
>> >
>> >I don't understand how that is a technical possibility.
>> >Helen
>> >
>> Technically, of course, it's not.
Exactly.
> Well that is open to discussion, correct me if I am wrong.
> to start with : "whats is a name"
> Any "bayer" 6MP camera has only 6MP sensors total
> to be divided for all colors.
Sort of.
> For each layer the bayer interpolation comes in to
> guess the missing pixels. That works remarkably well,
> especially in gradually changing colors.
> It goes wrong in sharp contrasts between black and white.
No. It works quite nicely on monochrome patterns. Look at the test charts on
dpreview, for example. The Bayer sensors correctly resolve up to about 2/3
the Nyquist freuquency. If you look at the Foveon test patters, you'll see
that basically that's all the Foveon resolves as well. Above 2/3 Nyquist,
the Foveon sensors are a mass of Moire. They have false responses at the
Nyquist frequency for frequencies at and above the Nyquist. (A digital file
can't even represent frequencies at and above the Nyquist, so the idea that
a camera could (or should) have a response at or above the Nyquist frequency
is completely ridiculous.)
The reason people think Foveon outresolves Bayer is that the Sigma Foveon
cameras don't have the (mathematically) required (for correct imaging)
antialiasing filter; if they did, the bogus response that the Foveon fans
think is resolution would be suppressed, and they'd look like any other
3.4MP camera.
> The so called Foveon sensor has 3 layers of 3.4 MP
> So each sensor (from the 3.4MP ) gives a real existing pixel.
That doesn't enable them to _correctly_ resolve any higher frequencies
(relative to the pixel spacing) than Bayer sensors for monochrome test
patterns. Essentially, there's zero advantage.
The place that there is an advantage is in test patterns that are red on
blue instead of black and white. But people can't see those patterns (if you
try, you'll get a headache real quick), either, since the human eye uses
essentially the same trick the Bayer sensors do (full color reconstructed
from multiple RGB sensors).
Basically, as long as you are producing images for human consumption, Foveon
offers no practical advantage.
David J. Littleboy
Tokyo, Japan
Percy D
>In message <43c9a3bc$0$11071$e4fe...@news.xs4all.nl>, "Bart van der Wolf"
Several months ago, someone posted this link to a site with some
scenery shots from a Sigma SD 10. Maybe someone could share some of
their samples that they think are better taken with their favorite
camera.
http://www.pbase.com/ghoerdt/panorama_cloudy_switzerland
Peter A. Stavrakoglou
news:GFkyf.10803$V.1102@fed1read04...
I don't think so, the post insinutate that there are none which is not
correct.
Peter A. Stavrakoglou
news:ovljs1tsk0t4d2797...@4ax.com...
I got the info directly from Sigma USA from two sources, one of them in the
#2 position at Sigma USA. Call them up - they'll tell you that the SD10 is
now discontinued and that a new DSLR will be released this year.
Peter A. Stavrakoglou
If your perception is that there are no professional photographers who use a
Sigma DSLR than your perception is incorrect.
Edmund
I have to admit I don't know what that Nyquist frequency means in
pictures. I do know the about Shannon/Nyquist in digital audio.
The following comes from: dpreview.com and doesn't help me much
in understanding.
The pictures however showing very good detail from de foveon chip.
In some parts of the pictures ( especially with the newer software),
the detail seems better better then the 6MP canon.
----------------------------Begin
* Nyquist frequency defined as the highest spatial frequency where the sensor
can still
faithfully record image detail. Beyond the Nyquist frequency aliasing occurs.
As we expected identical results to the SD9, and with good reason considering
the design of the sensor and the fact that the pixel count hasn't changed.
That's to a lack of a color filter the X3 sensor doesn't suffer from color
moire at the limits of resolution.
Because the X3 sensor doesn't employ an anti-alias (low pass / blur) filter it
continues to deliver detail past Nyquist. Taking the vertical resolution bar as
an example we can count all nine lines up to our 'absolute resolution'
measurement of around 1550 LPH, after this point lines begin to merge and by
2000 LPH we can count five obviously combined lines. In a real image this could
be the detail of leaves on a distant tree or bricks on a distant wall. At the
time of my SD9 review there was much debate over the 'validity' of this extra
detail. My opinion on this matter is that this detail is at least of
photographic merit, it is the correct color and represents detail that the
human eye although not able to distinguish perfectly would also see as a broken
texture, it's certainly better than the blurred area we would get from the
anti-alias filter of a Bayer sensor camera.
--------------------End
>
> The reason people think Foveon outresolves Bayer is that the Sigma Foveon
> cameras don't have the (mathematically) required (for correct imaging)
> antialiasing filter; if they did, the bogus response that the Foveon fans
> think is resolution would be suppressed, and they'd look like any other
> 3.4MP camera.
>
> > The so called Foveon sensor has 3 layers of 3.4 MP
> > So each sensor (from the 3.4MP ) gives a real existing pixel.
>
> That doesn't enable them to _correctly_ resolve any higher frequencies
> (relative to the pixel spacing) than Bayer sensors for monochrome test
> patterns. Essentially, there's zero advantage.
I can understand the frequencies relative to the pixel spacing part,
but how about the bayer interpolation that has to guess the color of
the pixels?
Like here http://www.foveon.com/article.php?a=70
>
> The place that there is an advantage is in test patterns that are red on
> blue instead of black and white. But people can't see those patterns (if
> you try, you'll get a headache real quick), either, since the human eye uses
> essentially the same trick the Bayer sensors do (full color reconstructed
> from multiple RGB sensors).
>
> Basically, as long as you are producing images for human consumption,
> Foveon offers no practical advantage.
I just wonder if there is a theoretical advantage, the whole
concept make sense to me. Maybe it needs to be refined a bit.
Bill Funk
The *final image* from the X3 has more pixels, because the camera
upsamples from 3.4 to 10+MP. The X3 sensor has 3.4MP.
We can't (or at the very least, shouldn't) take Foveon's word for the
fact that those images actually represent the same results anyone else
would get with the hardware they used. Foveon doesn't even say what
Bayer-filtered sensor they used.
Foveon is the company that makes the X3 sensor. Marketing being what
it is, claims made by Foveon, especially claims made with unspecified
hardware, must be seen as claims, not hard facts.
Bill Funk
>In message <9qjjs1ta0vuq3logf...@4ax.com>, Bill Funk wrote:
>> On Sat, 14 Jan 2006 17:44:24 -0000, "Gormless" <sim...@gormless.com>
>> wrote:
>>
>> >
>> >"Peter A. Stavrakoglou" <nto...@optonline.net> wrote in message
>> >news:rQ8yf.842$IM2...@fe12.lga...
>> >> it's 3.4 MP
>> >> sensor resolves as well as any 6MP Bayer, sometimes even better.
>> >>
>> >
>> >I don't understand how that is a technical possibility.
>> >Helen
>> >
>> Technically, of course, it's not.
>
>Well that is open to discussion, correct me if I am wrong.
>to start with : "whats is a name"
>Any "bayer" 6MP camera has only 6MP sensors total
>to be divided for all colors.
>For each layer the bayer interpolation comes in to
>guess the missing pixels. That works remarkably well,
>especially in gradually changing colors.
What "missing" pixels? A 6MP Bayer sensor outputs 6MP images. No
missing pixels.
What Bayer-filtered sensors do is send 6MP of data to the camera's
firmware, which interpolates the chroma data to make the colors; the
luminance is used very much as-is.
>It goes wrong in sharp contrasts between black and white.
>The so called Foveon sensor has 3 layers of 3.4 MP
>So each sensor (from the 3.4MP ) gives a real existing pixel.
>
Yes, that's 3.4MP. Not 10+, as many claim.
It uses three layers of sensors, all in the same place on the image
plane. That's 3.4 different, individual, points on the image plane,
while that 6MP Bayer-filtered sensor has 6 million different
individual points on the image plane.
Bill Funk
"Guessing" is what you'd do if someone showed you a featureless box,
and asked you to guess what's inside. It could be solid, liquid, or
gas. Animal, vegetable, or mineral. You have no idea, so you *guess*.
However, Bayer-filtered sensor cameras don't *guess* about colors;
they folow rules in the firmware to determine what each pixel's color
is. If they did guess, the colors would bear no resemblence to what
the lens saw.
The word "guess" is used wrongly in this context.
>>
>> The place that there is an advantage is in test patterns that are red on
>> blue instead of black and white. But people can't see those patterns (if
>> you try, you'll get a headache real quick), either, since the human eye uses
>> essentially the same trick the Bayer sensors do (full color reconstructed
>> from multiple RGB sensors).
>>
>> Basically, as long as you are producing images for human consumption,
>> Foveon offers no practical advantage.
>
>I just wonder if there is a theoretical advantage, the whole
>concept make sense to me. Maybe it needs to be refined a bit.
I agree; the concept makes sense. So far, the application has not
borne out the potential.
And, as far as I can see, neither Foveon nor Sigma has made any steps
towards advancing the technology lately.
>>
>> David J. Littleboy
>> Tokyo, Japan
>>
>>
>>
>
--
Mark˛
Whatever the possibility of insinuation, the key question should probably
be: Why aren't there large numbers of Sigma cameras being used by pros (and
others)? There are several possible reasons, many of which COULD have
little to do with body deficiencies (lack of existing pro arsenals of
Sigma-mount lenses...who owned those??). But there are other possibles
reasons which could relate directly to foveon deficiencies. It is in the
REASON BEHIND the distinctly rare presence of Sigma bodies that forms the
significance of the question as to professional use.
Since it is impossible to reliably describe or know the general cause of
this general absence (not 100% absence...but general absence), we could go
round and round in circles, forever ascribing the cause to suit whichever
side one is on in this discussion.
Personally, I am convinced that the Sigmas simply don't offer what most pros
need...even if they DID have an arsonal of Sigma glass (which has its own
list of reasons for not being the reality). The causes for that are equally
open to wildly differing explanations, depending on one's bias/experience,
etc.
Sigma has some particular strenghts...yes.
But until those strengths are no longer counter-balanced by significant
negatives (color rendition questions, low light performance, texture
issues...etc.), I think it is a fair assessment to say that it's not
reaching general professional acceptance because of it's own deficiencies.
Again...note the word "general." I know there may be some pros using it,
but that's not really key to the question. Besides...some of the worst
photography I've ever seen was created by shooters who were paid to
shoot...and were then deemed, "pro." :) So... Are there "pros" using a
particular camera? Great! So what. Lets see the images. When the images
don't exhibit what is needed, in the fullness of general professional
requirements...it gets left in the dust.
I do think its significant that production has stopped...LONG before a
replacement is available.
This is very different from most camera manufacturing patterns. With Canon
and Nikon, for example...it is rarely the lacking sales of manufacture of a
current model that leads to the new. Rather, it is the effort of those
manufacturers to out-do not only the competition...but also out-do their
existing, for-sale, current model...that prompts them to release new stuff.
In the case of Sigma, it appears that they ceased production because they
just couldn't sell them...and they're now trying to produce a more generally
saleable product.
-Mark
Bill Funk
Would you accept that the number of pros using SDx cameras is very
small compare to Nikon and Canon users?
Mark˛
"No professionals" in an absolute sense is incorrect. Right.
But "no professionals" in a coparative sense, is indeed correct.
Ya, there are pros using them.
Are there substantial enough numbers using them to indicate gernal
confidence in them? No, there are not.
So in these two ways, you are both correct.
To me, the more important question is the latter.
-Mark
Mark˛
That would have little to do with acceptance...since it is a fact.
:)
Bill Funk
<nto...@optonline.net> wrote:
>"Bill Funk" <Big...@pipping.com.com> wrote in message
>news:ovljs1tsk0t4d2797...@4ax.com...
>> On Sat, 14 Jan 2006 19:00:51 -0500, "Peter A. Stavrakoglou"
>> <nto...@optonline.net> wrote:
>>
>>>> and Canon RebelXT/350D, and seriously dusted off by cameras like the
>>>> Nikon
>>>> D200 (at the price point the Sigma SD-10 was introduced) and the
>>>> admittedly much more expensive Canon 5D.
>>>
>>>C'mon Skip, comparing current DSLR prices with one that is over two years
>>>old. The SD10 is much less than a D200, half the price or less, if you
>>>can
>>>find one. It's been discontinued since there's a new one on the way.
>>>
>>
>> You said this before; can you point to any online rumors/reports of an
>> upcoming SD camera using Foveon technology?
>> Does this mean a new Foveon sensor?
>> With (hopefully) an AA filter? :-)
>
>I got the info directly from Sigma USA from two sources, one of them in the
>#2 position at Sigma USA. Call them up - they'll tell you that the SD10 is
>now discontinued and that a new DSLR will be released this year.
>
Give me his number.
Neither Sigma nor Foveon has given any information about this at all
that I can find. No indication that any advancement of this technology
will bear any fruit in the near future. Not a peep on either Foveon's
nor Sigma's sites.
I'm sure you'll understand if I continue to wonder if there will be
another SD camera any time soon.
I'm not anti-Sigma or -Foveon; I like the terchnology Foveon has come
up with. I think competition would be good; both types of sensor would
benefit from competition, with the users benefiting, too.
But it does not seem to me that Foveon and Sigma are in any hurry at
all in marketing any advances, and certainly aren't interested in
advertising any advantages that may have come up with; a very peculiar
thing for tech companies. The fairly obvious conclusion is that the
advances aren't there.
Gormless
> >>
> > Without suggesting you're a liar - I'm sure you're not - my own personal
> > and
> > extensive experience indicates that your perception of the truth in this
> > particular matter is not equal to mine.
> > Helen
>
> If your perception is that there are no professional photographers who use
a
> Sigma DSLR than your perception is incorrect.
>
the sum of what I have subjectively perceived. I have personally never
visually perceived any professional photographer with a Sigma camera. Put
it simply - I ain't seen one.
Helen
Bart van der Wolf
"Edmund" <nom...@hotmail.com> wrote in message
news:dqdiut$4pd$1...@azure.qinip.net...
SNIP
I've never seen such poor Bayer CFA results as on that page, but then
the page is supposed to make the Foveon look good, so I guess anything
is allowed ...
I'm not saying it is impossible to get CFA demosaicing artifacts, but
a lot depends on the Raw converter used. Most modern Raw converters
have a False Color removal option, either user selectable or
automatic.
Bart
David J. Littleboy
"Edmund" <nom...@hotmail.com> wrote:
>
> Because the X3 sensor doesn't employ an anti-alias (low pass / blur)
> filter it continues to deliver detail past Nyquist.
Note that this is simply a mathematical impossibility. The digital file
can't even represent such detail. What it delivers is bogus; in the test
charts, it shows patterns in the image with fewer lines than the test chart
has. The mathematical term for this is "aliasing".
>> That doesn't enable them to _correctly_ resolve any higher frequencies
>> (relative to the pixel spacing) than Bayer sensors for monochrome test
>> patterns. Essentially, there's zero advantage.
>
> I can understand the frequencies relative to the pixel spacing part,
> but how about the Bayer interpolation that has to guess the color of
> the pixels?
> Like here http://www.foveon.com/article.php?a=70
If you read marketing hype, you'll get confused. Especially if you don't
understand the math.
>> The place that there is an advantage is in test patterns that are red on
>> blue instead of black and white. But people can't see those patterns (if
>> you try, you'll get a headache real quick), either, since the human eye
>> uses
>> essentially the same trick the Bayer sensors do (full color reconstructed
>> from multiple RGB sensors).
>>
>> Basically, as long as you are producing images for human consumption,
>> Foveon offers no practical advantage.
>
> I just wonder if there is a theoretical advantage, the whole
> concept make sense to me. Maybe it needs to be refined a bit.
There's no "theoretical" advantage. Theory is quite clear. Correct imaging
requires antialiasing. All Foveon gives you is the ability to handle images
that both Bayer and the human eye have trouble with.
The problem is that people don't understand how good Bayer actually is. It
basically gives you color for free. The only way Foveon manages to look
somewhat better is by not including the mathematically required AA filter,
and thus allowing aliasing. It's cynical marketing taking advantage of
people's mathematical naiveté.
Peter A. Stavrakoglou
Call Sigma USA, the number's on their site. Ask them.
Peter A. Stavrakoglou
Of course, but to state that there are *no* pros using them is untrue.
Peter A. Stavrakoglou
Production has stopped yes, but you cannot with certainty say that it is
LONG before a replacement is available, unless you have some inside
information you're not sharing. A replacement could be available next
month. When the SD10 was announced it was available for sale, not months
later. I suspect the same will happen with its successor.
Mark˛
For the sake of argument, yes. That's a possibility. But again, this
wouldn't fit with the patterns of any other major commpanies. Rumors of
replacements...and even the official announcements... always come months (or
longer) before the the current model becomes scarce. And even if it IS
scarce...the cause is attributed to easy sales of it. There may still be
Sigma bodies for sale, but not because of huge production, rather a lack of
demand for them.
>A replacement could be available next
> month. When the SD10 was announced it was available for sale, not
> months later.
Right. And that particular upgrade indeed fit the normal pattern.
The breakdown came AFTER the SD10, and is what I am commenting on.
I think their hope was that the SD10 would be their break-throgh product.
It wasn't. The foceon remained a fringe item, that still wasn't ready for
general consumption.
>I suspect the same will happen with its successor.
That's no possible, since there has been no indication of a successor.
While I have little doubt another offering will arrive...by comments about a
pattern "gap" indicating trouble...still holds water at this point.
Honestly, I'd love to see Foveon come out with the "killer sensor (killer
app)!" As I've ALWAYS maintained, this could only be a good thing for all
of us...even those who never adopt it. Competition is good, so I will
continue to hope Foveon has a blockbuster. I don't see it in the works,
though, even if the idea was/is interesting and innovative. I'd love to be
proven wrong in my assessment.
-Mark
Mark˛
Right. Your are being literal in the conversation, rather than seeing his
point...which was..."Pros, in general...though not entirely...have rejected
(or at least failed to widely adopt) the Foveon at this point."
Mark˛
> "Edmund" <nom...@hotmail.com> wrote:
>> I just wonder if there is a theoretical advantage, the whole
>> concept make sense to me. Maybe it needs to be refined a bit.
>
> There's no "theoretical" advantage. Theory is quite clear. Correct
> imaging requires antialiasing.
This is it in a nutshell.
Foveon fails whenever there are visual elements that require representation
in the image...but that do not occupy more than a single pixel either in its
full dimension, or in the transition to other colors in adjacent pixels.
This would explain many (if not all) of the strange texture renditions, and
odd-lloking color transitions.
>All Foveon gives you is the ability to
> handle images that both Bayer and the human eye have trouble with.
I suspect the Foveon might make a very decent astro photography
camera...much like the 20Da...which intentionally lacks the AA filter.
> The problem is that people don't understand how good Bayer actually
> is. It basically gives you color for free. The only way Foveon
> manages to look somewhat better is by not including the
> mathematically required AA filter, and thus allowing aliasing. It's
> cynical marketing taking advantage of people's mathematical naiveté.
Bing!
:)
-Mark
Mark˛
Sigma very intentionally selected examples of weakness in bayer sensors,
including a very careful selection of odd color-combinations that don't lead
to Foveon troubles. Note the lack of skin tone inclusion, and patterns that
can be distinctly used to show realism in naturally occurring items (rocks,
trees, skin).
The marketers at Foveon are VERY good, and also very manipulative.
You can hardly blame them. They are, after all, tasked with selling the
product...not researching it's full, generalized effectiveness.
-Mark
Mark˛
Exactly NONE of those (beautiful) scenes offer the ability to judge the
Sigma problem areas (skin tones, and textures that are rendered closely
enough for surface rendition (tree bark, close boulders, etc.). When you
have such tiny stiched photos, it is quite impossible to judge detail, since
downsizing has eliminated differences that might otherwise be notable.
Skip M
>>
>> "Peter A. Stavrakoglou" <nto...@optonline.net> wrote in message
>> news:%Vfyf.978$IM2...@fe12.lga...
>> SNIP
>> > The Foveon can resolve detail that a Bayer sensor cannot.
>>
>> That "detail" is called "aliasing artifacts" and they have little to
>> do with actual detail. They are larger than life aberrations caused by
>> the simple inability of any(!) discrete sampling device (like a photo
>> sensitive sensor element) to reliably reproduce detail that's smaller
>> than 1 pixel, unless an anti-aliasing filter is used in front of the
>> sensor.
>>
>> Bart
>
> Like to see your comment on this site.
>
> http://www.foveon.com/article.php?a=70
>
>
that tells us nothing about the number of pixels. The next phrase says the
"X3 direct image sensor has three times as many pixels" as the Bayer, or
"Mosaic" as they call it. So, that is either a 1.1 mp sensor, given that
the Foveon is strictly a 3.4mp sensor, or it is a 3.4 mp sensor itself,
given Foveon's propensity to claim 10.2 mp for themselves. Whatever the
case, I've not seen a DSLR produce images as bad as the ones they present,
all in the name of marketing.
--
Skip Middleton
http://www.shadowcatcherimagery.com
Matt Ion
> As far as I understand things, this foveon sensor
> looks the best technology to me.
> Why is it hardly anybody uses it?
Because nobody keeps domesticated Foveons anymore?
---
avast! Antivirus: Outbound message clean.
Virus Database (VPS): 0603-0, 01/15/2006
Tested on: 1/15/2006 7:44:38 PM
avast! - copyright (c) 1988-2005 ALWIL Software.
http://www.avast.com
Skip M
news:PEEyf.354898$ki.151325@pd7tw2no...
> Edmund wrote:
>> As far as I understand things, this foveon sensor
>> looks the best technology to me.
>> Why is it hardly anybody uses it?
>
> Because nobody keeps domesticated Foveons anymore?
>
>
Yeah, it went the way of the Elusive Nauga, which cut off the supply of
American Naugahyde...
;-)
Percy D
Who cares about skin tones in a scenery shot. Maybe someone can offer
a sample of similar "tiny stitched photos" of scenery with another
camera for comparison.
Ray Fischer
>"Gormless" <sim...@gormless.com> wrote in message
>> > sensor resolves as well as any 6MP Bayer, sometimes even better.
>> >
>>
>> I don't understand how that is a technical possibility.
>> Helen
>
False.
--
Ray Fischer
rfis...@sonic.net
Ray Fischer
>As far as I understand things, this foveon sensor
>looks the best technology to me.
>Why is it hardly anybody uses it?
Because it's great in theory but problematic in practice.
In general you _can_ separate colors by how far they penetrate the
sensor. That's the principle behind the technology. The problem
is that it's really hard to control the variables. The differences
in the amount of penetration of the different light frequencies isn't
that great and small differences tend to get magnified. The result is
somewhat uneven and occasionally blotchy color.
Bayer sensors, while someone lower in sharpness, have the benefit that
the color filters in front of each sensor are easier to control and so
the light level isn't attenuated as much and the color is more
accurate and more consistent.
Besides, people are happy with Bayer sensors. They work. They work
well. And they're cheaper to make.
--
Ray Fischer
rfis...@sonic.net
Mark˛
You missed the point entirely.
If I have to explain it, perhaps its not worth effort...
It's sort of like testing the abilities of a ferrari by driving 55 mph in
the slow lane.
Within that limitation, a Ferrari acts rather the same as a VW
Beetle...except that the Beetle probably rides a little nicer.
Point?
The sample you point to doesn't allow scrutiny that relates to issues facing
Foveon.
David J. Littleboy
Bad example. Someone tries to sell you a hand-made car (Stigma) telling you
that it performs like a ferrari (6MP dSLR). Testing at 55 in the slow lane
doesn't tell you if it does.
Dave Martindale
>I have to admit I don't know what that Nyquist frequency means in
>pictures. I do know the about Shannon/Nyquist in digital audio.
It's almost the same. In digital audio, the independent variable is
time, the sampling rate is in samples/second, and the Nyquist limit is
the point at which the signal frequency is half the sample frequency.
In other words, you need more than two samples per cycle to resolve
a signal without aliasing.
In imaging, the independent variable is space, and the signal is 2D
instead of 1D. So the sample rate is in samples per mm or samples per
inch, and the Nyquist limit is the point at which the signal frequency
is half the sample frequency. In other words, the wavelength is twice
the pixel spacing at the Nyquist limit, or you need more than two pixels
per cycle to resolve a signal without aliasing.
>As we expected identical results to the SD9, and with good reason considering
>the design of the sensor and the fact that the pixel count hasn't changed.
>That's to a lack of a color filter the X3 sensor doesn't suffer from color
>moire at the limits of resolution.
The X3 sensor doesn't suffer from colour moire on B&W images, and that is
a real advantage. But it *does* suffer from B&W moire on fine B&W patterns
because it lacks an anti-aliasing filter, and that's just incorrect.
>Because the X3 sensor doesn't employ an anti-alias (low pass / blur) filter it
>continues to deliver detail past Nyquist. Taking the vertical resolution bar as
>an example we can count all nine lines up to our 'absolute resolution'
>measurement of around 1550 LPH, after this point lines begin to merge and by
>2000 LPH we can count five obviously combined lines. In a real image this could
>be the detail of leaves on a distant tree or bricks on a distant wall. At the
>time of my SD9 review there was much debate over the 'validity' of this extra
>detail. My opinion on this matter is that this detail is at least of
>photographic merit, it is the correct color and represents detail that the
>human eye although not able to distinguish perfectly would also see as a broken
>texture, it's certainly better than the blurred area we would get from the
>anti-alias filter of a Bayer sensor camera.
That's a matter of taste. I, at least, would prefer that a camera
render detail too fine to resolve as the average colour, not as some
different detail that the camera would have been able to resolve had the
image contained that. I never want 9 pickets in my picket fence imaged
as 5 boards - ever. The eye blurs detail it can't resolve, and so does
film. The aliased images from the Sigma camera just look unnatural to
me.
>I can understand the frequencies relative to the pixel spacing part,
>but how about the bayer interpolation that has to guess the color of
>the pixels?
>Like here http://www.foveon.com/article.php?a=70
Calling it "estimating" would be more accurate than "guessing". But
Bayer colour estimation works remarkably well in practice for most
subjects on a modern camera; it requires some effort to find images
where it will fail.
>I just wonder if there is a theoretical advantage, the whole
>concept make sense to me. Maybe it needs to be refined a bit.
There is a theoretical advantage, if the 3 layers of the sensor have
similar colour response and sensitivity to the filtered cells in a Bayer
sensor. And in fact some video cameras use 3 CCDs and a prism colour
separation block, and Foveon themselves once sold such a camera. This
design is expensive but works very well. The Foveon X3 is an attempt to
get the same advantages from a single flat sensor - but so far it has
not worked very well.
Dave
Mark˛
Ya, OK...the same example, but in reverse.
Thought that's what I said, but you said what I meant to say better than how
I said what I said when I said it the first time it was said.
:)
Edmund
Well I know thare is at least one out there :-)
http://www.foveon.com/gallery.php the picture of the child
in made by him. Not bad in detail IMHO.
Edmund
> On Sun, 15 Jan 2006 14:14:59 +0000, Edmund <nom...@hotmail.com> wrote:
>
> >In message <9qjjs1ta0vuq3logf...@4ax.com>, Bill Funk wrote:
> >> On Sat, 14 Jan 2006 17:44:24 -0000, "Gormless" <sim...@gormless.com>> >> >news:rQ8yf.842$IM2...@fe12.lga...
> >> >> it's 3.4 MP
> >> >> sensor resolves as well as any 6MP Bayer, sometimes even better.
> >> >>
> >> >
> >> >I don't understand how that is a technical possibility.
> >> >Helen
> >> >
> >
> >Well that is open to discussion, correct me if I am wrong.
> >to start with : "whats is a name"
> >Any "bayer" 6MP camera has only 6MP sensors total
> >to be divided for all colors.
> >For each layer the bayer interpolation comes in to
> >guess the missing pixels. That works remarkably well,
> >especially in gradually changing colors.
>
> What "missing" pixels?
For instance, green ones on the place where there is NO green
sensor, and the red ones on places where are no
red sensors.
>A 6MP Bayer sensor outputs 6MP images. No
> missing pixels.
That is because the bayer interpolation, interpolates
the assumed color of "missing" pixels.
That is what it puts out.
> What Bayer-filtered sensors do is send 6MP of data to the camera's
> firmware, which interpolates the chroma data to make the colors; the
> luminance is used very much as-is.
> >It goes wrong in sharp contrasts between black and white.
> >The so called Foveon sensor has 3 layers of 3.4 MP
> >So each sensor (from the 3.4MP ) gives a real existing pixel.
> >
> Yes, that's 3.4MP. Not 10+, as many claim.
Yes that is true, and it is a matter of what definition one
use, I prefer to call it 3.4 MP full color.
In that Bayer case there is on a 6MP camera a
6MP full color but most of these pixels are guessed.
> It uses three layers of sensors, all in the same place on the image
> plane. That's 3.4 different, individual, points on the image plane,
> while that 6MP Bayer-filtered sensor has 6 million different
> individual points on the image plane.
Right, but not for all colors.
Suppose the next numbers represent bayer sensors and a picture
must be taken from a chart which has this resolution too.
The colors of that chart are
1= red 2 white 3= black 4= white
1234567
Then a bayer interpolation will never produce an output with
pixel 2 white, which it must be.
The Foveon sensor however can produce the white here.
It is a bit oversimplified but that is how it works as
far as I understand.
Edmund
> Edmund <nom...@hotmail.com> writes:
>
> >I have to admit I don't know what that Nyquist frequency means in
> >pictures. I do know the about Shannon/Nyquist in digital audio.
>
> It's almost the same. In digital audio, the independent variable is
> time, the sampling rate is in samples/second, and the Nyquist limit is
> the point at which the signal frequency is half the sample frequency.
> In other words, you need more than two samples per cycle to resolve
> a signal without aliasing.
At least exactly 2 samples per cycle in fact.
>
> In imaging, the independent variable is space, and the signal is 2D
> instead of 1D. So the sample rate is in samples per mm or samples per
> inch, and the Nyquist limit is the point at which the signal frequency
> is half the sample frequency. In other words, the wavelength is twice
> the pixel spacing at the Nyquist limit, or you need more than two pixels
> per cycle to resolve a signal without aliasing.
Thank for explaining, can I find more info on this on the Internet?
I like to know what kind of errors are made without these anti-aliasing.
In audio e.g. the SACD they just sample very high and don't use anti
aliasing anymore.
You are right, estimating is a much better word, but hey
I am Dutch :-)
>
> But
> Bayer colour estimation works remarkably well in practice for most
> subjects on a modern camera; it requires some effort to find images
> where it will fail.
>
> >I just wonder if there is a theoretical advantage, the whole
> >concept make sense to me. Maybe it needs to be refined a bit.
>
> There is a theoretical advantage, if the 3 layers of the sensor have
> similar colour response and sensitivity to the filtered cells in a Bayer
> sensor. And in fact some video cameras use 3 CCDs and a prism colour
> separation block, and Foveon themselves once sold such a camera. This
> design is expensive but works very well. The Foveon X3 is an attempt to
> get the same advantages from a single flat sensor - but so far it has
> not worked very well.
That popped to my mind too, to use a prism for this purpose.
But that is already done, OK.
>
> Dave
Peter A. Stavrakoglou
news:OmDyf.12423$V.6095@fed1read04...
Isn't "he" a "she"?
Edmund
>
> "Edmund" <nom...@hotmail.com> wrote:
> >
> > Because the X3 sensor doesn't employ an anti-alias (low pass / blur)
> > filter it continues to deliver detail past Nyquist.
>
> Note that this is simply a mathematical impossibility. The digital file
> can't even represent such detail. What it delivers is bogus; in the test
> charts, it shows patterns in the image with fewer lines than the test chart
> has. The mathematical term for this is "aliasing".
I appreciate what you are saying I just quoted dpreview here.
Apart from that, I nowhere saw any detail on from a Foveon
sensor that wasn't on the chart.
Do you have such examples?
>
> The problem is that people don't understand how good Bayer actually is. It
> basically gives you color for free. The only way Foveon manages to look
> somewhat better
I interrupt you right here, because that is what it's all about, "better
looking pictures."
> is by not including the mathematically required AA filter,
> and thus allowing aliasing. It's cynical marketing taking advantage of
> people's mathematical naiveté.
Maybe it comes down to what extent the picture is disturbed by
allowing aliasing or by the AA filter.
Maybe the need for AA is ( much ) bigger in Bayer configuration
then with Foveon?
Percy D
>>>> http://www.pbase.com/ghoerdt/panorama_cloudy_switzerland
>>>
>>> Exactly NONE of those (beautiful) scenes offer the ability to judge
>>> the Sigma problem areas (skin tones, and textures that are rendered
>>> closely enough for surface rendition (tree bark, close boulders,
>>> etc.). When you have such tiny stiched photos, it is quite
>>> impossible to judge detail, since downsizing has eliminated
>>> differences that might otherwise be notable.
>>>
>> Who cares about skin tones in a scenery shot. Maybe someone can offer
>> a sample of similar "tiny stitched photos" of scenery with another
>> camera for comparison.
>
>You missed the point entirely.
>
>If I have to explain it, perhaps its not worth effort...
>
>It's sort of like testing the abilities of a ferrari by driving 55 mph in
>the slow lane.
>Within that limitation, a Ferrari acts rather the same as a VW
>Beetle...except that the Beetle probably rides a little nicer.
>
>Point?
>The sample you point to doesn't allow scrutiny that relates to issues facing
>Foveon.
am saying is that it is capable of some impresive performance in
skillfull hands. To my un-educated eye, they look almost as good as
something that might be seen in National Geographic or Arizona
Highways mag.
Bill Funk
>In message <79cls11icoge8mi3t...@4ax.com>, Bill Funk wrote:
>> On Sun, 15 Jan 2006 14:14:59 +0000, Edmund <nom...@hotmail.com> wrote:
>>
>> >In message <9qjjs1ta0vuq3logf...@4ax.com>, Bill Funk wrote:
>> >> On Sat, 14 Jan 2006 17:44:24 -0000, "Gormless" <sim...@gormless.com>
>> >> wrote:
>> >>
>> >> >
>> >> >"Peter A. Stavrakoglou" <nto...@optonline.net> wrote in message
>> >> >news:rQ8yf.842$IM2...@fe12.lga...
>> >> >> it's 3.4 MP
>> >> >> sensor resolves as well as any 6MP Bayer, sometimes even better.
>> >> >>
>> >> >
>> >> >I don't understand how that is a technical possibility.
>> >> >Helen
>> >> >
>> >> Technically, of course, it's not.
>> >
>> >Well that is open to discussion, correct me if I am wrong.
>> >to start with : "whats is a name"
>> >Any "bayer" 6MP camera has only 6MP sensors total
>> >to be divided for all colors.
>> >For each layer the bayer interpolation comes in to
>> >guess the missing pixels. That works remarkably well,
>> >especially in gradually changing colors.
>>
>> What "missing" pixels?
>
>For instance, green ones on the place where there is NO green
>sensor, and the red ones on places where are no
>red sensors.
That's chroma data, not pixels.
And 'missing' is maybe the wrong word; it's determined in the
interpolation.
>
>>A 6MP Bayer sensor outputs 6MP images. No
>> missing pixels.
>
>That is because the bayer interpolation, interpolates
>the assumed color of "missing" pixels.
>That is what it puts out.
Chroma, not luminance. You seem to be confusng chroma data with pixel.
>
>> What Bayer-filtered sensors do is send 6MP of data to the camera's
>> firmware, which interpolates the chroma data to make the colors; the
>> luminance is used very much as-is.
>
>> >It goes wrong in sharp contrasts between black and white.
>> >The so called Foveon sensor has 3 layers of 3.4 MP
>> >So each sensor (from the 3.4MP ) gives a real existing pixel.
>> >
>> Yes, that's 3.4MP. Not 10+, as many claim.
>
>Yes that is true, and it is a matter of what definition one
>use, I prefer to call it 3.4 MP full color.
>
>In that Bayer case there is on a 6MP camera a
>6MP full color but most of these pixels are guessed.
No, not "guessed".
You guess whenyou have no idea of the answer. Tjhe color ion a Bayer
sensor isn't guessed, it's interpolated using data from other,
adjacent, sensels.
You would have to "guess" what's in a featureles box. Thge camera
doesn't have to "guess", because it has actual data to go on.
>
>> It uses three layers of sensors, all in the same place on the image
>> plane. That's 3.4 different, individual, points on the image plane,
>> while that 6MP Bayer-filtered sensor has 6 million different
>> individual points on the image plane.
>
>Right, but not for all colors.
Exactly.
>
>Suppose the next numbers represent bayer sensors and a picture
>must be taken from a chart which has this resolution too.
>The colors of that chart are
>1= red 2 white 3= black 4= white
>
>1234567
>
>Then a bayer interpolation will never produce an output with
>pixel 2 white, which it must be.
>The Foveon sensor however can produce the white here.
>
>It is a bit oversimplified but that is how it works as
>far as I understand.
On an individual, one-pixel level, yes.
But we don't see that way.
That one small white dot wouldn't be seen by the eye either, so Bayer
interpolation works very well.
OTOH, if the Foveon sensor showed that white pixel, the eye would see
it as false information, because the eye wouldn't see it in the
original scene. That's the problem with the lack of an AA filter;
false information that's claimed as better detail. That detail is
false detail.
--
Bill Funk
Replace "g" with "a"
funktionality.blogspot.com
Bill Funk
number here)@cox..net> wrote:
>Bill Funk wrote:
>> On Sun, 15 Jan 2006 12:44:47 -0500, "Peter A. Stavrakoglou"
>> <nto...@optonline.net> wrote:
>>
>>> "Gormless" <sim...@gormless.com> wrote in message
>>> news:MIudnZSXSYz...@pipex.net...
>>>>
>>>> "Peter A. Stavrakoglou" <nto...@optonline.net> wrote in message
>>>> news:K1kyf.5573$Aq5...@fe08.lga...
>>>>
>>>>>> Again, and for the record, during over 7 years of varied
>>>>>> professional photography in both the UK and the USA I am yet to
>>>>>> see or hear of a pro using Sigma cameras. Lenses yes, I'll grant
>>>>>> you, but not a camera. As I apparently correctly predicted above
>>>>>> - "though its adherents will swear differently".
>>>>>
>>>>> So I'm lying?
>>>>>
>>>> Without suggesting you're a liar - I'm sure you're not - my own
>>>> personal and
>>>> extensive experience indicates that your perception of the truth in
>>>> this particular matter is not equal to mine.
>>>> Helen
>>>
>>> If your perception is that there are no professional photographers
>>> who use a Sigma DSLR than your perception is incorrect.
>>>
>>>
>> Would you accept that the number of pros using SDx cameras is very
>> small compare to Nikon and Canon users?
>
>That would have little to do with acceptance...since it is a fact.
>:)
>
I was responding to Peter, who has lapsed into a semantic argument.
Bill Funk
<nto...@optonline.net> wrote:
OK, we've solved the semantics problem.
Bill Funk
<nto...@optonline.net> wrote:
Well, *I'm* a he.
Who are you referring to?
Bill Funk
<nto...@optonline.net> wrote:
>"Bill Funk" <Big...@pipping.com.com> wrote in message
>news:rhdls1158tk56mn5p...@4ax.com...
>> On Sun, 15 Jan 2006 12:36:11 -0500, "Peter A. Stavrakoglou"
>> <nto...@optonline.net> wrote:
>>
>>>"Bill Funk" <Big...@pipping.com.com> wrote in message
>>>news:ovljs1tsk0t4d2797...@4ax.com...
>>>> On Sat, 14 Jan 2006 19:00:51 -0500, "Peter A. Stavrakoglou"
>>>> <nto...@optonline.net> wrote:
>>>>
>>>>>> and Canon RebelXT/350D, and seriously dusted off by cameras like the
>>>>>> Nikon
>>>>>> D200 (at the price point the Sigma SD-10 was introduced) and the
>>>>>> admittedly much more expensive Canon 5D.
>>>>>
>>>>>C'mon Skip, comparing current DSLR prices with one that is over two
>>>>>years
>>>>>old. The SD10 is much less than a D200, half the price or less, if you
>>>>>can
>>>>>find one. It's been discontinued since there's a new one on the way.
>>>>>
>>>>
>>>> You said this before; can you point to any online rumors/reports of an
>>>> upcoming SD camera using Foveon technology?
>>>> Does this mean a new Foveon sensor?
>>>> With (hopefully) an AA filter? :-)
>>>
>>>I got the info directly from Sigma USA from two sources, one of them in
>>>the
>>>#2 position at Sigma USA. Call them up - they'll tell you that the SD10
>>>is
>>>now discontinued and that a new DSLR will be released this year.
>>>
>> Give me his number.
>> Neither Sigma nor Foveon has given any information about this at all
>> that I can find. No indication that any advancement of this technology
>> will bear any fruit in the near future. Not a peep on either Foveon's
>> nor Sigma's sites.
>>
>> I'm sure you'll understand if I continue to wonder if there will be
>> another SD camera any time soon.
>>
>> I'm not anti-Sigma or -Foveon; I like the terchnology Foveon has come
>> up with. I think competition would be good; both types of sensor would
>> benefit from competition, with the users benefiting, too.
>> But it does not seem to me that Foveon and Sigma are in any hurry at
>> all in marketing any advances, and certainly aren't interested in
>> advertising any advantages that may have come up with; a very peculiar
>> thing for tech companies. The fairly obvious conclusion is that the
>> advances aren't there.
>
>Call Sigma USA, the number's on their site. Ask them.
>
And I wouild fully expect Foveon, who is still trying to sell a
product, to say everything is fine, new product is in the pipeline,
etc.
This is the usual line offered by companies right up to the instant
they declare bankruptcy and close their doors.
It's called marketing.
Do you have anything to actually disuade me from this line of
thinking? Or just assurances that a new camera is coming out next
month?
As I've said, the lack of *any* info on ongoing research or production
does not bode well in this industry. If Nikon or Canon or Intel or AMD
ceased all press releases, it would be only natural to think that
something was amiss; why should Foveon and Sigma be different? Sigma
does, in fact, continue normal press releases about their lens line.
Do you see the point?