Why is Foveon sensor bad in low light ?

Started Mar 27, 2008 | Discussions
popeye69
Contributing MemberPosts: 507
Like?
Why is Foveon sensor bad in low light ?
Mar 27, 2008

Hi, I dont want to troll sigma, I am a big fan of their inovating film-like sensor & will surelly buy a DP1 for landscape photography.

But I wondered why is the foveon sensor bad in low light. Is it due to the smaller sensor size regarding other DSLR sensors ? olympus 4/3 are smaller, but doing ok, not ? (I am not talking about the dynamic range of course).

What could Sigma do to improve this weakness that surely affects the AF accuracy also on DP1 ?

Zone8
Forum ProPosts: 15,483
Like?
Re: Why is Foveon sensor bad in low light ?
In reply to popeye69, Mar 27, 2008

This is repeated ad infinitum yet there have been many posts of images in low light and higher ISO ratings that belie this "statement".

A check through even just most recent posts will bring up plenty of examples to refute this commonly claimed "defect".

-- hide signature --

Zone8

The photograph isolates and perpetuates a moment of time: an important and revealing moment, or an unimportant and meaningless one, depending upon the photographer's understanding of his subject and mastery of his process. -Edward Weston
http://www.photosnowdonia.co.uk/ZPS

Reply   Reply with quote   Complain
CSJC
Regular MemberPosts: 477
Like?
is it really?
In reply to popeye69, Mar 27, 2008
Reply   Reply with quote   Complain
little laker
Senior MemberPosts: 1,536
Like?
Re: Why is Foveon sensor bad in low light ?
In reply to popeye69, Mar 27, 2008

From what I've seen it really isn't all that bad. Although I haven't tried night photography with it yet.

For some reason I can get a better pic in low light conditions while hand holding it than I can with my Pentax K100D. Which is considered as a pretty good low light camera.

However when I have the camera on the tripod the K100D gives me better pic's.
I've never figured out why.
--
http://www.bcscenicwonder.com

Reply   Reply with quote   Complain
popeye69
Contributing MemberPosts: 507
Like?
Re: Why is Foveon sensor bad in low light ?
In reply to little laker, Mar 27, 2008

I dont want to make a strike between one brand or another. just want to have some clue thoughts about what Sigma should do to improve their only weekness (other sensors have color or dynamicrange - lack weekness ) .

Dont tell me that it is possible to have the same non-grainy pics at 400 or 800 ASA than on a 30D or Canon 5D for example .(that , I insist, have also their own problems). All examples show the opposite. Sigma'sensor is'nt for low light photography , for the moment of course.
But WHY ? (technically)

Reply   Reply with quote   Complain
spalbird
Senior MemberPosts: 2,493Gear list
Like?
Re: Why is Foveon sensor bad in low light ?
In reply to popeye69, Mar 27, 2008

You should show a sample just that I get an impression what you mean with low light, a bad Sigma picture and a good picture from another camera under the same conditions.

It is typical that all effects are mixed up. To make it short, under natural light many feel the Foveon sensor is not bad at all, very difficult are incandescant situations and colors go bad with flourescent.

Then, low light, bad light and available light is mixed into one bag. There are people who feel that the camera should repair bad light which in fact it can't, the photographer is responsible for this. Yes, it is possible that the Foveon is not yet best under every available light situation, but most probably the ingenieurs will stretch the limits with new generations.

For me it seems photography just wasn't possible before the advent of the D3.

This keeps me here with some headache...

Best regards

Wolfgang

 spalbird's gear list:spalbird's gear list
Sigma DP1s Sigma SD10 Sigma SD14 Sigma 10-20mm F4-5.6 EX DC HSM Sigma 18-200mm F3.5-6.3 DC OS HSM +6 more
Reply   Reply with quote   Complain
Tiffles
Veteran MemberPosts: 4,944Gear list
Like?
Re: Why is Foveon sensor bad in low light ?
In reply to popeye69, Mar 27, 2008
 Tiffles's gear list:Tiffles's gear list
Sigma DP1 Sigma DP1 Merrill Kodak DCS Pro 14n Nikon D700 Sigma SD10 +4 more
Reply   Reply with quote   Complain
popeye69
Contributing MemberPosts: 507
Like?
Re: Why is Foveon sensor bad in low light ?
In reply to spalbird, Mar 27, 2008

So , if I understand Foveon has more diffcutlties with fluorecent light or mixed lights (with indoor lightings) and underexposure shots.

OK, but that still doesnt answer my initial question , what can Sigma technicaly do to improve this Sensor ? (what is physically possible?) Why, regarding CCD or CMOS technology, Does the Foveon has more problems with those indoor lights ?

I really like the idea of foveon, again, and just want it to become a number 1 seller , as Sigma is for me a great inovator in this field.

Reply   Reply with quote   Complain
Ed_S
Veteran MemberPosts: 8,073
Like?
Re: Why is Foveon sensor bad in low light ?
In reply to popeye69, Mar 27, 2008

popeye69 wrote:

So , if I understand Foveon has more diffcutlties with fluorecent
light or mixed lights (with indoor lightings) and underexposure shots.

OK, but that still doesnt answer my initial question , what can Sigma
technicaly do to improve this Sensor ? (what is physically possible?)
Why, regarding CCD or CMOS technology, Does the Foveon has more
problems with those indoor lights ?

I really like the idea of foveon, again, and just want it to become a
number 1 seller , as Sigma is for me a great inovator in this field.

There have been many theories about "why". One if I recall is that since each layer is not a discrete color sensor but rather captures all colors to that "depth" http://www.foveon.com/article.php?a=67 , and since the full spectrum is missing in low - narrow spectrum lighting, it is difficult for the mathematical algorithms to "sort out" the algebra to derive discrete colors. Most often maligned seems to be the nominally "blue" layer through which apparently all wavelengths must pass. No or little blue light - difficult to derive a proper color balance. Therefore incorrect colors and/or noise. I didn't create the Foveon so I'm clearly not smart enought to say what would fix it.

Kind regards,
--
Ed_S
http://www.pbase.com/ecsquires

Reply   Reply with quote   Complain
mike earussi
Veteran MemberPosts: 4,744
Like?
Re: Why is Foveon sensor bad in low light ?
In reply to Tiffles, Mar 27, 2008

IMHO it is because light has to pass through three layers whereas the Bayer just has one. Even though you can get "decent" quality photos at ISO 800 the Foveon chip is nowhere as good as the Canon 40D or Nikon D300 in low light capability. Where they excell, and produce better images than any other, is in good lighting conditions at 5500K.

Reply   Reply with quote   Complain
Kendall Helmstetter Gelner
Forum ProPosts: 18,405Gear list
Like?
Why not, if it is? And what about low ISO noise.
In reply to popeye69, Mar 27, 2008

popeye69 wrote:

Dont tell me that it is possible to have the same non-grainy pics at
400 or 800 ASA than on a 30D or Canon 5D for example .(that , I
insist, have also their own problems). All examples show the
opposite.

Not the DP-1 examples:

http://www.sigma-dp1.com/jp/sample-photo/

Look at 33 and 34 for ISO 400 and 800 shots.

And by now, there are other examples all over that show the camera looking really good.

As for size, the Foveon chip has the same Photosite size as the 5D. So you tell me - why are ISO 100 pictures noisier on the 5D (and there are plenty of examples I've seen of that).

 Kendall Helmstetter Gelner's gear list:Kendall Helmstetter Gelner's gear list
Sigma 70-200mm F2.8 EX DG OS HSM Sigma 8-16mm F4.5-5.6 DC HSM Sigma 24-70mm F2.8 EX DG HSM Sigma 50-500mm F4.5-6.3 DG OS HSM Sigma 85mm F1.4 EX DG HSM +4 more
Reply   Reply with quote   Complain
Amin Sabet
Veteran MemberPosts: 6,712Gear list
Like?
Unfortunately, yes it is.
In reply to CSJC, Mar 27, 2008

Your images show quite a bit of blotchy color noise in the shadows. Much more than cameras with current APS-C Bayer sensors show at those ISOs. This situation goes from suboptimal to poor depending on the type of light. The DP1 samples I have seen have not disappointed, and I fully plan to buy one as soon as I can find one in stock somewhere for $800. However, high ISO color performance is just okay - better than small sensor cameras at least. B&W performance, OTOH, seems very good.

 Amin Sabet's gear list:Amin Sabet's gear list
Fujifilm XF1 Olympus OM-D E-M5 Leica M Typ 240 Panasonic Lumix G 14mm F2.5 ASPH Leica Summilux-M 35mm f/1.4 ASPH +4 more
Reply   Reply with quote   Complain
popeye69
Contributing MemberPosts: 507
Like?
Re: Why not, if it is? And what about low ISO noise.
In reply to Kendall Helmstetter Gelner, Mar 27, 2008

yeap, I already saw this official gallery, very stunning, it feels like medium format film - like pics.

Hoping that one day A foveon 30mm/30mm (or higher) sensor cam will come out , to get nice squared digital picts with it !

Thanks by the way for the technical explanations. Would'nt it be better if they used a divided sensor in 3 layers , on the same sensor , than in the depht of it?

Reply   Reply with quote   Complain
spalbird
Senior MemberPosts: 2,493Gear list
Like?
Re: Why is Foveon sensor bad in low light ?
In reply to popeye69, Mar 27, 2008

popeye69 wrote:

So , if I understand Foveon has more diffcutlties with fluorecent
light or mixed lights (with indoor lightings) and underexposure shots.

The pictures turned to B&W at these situation are often very pleasing.

Florescent light, for example is bad light as it consists not of a continous spectrum, and normally has a heavy peak with the greens, which turns skin greenish in such situations.

OK, but that still doesnt answer my initial question , what can Sigma
technicaly do to improve this Sensor ? (what is physically possible?)
Why, regarding CCD or CMOS technology, Does the Foveon has more
problems with those indoor lights ?

They are improving from model to model, bringing less and more pleasing noise. I have seen pictures from 350D which also has problems with indoor lights, and the Canon way is to turn everything more reddish. Why don't you ask why other camera makers don't improve their good light performance to that of the Foveon?

I really like the idea of foveon, again, and just want it to become a
number 1 seller , as Sigma is for me a great inovator in this field.

The skills of a photographer are more important than the capabilities of the camera, I have heard. If you like the technology, why don't give it a try. If you love your equipment there is a little less between you and a good picture

Best regards

Woflgang

 spalbird's gear list:spalbird's gear list
Sigma DP1s Sigma SD10 Sigma SD14 Sigma 10-20mm F4-5.6 EX DC HSM Sigma 18-200mm F3.5-6.3 DC OS HSM +6 more
Reply   Reply with quote   Complain
Eamon Hickey
Veteran MemberPosts: 3,156
Like?
its color separation method
In reply to popeye69, Mar 27, 2008

popeye69 wrote:

So , if I understand Foveon has more diffcutlties with fluorecent
light or mixed lights (with indoor lightings) and underexposure shots.

OK, but that still doesnt answer my initial question , what can Sigma
technicaly do to improve this Sensor ?

Folks whose informed opinions seem credible to me say that the Foveon's color separation method -- i.e. relying on wavelength penetration depth in silicon -- is not as precise as the color filters used in a Bayer sensor, which give a very sharp, specific, and predictable wavelength cutoff for each of the three colors captured. Another factor may be signal attenuation -- i.e. the signal that is being detected in the lower portions of the Foveon pixel (red) is weaker than the same color signal when captured by a Bayer pixel (and therefore harder to distinguish from noise).

Apparently, Foveon must use complex (and perhaps aggressive) color transforms to compensate for these factors when outputting a finished color image and that adds up to more noise, especially chroma noise, when the overall signal started out lower to begin with (i.e. whenever you use a higher ISO).

Of course, Foveon's color separation method is what gives it the unique advantage of capturing all three colors at one location. It's a tradeoff that proves the old adage that ain't nuthin' perfect. I think Foveon and Sigma have clearly demonstrated that the Foveon chip can be used to do a wide variety of excellent photography. And in some ways, it's superior. But it has weaknesses, too, and for some other kinds of things, it may underperform compared to alternative technologies.

If the synopsis above is more or less accurate, then it seems probable that Foveon will always lag CFA-based sensors (either Bayer or the alternative patterns developed by Kodak) in high-ISO noise performance. But Foveon made a notable improvement when they added microlenses to the chip in the SD-10 and maybe they've got other tweaks up their sleeve.

Reply   Reply with quote   Complain
Amin Sabet
Veteran MemberPosts: 6,712Gear list
Like?
Re: Why not, if it is? And what about low ISO noise.
In reply to Kendall Helmstetter Gelner, Mar 27, 2008

Kendall Helmstetter Gelner wrote:

Not the DP-1 examples:

http://www.sigma-dp1.com/jp/sample-photo/

Look at 33 and 34 for ISO 400 and 800 shots.

And by now, there are other examples all over that show the camera
looking really good.

Compared to other compact cameras, yes, very good. Compared to other large sensor cameras, then not so good.

As for size, the Foveon chip has the same Photosite size as the 5D.
So you tell me - why are ISO 100 pictures noisier on the 5D (and
there are plenty of examples I've seen of that).

ISO 100 noise on 5D samples is never significant unless an underexposed area is "pushed" in postprocessing. OTOH, ISO 800 and up color images from the DP1 frequently show blotchy color problems in the shadow regions, even when processed from RAW. It's not a big deal to me, because this is still much better than one can get from other compact cameras. However, any comparison between noise profiles of color DP1 images and color 5D images, from a practical standpoint, is really a non-starter.

 Amin Sabet's gear list:Amin Sabet's gear list
Fujifilm XF1 Olympus OM-D E-M5 Leica M Typ 240 Panasonic Lumix G 14mm F2.5 ASPH Leica Summilux-M 35mm f/1.4 ASPH +4 more
Reply   Reply with quote   Complain
spalbird
Senior MemberPosts: 2,493Gear list
Like?
Re: Why not, if it is? And what about low ISO noise.
In reply to popeye69, Mar 27, 2008

You ask for high demand without probably needing it. Sigma and Foveon delivers what they can sell in small quantities. If they satisfy just you they run into the risk not to sell a single sample, because it is too exotic and expensive.

The more people support Sigma and Foveon, the more easy for them to bring more value for money. It is easy like this. For now you have to live with what they have on market.

Wolfgang

popeye69 wrote:

yeap, I already saw this official gallery, very stunning, it feels
like medium format film - like pics.

It seems this quality is not enough, tell me why...

Hoping that one day A foveon 30mm/30mm (or higher) sensor cam will
come out , to get nice squared digital picts with it !

Thanks by the way for the technical explanations. Would'nt it be
better if they used a divided sensor in 3 layers , on the same sensor
, than in the depht of it?

It is something like that already. Light can't go through silicon I guess.

 spalbird's gear list:spalbird's gear list
Sigma DP1s Sigma SD10 Sigma SD14 Sigma 10-20mm F4-5.6 EX DC HSM Sigma 18-200mm F3.5-6.3 DC OS HSM +6 more
Reply   Reply with quote   Complain
boothrp
Forum MemberPosts: 66
Like?
Re: Why is Foveon sensor bad in low light ?
In reply to popeye69, Mar 27, 2008

As far as low light DP1 photograpy goes, have a look at.

http://www.rytterfalk.com/2008/03/19/dp1-8s-at-iso800/

OK, so it's got grain, but it doesn't ruin the picture for me.

You would have to go to a significantly bigger camera or longer exposure to obtain such a good or better result.

Most of the debating over the DP1 seems to miss what to me is the main point -

You won't find a digital camera out there that gives such good image quality in such a small package.

Trying to find the ideal camera is a sure route to insanity, it will either be too big/heavy/expensive or it will be lacking in features/quality.
Most of the posts I see relating to the DP1 do not seem to acknowledge this.

I don't claim my logic to be 100% watertight, but hope this is seen as helpful & neutral.
Paul.

Reply   Reply with quote   Complain
ShutterFlutter
Regular MemberPosts: 224
Like?
agreed
In reply to Amin Sabet, Mar 27, 2008

I routinely shoot at 1600 ISO with my 5D yet I still hesitate DP1 at 800ISO. It might change with learning how to use it but I would say that DP1's ISO 800 barely compares with 5D's ISO 1600. For my taste 5D at 3200 ISO starts to fall apart and I rarely ever shoot at that ISO setting.

Here is one very low light DP1 pic that I took today and butchered in post processing

Reply   Reply with quote   Complain
SandyF
Forum ProPosts: 14,875Gear list
Like?
Re: ISO's & EV's
In reply to ShutterFlutter, Mar 27, 2008

for best image quality on the DP1 -- in fact for any camera -- keep the ISO as low as you can. PHotography 101.

Yesterday experimenting with the DP1 indoors, mixed light, shopping mall, I had 'better' results at ISO400 +EV.03 and +EV.07 than I did at ISO800, just as an example. Experiment with your settings, learn how it handles best in particular situations.
Best regards, Sandy
sfleischmann@gilder.com
http://www.pbase.com/sandyfleischman
http://www.flickr.com/photos/sandyfleischmann

 SandyF's gear list:SandyF's gear list
Sigma DP2 Sigma DP1 Sigma DP2 Merrill Canon EOS 5D Mark II Sigma SD9 +5 more
Reply   Reply with quote   Complain
Keyboard shortcuts:
FForum MMy threads