-
Welcome to the new forums! Please read this first. For known issues we are working to resolve, click here.
What will it take to make the Fovean naysayers to S.U.
- Thread starter Edwaste
- Start date
You pose a very important question to this forum. If available (and reasonable) to all manufacturers, a six megapixel Fovean could possibly dominate the prosumer marketplace in a couple of years and become the defacto standard (my guess is based on the good showing so far of the "near production ready" Sigma prototype).
I think the pro marketplace will always be dominated by Nikon and Canon glass in conjunction with low noise sensors. Between the two, Nikon would probably embrace a Fovean sensor before Canon. What do you think?
Joe Kurkjian
I think the pro marketplace will always be dominated by Nikon and Canon glass in conjunction with low noise sensors. Between the two, Nikon would probably embrace a Fovean sensor before Canon. What do you think?
Joe Kurkjian
Valliesto Bailey
Senior Member
I have been very harsh on the X3/SD_9 in the past. That was mostly out of my dislike for vaporware, since the chip has since materialized I have been much more objective. The samples look very good and I am impressed by the technology behind them. I am not impressed by the Sigma body, but we can't have everything I guess.
While I am very impressed with the image quality, they are not perfect by any stretch of the imagination. I think that those who point out the flaws they find in these samples do us all a service. Would you have nothing but "yes men" in these forums just praising the SD-9 blindly?
The SD-9 has shown that there is promise in the X3 technology, but it (the SD-9) is not the end all and be all of digital imaging.
Valliesto
'A hero is no braver than an ordinary man, but he is brave five
minutes longer.'
While I am very impressed with the image quality, they are not perfect by any stretch of the imagination. I think that those who point out the flaws they find in these samples do us all a service. Would you have nothing but "yes men" in these forums just praising the SD-9 blindly?
The SD-9 has shown that there is promise in the X3 technology, but it (the SD-9) is not the end all and be all of digital imaging.
--
Valliesto
'A hero is no braver than an ordinary man, but he is brave five
minutes longer.'
- R.W. Emerson
E
Erik Magnuson
Guest
About the same as the current state of the art 11 - 14 MP Bayer? They've only roughly matched 4 - 6 month old cameras at low ISO - they are still half a generation behind.
--
Erik
Kind of hard to tell. Since Nikon bodies, or least parts of them are used with Kodak's new digital camera, I would tend to think that Nikon will embrace the Fill Factory full frame 14 megapixel.
Perhaps Nikon will employ the Fovean in the Coolpix series.
TJ66308
Forum Enthusiast
This is their first sensor. I'll bet they are working on others. Look how Faveon's sensor compares to Nikon, Canon, etc, etc first trys. I think Faveon has done a major end-run on the industry and the sensor has many other advantages for digital SLR manufacturers that we will not really be talking about on these forums. We will see.
Regards
Ted
Regards
Ted
R
Ron Parr
Guest
I agree that it's not a touchdown - probably more like field goal. However, they're definitely in the game now.
They will be criticized for their shortcomings, but I don't think they can be dismissed. If they can improve on this for the next generation, things will get very, very interesting.
--
Ron Parr
FAQ: http://www.cs.duke.edu/~parr/photography/faq.html
Gallery: http://www.pbase.com/parr/
DavidMillier
Forum Pro
You don't need to overhype...
It's quite possible that with the SD9 Foveon have just sent a signal that the days of Bayer sensors are numbered.
But, you're not saying that the SD9 images outclass the Kodak 14MP or even the 1Ds are you?
There's obviously a lot of promise in the new sensor and it will certainly give the 6MP cameras a run for their money AND it's very cheap.
All good stuff, but overall Foveon superiority is going to have to wait for a much bigger sensor...
It's quite possible that with the SD9 Foveon have just sent a signal that the days of Bayer sensors are numbered.
But, you're not saying that the SD9 images outclass the Kodak 14MP or even the 1Ds are you?
There's obviously a lot of promise in the new sensor and it will certainly give the 6MP cameras a run for their money AND it's very cheap.
All good stuff, but overall Foveon superiority is going to have to wait for a much bigger sensor...
No way am I saying that. I was sick of hearing the word vaporware.
I think the only vaporware in this case was between their ears.
!4 megapixels is nice... especially for the manufacturers of portable storarge. I would rather shoot a 3 megapixel RAW file from a fovean than a RAW 6 megapixel from a bayer chip.
E
Erik Magnuson
Guest
Neither side can afford to stand still. Regardless of which camera you prefer, it will be both better and cheaper because of this competition.
--
Erik
Paul W. Walters
Senior Member
The 14 megapixel Fill Factory is being produced only for Kodak. They have an exclusive deal for THAT sensor, but they can manufacture different sensors for Nikon.
My feeling with this whole thing is that Foveon DOES have great potential! However, when you look two years into the future, you just can't see how much CCD's will improve by then. They currently are not at the top level, so they not only have to get to the top level of what's available TODAY, but they have to do it for what's available at the time they reach that level!
I'm hoping for the best, but satisfied with my S2 in the meantime.
Paul
My feeling with this whole thing is that Foveon DOES have great potential! However, when you look two years into the future, you just can't see how much CCD's will improve by then. They currently are not at the top level, so they not only have to get to the top level of what's available TODAY, but they have to do it for what's available at the time they reach that level!
I'm hoping for the best, but satisfied with my S2 in the meantime.
Paul
E
Erik Magnuson
Guest
Of course. The "next" sensor has already been announced and is expected to be in cameras by spring. (I'm talking about the 1/2" F10.) In fact, a potential problem for Foveon will be trying to decide which markets they want to compete in. Are they going to make a run at the low end (the oft-speculated "digital disposable"?) Stick with the high-end, higher margin sensors? Do they have the resources to do both?
I think that "the industry" is more surprised by the FillFactory full-frame CMOS to sell in a camera for $4k than Foveon.
--
Erik
Paul W. Walters
Senior Member
They made a very reputable showing with digital cameras, but if they stick this technology into a digital camcorder, it's WAY ahead of the game! I'm anxious to see them use something like the F10 Image Sensor in a video camera. It will compete with 3 CCD digital cameras and quite posibly surpass them! But, we'll have to wait and see. One way or the other, I don't believe Foveon will fade away anytime soon.
Though I also think it will be a VERY difficult battle for them to actually PASS UP the quality of the best CCD sensors at any given time for digital cameras.
Paul
Though I also think it will be a VERY difficult battle for them to actually PASS UP the quality of the best CCD sensors at any given time for digital cameras.
Paul
L. Kraven
Well-known member
I don't think I want to hear all the criticism suddenly stop. Just because there is a lot of criticism doesn't mean some of it isn't unwarranted.
I don't think anyone can truthfully say that Foveon's innovation isn't good, but neither can you say that it suddenly makes all other (bayer included) technology obsolete.
And your point would seem silly if I put it in the reverse: If this is what an 11megapixel bayer-interpolated sensor can do, what can a 22megapixel bayer-interpolated sensor do?
I think that if you shrunk down a large 6mp or 11mp image to the same dimensions of the Foveon, you'd end up with a sharper image and much less noise (because the downsize would act as a rudimentary denoising filter).
My very untechnical observation of the the Foveon's advantage works something like this: Assuming that the actual pixel sizes are the same, a 3mp X3 sensor should resolve the same resolution as a 3mp bayer interpolated sensor, BUT, get 3x as much color data. Now, to our eyes, that color data could translate into sharpness (simply because, between pixels, there can be a greater amount of contrast, and contrast is how we visually determine sharpness), but the truth is that actual resolution should be the same, lens, etc. being equal.
The problem is, Sigma lenses are not Nikon or Canon or Zeiss lenses, they're Sigma lenses. So that's one legitimate negative.
The other problem is you're already stretching the 3mp Foveon sensor which isn't on the market yet to the "futureware" 6mp Foveon sensor which who knows when will be available.
Foveon is behind in the ballgame if they want to compete with the 10+mp crowd, and my gut feeling is that while the 3mp X3 sensor may be the equal or slight superior of an S2 or D60 image, the image will be noisier based on what I've seen.
No one is knocking the technology, but I really hope this launches it and they start moving on it right away, else, they'll go the way of the betamax and Orb drive. (The Orb drive, a 2gb drive sounded GREAT when they announced it and it was competing with only the 100mb Zip drive... but when they delivered it 18 months later, the 2gb Jazz drive was already everywhere... though the Orb was superior technology.)
I just don't think it's wise to hang your hat on a "6mp X3" when the 3mp one isn't in anyone's hands yet, and who knows what can and will happen by the time the 6mp X3 hits the market. Here's hoping it never does, and they go straight to something like a 12mp X3.
--
L. Kraven
I've printed some of the samples and while I'm impressed with the output I'll wait for the 6mp version without the Sigma mount. I'll keep my S2 for now since I don't see any SIGNIFICANT improvement in print quality over my S2. I wonder if Foveon offered the chip to Nikon or Canon? I just don't see them selling a lot of sd9's for $1800.......If they included a lens and out the door for $1200, maybe Sigma could grab a large chunk of the $1000 niche.
Boris
http://public.fotki.com/borysd/
Boris
http://public.fotki.com/borysd/
tim martin
Senior Member
In July or so I thought I heard something about a story in "IIEE",
or something like that, saying National Semiconductor was having
"crosstalk" problems with the sensor.
Anybody know anything about that ? If so, would a problem
like that be worse on a sensor with more pixels ?
In summary, I wonder if Foveon and National Semiconductor
will have problems with a bigger sensor.
Not that this would be an issue for me. There's really only one
issue left for me, price. That is, I want a Foveon DSLR, and I
want to pay no more than say, $1100 or something.
If after 18 months the prices don't seem to be dropping, then
maybe I would cough up all the money, but for now I will
just wait, and rely on my G2. It's not really so bad.
E
Erik Magnuson
Guest
tim martin wrote:
Note: this is speculation. I'll let you decide how knowledgeable it is.
On Foveon, color cross talk could be one of two things:
--
Erik
Note: this is speculation. I'll let you decide how knowledgeable it is.
For Bayer sensors, color cross talk is when light passes through the filter for one color but causes a change in charge for the cell of neighbor. This is why the photosite is usually surrounded by a metal shield. Microlenses also help reduce this by focusing the light on a single photosite.
On Foveon, color cross talk could be one of two things:
- a problem in deciding what color to assign light that is absorbed at a depth that is half-way between two layers
- or, similar to the Bayer, light strikes at a angle: the blue light is registered at hitting one pixel, but the green/red light is absorbed by a neighboring pixel (e.g. the light continues diagonally through the layers).
I don't think that this problem would be better/worse with a bigger sensor (e.g.full-frame 24 x 36mm). It may have to be addressed to build a denser sensor (e.g. a 6MP X3 4/3" sensor).
--
Erik
Dana Jacobsen
Leading Member
Compare the airport scene with the SD9 here:
http://www.digitalcamera.jp/report/SD9-020927/
with the D60 and 1D images here:
http://www.digitalcamera.jp/report/EOS-1Ds-020924/index.htm
or the D60 and S2Pro images here:
http://www.digitalcamera.jp/report/S2Pro-020602/index.htm
Ignore the horrible chromatic abberation on the SD9 shot -- I'm going to assume this is the lens, though it is strange that it is a different lens (20mm) than Phil used in the bike shot (28-70 @ 30mm) and we get the same extreme problems. You could argue that the SD9 shot is underexposed which makes the comparison not as valid, especially with regards to noise.
When I downsample the D60 image to the size of the SD9 image, I see little difference -- perhaps more noise in the SD9 image, especially in the shadows. When the SD9 image is upscaled to the D60 size, the D60 has noticibly more detail. Yes, I see more resolution! And for something that is sure to get me flamed, I think the 4MP 1D image has more detail as well -- regardless od detail it looks much better to me. Maybe it is just a really bad lens on the Sigma. Note that I've applied appropriate sharpening to both images before comparing.
The one advantage the Foveon has here is the lack of chroma artifacts. If you look carefully at the buildings on the other cameras you can see some artifacts. They're pretty subtle except for the 1Ds and D100 images which are much more visible at full resolution.
It's funny -- you guys think of us as naysayers, while I think of you as overenthusiastic. 'Bayer suxx' and such messages don't exactly lead to rational discussions. I look at the one set of comparison images and I don't see amazing results. If the Foveon can't beat a 4MP Bayer then this strikes me as indicating a possible test problem -- Phil's tests are much more controlled so we'll see next month.
We're comparing the first Foveon product by Sigma to a second generation product by Canon. So in terms of real world implementations, Sigma/Foveon is at a disadvantage. Do they have nice images? Yes. Do they make all existing digital camera obsolete? No. Would I like a 6MP low-noise non-CFA (e.g. Foveon) sensor in my next Canon-mount camera? Sure.
Dana Jacobsen
Leading Member
I just wanted to make it clear that my subject 'Better images' was in direct answer to Edwaste's query of what it would take. The examples I linked to are not meant to be any sort of this-shows-the-final-answer sort of thing. It was meant to show that there are good and bad points on both sides and the answer isn't clear cut.
I suspect that when the Sigma is shipped and we see lots of lengthy reviews that we can get a better answer. If the images show a marked improvement over the corresponding CFA sensor based cameras, then we'll all agree. Unfortunately life isn't going to be quite so simple (e.g. what is the proper corresponding sensor size?, what about cost?, what about actual product choices?, how important is resolution vs. chroma to you?).
Karl Guttag
Veteran Member
The original "hype" was that the X3 was a dramatic breakthrough that was going to obsolete all existing cameras. I think even the so called naysayers agree that the X3 concept is very interesting and admit that the Bayer is a hack.
The issue is that the images resolution appear to be between a 3MP D30 and a 6MP D60. Furthermore the SD9 only goes to ISO400 where the D60 goes to ISO1000. Any REASONABLE person would realize that the SD9/current-X3 is not doing well in terms of NOISE.
Furthermore, as expected there are other artifacts in the images. We don't know if these are software or the imager or some combinations. Even Phil Askey has said that Foveon/Simga know that there are problems in the images, at least some of which they think they can fix in software/firmware, yet the foveon enthusiasts claim not to see anything.
None of this was unexpected to those of us with some technical background. A 3.4MP X3 was expected to be about a good in resolution as a 6MP bayer. We knew that a Bayer "megapixel" had to be discounted about 1.5 to 2X, but it was not the 3X to 4X as some Foveon enthusiasts ludicrously claimed.
We also had to figure that the "stacked" sensors had to be compromised compared to a simpler unstacked sensor that had a filter on top of it. This shows up in the ISO. If the stack sensor was as good as the Bayer sensor, then the X3 should have had BETTER low noise performance since in theory it did not throw away 1/3rd of the light of the filter. Instead the X3 has worse low light/high-ISO performance.
There was a lot of concern that the Canon D60 had only ISO1000 when the D30 when to ISO1600. Image if the D60 only went to ISO400. Only the still life people would have considered it a clear step forward over the D30.
While I/we know that the Bayer filtering is a compromise, I/we think WAY too much has been made over it. If the X3 is going to clearly beat/obsolete the Bayer sensors, it has to get nearer to the megapixels of the Bayer sensors AND have the same or less noise at the same ISO, while at the same time it has to show that it can ship reliably in high volume (yet to be seen).
Karl
Karl
The issue is that the images resolution appear to be between a 3MP D30 and a 6MP D60. Furthermore the SD9 only goes to ISO400 where the D60 goes to ISO1000. Any REASONABLE person would realize that the SD9/current-X3 is not doing well in terms of NOISE.
Furthermore, as expected there are other artifacts in the images. We don't know if these are software or the imager or some combinations. Even Phil Askey has said that Foveon/Simga know that there are problems in the images, at least some of which they think they can fix in software/firmware, yet the foveon enthusiasts claim not to see anything.
None of this was unexpected to those of us with some technical background. A 3.4MP X3 was expected to be about a good in resolution as a 6MP bayer. We knew that a Bayer "megapixel" had to be discounted about 1.5 to 2X, but it was not the 3X to 4X as some Foveon enthusiasts ludicrously claimed.
We also had to figure that the "stacked" sensors had to be compromised compared to a simpler unstacked sensor that had a filter on top of it. This shows up in the ISO. If the stack sensor was as good as the Bayer sensor, then the X3 should have had BETTER low noise performance since in theory it did not throw away 1/3rd of the light of the filter. Instead the X3 has worse low light/high-ISO performance.
There was a lot of concern that the Canon D60 had only ISO1000 when the D30 when to ISO1600. Image if the D60 only went to ISO400. Only the still life people would have considered it a clear step forward over the D30.
While I/we know that the Bayer filtering is a compromise, I/we think WAY too much has been made over it. If the X3 is going to clearly beat/obsolete the Bayer sensors, it has to get nearer to the megapixels of the Bayer sensors AND have the same or less noise at the same ISO, while at the same time it has to show that it can ship reliably in high volume (yet to be seen).
Karl
--
Karl
Keyboard shortcuts
- f
- Forum