What are SD9's shortcomings? Objective comments please.THANKS!

Erik Magnuson · Sep 10, 2003

sg10 said:
It is indeed pointless, you will never convince me that a 1 color
render of a complex 36-bit data file has the same dynamic range as
a 36-bit render.

Sensors are NOT renders:

Let's talk about a hypothetical sensor. Sensors measure light by measuring electrical charge. So our unit of measure is going to be electrons. This sensor has a noise floor of 100 electons. That means no matter how little light hits the sensor, it will always read out at least 100 electrons. And let say that this sensor has a saturation limit of 25,500 electrons. Each sensor well cannot hold any more electrons than that as they will spill out into the anti-blooming drains. The sensor's dynamic range is 255:1 or about 8 stops (pretty good sensor, right?) Note that to make the math simple, I'm just dealing with a single color. It does not matter is the 3 colors are stacked or adjacent for this example.

If we use an 8-bit ADC scaled for the full sensor range, then we will measure in units of about 100 electrons. 100 = 1, 200 = 2, ... 25,500 = 255. Now let's use a 12-bit ADC instead scaled for the full sensor range. 6 electrons = 1. But wait, what's the first value? That's right 100 = 16. It does not matter that the ADC can decode only 6 electrons as it will never see only 6 electrons (100 is the minimum, right?). The maximum would be 25,500 = 4095. Now we could use the same scaling as the 8-bit ADC (100=1), but there will never be more than 25,500 electrons to be measured so 255 would still be the upper limit. So the dynamic range is either way is 4095:16 or 255:1. In other words, still 8 stops.

So, for the purpose of dynamic range calculations, it does not matter that the 12-bit ADC can measure values like 106 electrons or 212 electrons. The limit OF THE SENSOR is still 8 stops. (Now, I'm not going to go into other sources of noise like how reliably does 106 electrons mean more light than 100 electrons or if units of 100 electrons are enough for a particular use as that has very little to do with dynamic range.)

--
Erik

Brian Slater · Sep 10, 2003

Forrest said:
Forrest41926 said:

Oooh, another troll has entered the realm. Do yourself (and others
who might be misslead by you) a favor and check out this forum for
the latest information on your missinformation. Then get some
hands-on experience with an SD-9. Finally, come on back and report
your findings to us. I'll lay you 10:1 odds that you'll be singing
another tune. I'll drink to that!

Click to expand...

Why is anyone who takes the work "objective" to heart a troll? Is
no one who won't sing the tune "SD-9 blows large format out of the
water" welcome here?

My experience of reading this forum for many weeks is that almost everyone has been very welcome. However, the particular individual referred to above appears to meet the definition of a troll quite well. It is hard to see how the word "objective" could be applied to a number of his posts, particularly one that at least had the appearance of being racist. On the other hand, you appear to generalize in a post in the Canon 300D forum entitled "Sigma trolls making everything up..."

Also, I haven't heard the tune you refer to. There are individuals that overstate the qualities of their preferred camera in most forums. The collective experience here is that, at least in the area of image sharpness, the SD9 does very well in comparison with Canon's 6MP offerings. I have seen nothing but encouragement in this forum for anyone willing to post full-sized images that can be compared "objectively" with those from the SD9. You will also find many ready to admit that the SD9 has it's weaknesses, and frequent admission of the fine qualities of other cameras. What seems to be fairly consistent in this forum though is a strong sense of satisfaction with the potential of the SD9 to deliver, and a realization that the photographer is generally far more limitied than the equipment.

Kevin Blair · Sep 10, 2003

And on top of that, I don't know if there's a good way of independently measuring the dynamic range of various cameras to confirm if any of the numbers quoted for a sensor actually correlate to real world performance. Any takers?

Kevin

ding1 said:
Hope the next generation of the Foveon sensor have bigger DN!

Kevin Blair said:

I know Phil tried to measure the dynamic range of cameras in his
early reviews but stopped because there was some question as to the
accuracy of the numbers. It would help in these types of
discussions if there was a way to independently verify the dynamic
range of various cameras in actual practice. (That would also help
settle the current discussion as well.)

Kevin

Kevin Blair said:

Kevin

ding1 said:

Just a note about the Foveon X3 sensor's dynamic range: according
to http://www.alt-vision.com/r/documents/AeroSense_2003_Oral.pdf
page 30, its dynamic range is 61 dB. Therefore, the use of 12 bit
A/D converters in SD9 is justified. The number you estimated (10
dB) is way too low; isn't it a typo?

Click to expand...

Click to expand...

Click to expand...

Laurence Matson · Sep 10, 2003

Brian Slater said:
Forrest said:

Brian Slater said:

Oooh, another troll has entered the realm. Do yourself (and others
who might be misslead by you) a favor and check out this forum for
the latest information on your missinformation. Then get some
hands-on experience with an SD-9. Finally, come on back and report
your findings to us. I'll lay you 10:1 odds that you'll be singing
another tune. I'll drink to that!

Click to expand...

Why is anyone who takes the work "objective" to heart a troll? Is
no one who won't sing the tune "SD-9 blows large format out of the
water" welcome here?

Click to expand...

My experience of reading this forum for many weeks is that almost
everyone has been very welcome. However, the particular individual
referred to above appears to meet the definition of a troll quite
well. It is hard to see how the word "objective" could be applied
to a number of his posts, particularly one that at least had the
appearance of being racist. On the other hand, you appear to
generalize in a post in the Canon 300D forum entitled "Sigma trolls
making everything up..."

Also, I haven't heard the tune you refer to. There are individuals
that overstate the qualities of their preferred camera in most
forums. The collective experience here is that, at least in the
area of image sharpness, the SD9 does very well in comparison with
Canon's 6MP offerings. I have seen nothing but encouragement in
this forum for anyone willing to post full-sized images that can be
compared "objectively" with those from the SD9. You will also find
many ready to admit that the SD9 has it's weaknesses, and frequent
admission of the fine qualities of other cameras. What seems to be
fairly consistent in this forum though is a strong sense of
satisfaction with the potential of the SD9 to deliver, and a
realization that the photographer is generally far more limitied
than the equipment.

--
Laurence Φ€

http://www.pbase.com/lmatson/sd9_images
http://www.pbase.com/sigmasd9/root
http://www.pbase.com/cameras/sigma/sd9
http://www.beachbriss.com (eternal test site)

Forrest41926 · Sep 10, 2003

Brian Slater said:
everyone has been very welcome. However, the particular individual
referred to above appears to meet the definition of a troll quite
well. It is hard to see how the word "objective" could be applied
to a number of his posts, particularly one that at least had the
appearance of being racist. On the other hand, you appear to

I haven't read any of his other posts ... although having extensively used Sigma's 15-30 and Canon's 16-35L ( which are very similar, and competing lenses ), while I can easily say the Sigma is quite a bit sharper than the Canon, I can also confirm a lot of what he said about color casts ... although I think he may have either exadurated it a little bit, or got a bad copy of the lens. While the 15-30 clearly outresolves the 16-35L at any aperture, it does have a warm cast to it's images.

Brian Slater said:
generalize in a post in the Canon 300D forum entitled "Sigma trolls
making everything up..."

Yes, it's true; 90% of what was presented as fact in the message I responded to, was made up by a Sigma troll in the Rebel D forum, and supported ( again, as fact ) by another one. People who take so much effort to create stunning, high-quality images ( carry heavy gear, a tripod, wait for MLU, ... ) would do well to study a bit of biology, and to know the basic facts. What bothers me is when people present bits of information that are patently false, as if they were established and widely accepted facts.

Brian Slater said:
Also, I haven't heard the tune you refer to. There are individuals
that overstate the qualities of their preferred camera in most
forums. The collective experience here is that, at least in the
area of image sharpness, the SD9 does very well in comparison with
Canon's 6MP offerings. I have seen nothing but encouragement in

I've seen quite a few people pop over into the other forums to point out how the output from an SD-9 is virtually indistinguishable from the output of a 1Ds. Usually this claim is made one the sole basis that the 1Ds has 11 mpx, while the Sigma has 10.5 million receptor sites.

Brian Slater · Sep 10, 2003

Forrest said:
Yes, it's true; 90% of what was presented as fact in the message I
responded to, was made up by a Sigma troll in the Rebel D forum,
and supported ( again, as fact ) by another one. People who take
so much effort to create stunning, high-quality images ( carry
heavy gear, a tripod, wait for MLU, ... ) would do well to study a
bit of biology, and to know the basic facts. What bothers me is
when people present bits of information that are patently false, as
if they were established and widely accepted facts.

I don't want to get into a slanging match here, nor to defend clearly erroneous statements made by others. However, your repsonse referred to included the following statement:

"Enlarging an SD-9 image to the size of the Rebel D's images will bring gobs of noise out." It bothers me too when people present patently false information, and this statement IS patently false.

Forrest41926 said:
I've seen quite a few people pop over into the other forums to
point out how the output from an SD-9 is virtually
indistinguishable from the output of a 1Ds. Usually this claim is
made one the sole basis that the 1Ds has 11 mpx, while the Sigma
has 10.5 million receptor sites.

I have never seen such a post.

Laurence Matson · Sep 10, 2003

for the trees.

Forrest, my friend,

If you truly believe that other forums are more objective or whatever, please stick to your beliefs. But the above poster is quite right in the general tone of this one.

Consider the source. If one of my comrades cruises into another forum and makes outlandish statements, deal with the issue then and there. But please do not generalize and extrapolate his pixel claim or other nonsense on to the entire gang of know-nothings here. You want to be respected as an individual; the feeling is mutual.

We have received more than our share of pontificating party crashers (to use the words one of them gave himself) about the pathetic nature of this camera, its worthless technology and the liars and idiots running the companies that give it to us. (If you need links to substantiate any of that, let me know.) Why not just let us stumble along in the light (can't shoot at night)?

Strange as the concept may seem, having the above drummed into you long enough and then discovering it is not true makes one a little stroppy. In short and speaking for myself, I get kind of sick of it. Here I am having a great time with some digital device, making nice shots, fielding tough phone calls from Decker loafing in Hawaii, and with amazing consistency someone cruises in here almost invariably from our friendly Canon supplier and bashes around for a while making it clear how dumb we are to have this etc. equipment. It is sort of like the schoolyard bully routine. Wouldn't you get tired of that? I sure do, and I can't blame others if they did too.

--
Laurence Φ€

http://www.pbase.com/lmatson/sd9_images
http://www.pbase.com/sigmasd9/root
http://www.pbase.com/cameras/sigma/sd9
http://www.beachbriss.com (eternal test site)

Laurence Matson · Sep 10, 2003

Brian,

Forrest is right about some ridiculous claims. They are everywhere, however, and are not reserved for addled SD9 users. Much of this began with the furious attacks back in February 2002, when the Cannonites lined up their big guns to blow Foveon out of the harbor before the camera was even launched. Captain G u t t a g was in charge back then.

Otherwise, I am with you. I print A3+ all the time. No noise except ooos and aaaahs from madame, who finally likes something that I do. The really big panos cause more consternation: "Looks great. Where do I put it?", but they are 3 meters long.

--
Laurence Φ€

http://www.pbase.com/lmatson/sd9_images
http://www.pbase.com/sigmasd9/root
http://www.pbase.com/cameras/sigma/sd9
http://www.beachbriss.com (eternal test site)

Brian Slater · Sep 10, 2003

Laurence Matson said:
Brian,

Forrest is right about some ridiculous claims. They are everywhere,
however, and are not reserved for addled SD9 users. Much of this
began with the furious attacks back in February 2002, when the
Cannonites lined up their big guns to blow Foveon out of the harbor
before the camera was even launched. Captain G u t t a g was in
charge back then.

I've read some of these, and even as recently as this month heated discussions have included claims such as "all the detail in SD9 images is the imaginary result of aliasing", and "the Foveon phenomenon is just a marketing trick" and so on ad nauseum.

Laurence Matson said:
Otherwise, I am with you. I print A3+ all the time. No noise except
ooos and aaaahs from madame, who finally likes something that I do.
The really big panos cause more consternation: "Looks great. Where
do I put it?", but they are 3 meters long.

--
Laurence ?€

3 meters. No wonder you get oos and ahs; perhaps you have been responding to the spam??

Forrest41926 · Sep 10, 2003

Laurence Matson said:
If you truly believe that other forums are more objective or
whatever, please stick to your beliefs. But the above poster is
quite right in the general tone of this one.

Now, I would never go so far as to say any of the other forums are more objective ... and I try to correct things I see that are patently false, no matter where they're passed off as fact. Quite a bit of this goes on in the Canon forums.

Laurence Matson said:
or other nonsense on to the entire gang of know-nothings here. You
want to be respected as an individual; the feeling is mutual.

You guys are anything but a group of know-nothings; I'm sure I'm hardly the only Canon SLR user who lurks here. Sigma lenses also fit Canon and Nikon mounts, and while there's a "Stigma" reputation for some of the zooms, others compete nicely at a fraction of the cost, and some are downright better than the Canon or Nikon equivelants. None of these lenses are without their issues ( ie filters for the 15-30 ), and you guys are the experts here.

Laurence Matson said:
need links to substantiate any of that, let me know.) Why not just
let us stumble along in the light (can't shoot at night)?

'Cause there's no light to stumble onto. The SD-9 is a dramatically better camera than most digitals on the market. It should trounce the upcoming E-1. But it's not without it's short-comings, either. I just thought it was a little, er, "funny" how one user was treated, for pointing out these short-comings to someone who asked about them.

But thanks for your response!!

Forrest41926 · Sep 10, 2003

Brian Slater said:
I've read some of these, and even as recently as this month heated
discussions have included claims such as "all the detail in SD9
images is the imaginary result of aliasing", and "the Foveon
phenomenon is just a marketing trick" and so on ad nauseum.

Someone should do a db analysis on Phil's forums.

Reading the Canon forums, every week someone jumps in from another forum to tell us all about how superior their camera is, and how we should all dump ours.

One week it's Sony's upcoming 828. After all, it has eight megapixels. Everything else be d@mned. Next week it's the Fuji S2 shooters. They have an interopelated twelve megapixels. But it's a superior form of interopelation, and sharpening artifacts are acceptable anyway ... the S2 is better than the 1Ds. When they get tired of it, it's Nikon's turn. Sure, the D100 has problems, but perceived front/back focus isn't one of them ... so naturally it's better than the 10D. And of course, there's the SD-9. It has 10.5 million pixels. You can resamle it to the 1Ds image size, and have more detail. When you run out of cameras to be superior to the D60, start at the top of the list, and make your way back through them. The SD-9, S2 Pro, and various Sonys are the most common.

Of course, you guys get more than your share of trolls, too. I wonder which forum gets more? Like I said, someone should do a db analysis. It'd be interesting to see...

tammons · Sep 10, 2003

Someone should do some TROLL reasearch and send it to MAD magazine for a writeup. What a bunch of nuts.

I just can't understand people that get off on stirring up trouble over a camera they have never used. Gees talk about a waste of time just to get an EGO boost. It also seems to get worse as the camera forum gets up into the dslr area. Take the Sony forum, mostly everyone is very nice and just out having fun with their cameras and generally happy. Go over to the Nikon or Canon forum and post a legitimate question and get ripped. This forum is pretty defensive, but there are a lot of trolls that feel like it is their right to rip everything about the SD9. I own it now and I like it, although the color is a little shakey. As far as which gets more trolls, I think its probably the Canon forum or another underdog like the oly slr forum. They get ripped pretty hard too, especially about the E1. There are a lot of troll alerts at the Canon forum as well. I owned a 10D (just sold) and I can tell you those people are in denial about the 10D soft images. Just wont admit it. Also I dont think the 10D color is that great either, but everybody over there seems to love it.

I dont know about you, but I have owned 5 different cameras at or over 5mp (+ the SD9) and I have not been 100% satified with any of them. They all have problems including the SD9.

All I want is a camera that does everything I want, but it does not exist except for maybe the 14n or 1ds but even those are a bit slow.

I want all the controls and speed of a Canon 1D, 1.3X crop, Canon or nikon lens mount, true color, not SD9 dull or 10D or S2 overcooked, but actual true color out of the camera, and images as sharp as the SD9, with a max $2500 pricetag. Maybe one day.

Maybe Sigma will come out with a pro model that will resolve some of the SD9 issues with more pixels.

Forrest said:
Forrest41926 said:

I've read some of these, and even as recently as this month heated
discussions have included claims such as "all the detail in SD9
images is the imaginary result of aliasing", and "the Foveon
phenomenon is just a marketing trick" and so on ad nauseum.

Click to expand...

Someone should do a db analysis on Phil's forums.

Reading the Canon forums, every week someone jumps in from another
forum to tell us all about how superior their camera is, and how we
should all dump ours.

One week it's Sony's upcoming 828. After all, it has eight
megapixels. Everything else be d@mned. Next week it's the Fuji S2
shooters. They have an interopelated twelve megapixels. But
it's a superior form of interopelation, and sharpening artifacts
are acceptable anyway ... the S2 is better than the 1Ds. When they
get tired of it, it's Nikon's turn. Sure, the D100 has problems,
but perceived front/back focus isn't one of them ... so naturally
it's better than the 10D. And of course, there's the SD-9. It has
10.5 million pixels. You can resamle it to the 1Ds image size, and
have more detail. When you run out of cameras to be superior to
the D60, start at the top of the list, and make your way back
through them. The SD-9, S2 Pro, and various Sonys are the most
common.

Of course, you guys get more than your share of trolls, too. I
wonder which forum gets more? Like I said, someone should do a db
analysis. It'd be interesting to see...

Cliff. Johnston · Sep 10, 2003

Laurence,

Laurence Matson said:
for the trees.

Forrest, my friend,

Strange as the concept may seem, having the above drummed into you
long enough ... someone cruises in here almost
invariably from our friendly Canon supplier and bashes around for a
while making it clear how dumb we are to have this etc. equipment.
It is sort of like the schoolyard bully routine. Wouldn't you get
tired of that? I sure do, and I can't blame others if they did too.

--
Laurence Φ€

It's known as the Willie Schickelgruber effect (my apologies if the spelling is incorrect). One tells a big lie and then tells it often enough so that eventually people begin to believe it. Then one can bully the minority of choice at the moment to no end. Of course that minority can change on a dime - such are the fickle breezes of fate.

Willie Schickelgruber aka Adolf Hitler.

Cliff.

Arvin Chang · Sep 10, 2003

Erik, you might be trying too hard to convince sg10 with your technical reasoning.

I could be wrong, but I don't think he has much background in computer graphics theory and so doesn't get it when you go into details. I mean I believe it is quite obvious that he doesn't understand the difference between quantization and dynamic range (and how reducing the bit depth changes the first and not the latter).

Stick with simple analogies - the yardstick versus meter stick was an excellent one.

sg10 · Sep 10, 2003

Kevin Blair said:
sg10 said:

This is the root of the misunderstanding, the extremes in a
particular image have nothing to do dynamic range Otherwise, if
you made the first pixel in every image white, and the last pixel
black, instantly all digital cameras would have the same dynamic
range.

Click to expand...

But the ability to render detail in the extremes IS what dynamic
range is all about. Dynamic range isn't about the ability to
differentiate colors in the middle. And no, simply setting the
first pixel to white and the last pixel to black doesn't give all
cameras the same dynamic range.

Ok, there are two misunderstandings, then... :^)

We aren't talking about digital cameras (creator of the message), we are talking about the same image after file format (the messenger) conversion, one format with an inherent dynamic range limit much lower than the other.

Kevin Blair said:
Kevin Blair said:

The reason you go from 3M unique colors in an SD-9 16-bit TIF to
300K in the best quality JPEG, is because now huge numbers of
pixels are simply the same color. You can't portray things well
same-color stuff.

Click to expand...

If you actually had 3M unique colors, the resulting JPEG would be
larger than 300K.

But those are real life numbers... try it.

You'll discover what I'm saying is quite true about JPEG. Even you convert a 24-bit TIF to a "24-bit" (effectively less, I'd argue) JPEG, you'll lose (merge) 1/2 to 2/3rds of the unique colors in the image. 36-bit to JPEG is heinous.

I've posted on this before here, with precise color counts.

Kevin Blair said:
Kevin Blair said:

Question: Which rendered image has better dynamic range?

Click to expand...

Answer: There's probably not that much dynamic range in either
photo so the difference, if any, between the two is slight.

There's a couple of errors in your thought experiment. First,
you're confusing bit depth with dynamic range.

It is true those two things can be confused, but not in this case.

Here's is how you confuse them, you convert a JPEG to a 48-bit format, thinking your dynamic range goes back to where it was orinigally. It doesn't, because now altough many more "place holders" exist, none in between the lower bit depth will bill filled with actual information.

Even if you then JPEG the new file, which will have the effect of a "unique color explosion," like JPEG'ing a JPEG often does--that's all noise, not optical.

Kevin Blair said:
Although that shiny
new SD-9 may use 12 bit A/D converters, the underlying sensor
probably has around 10 dB of dynamic range. Which makes sense since
they probably wanted to oversample to reduce measurement error. So
instead of 4096 unique values, you're probably getting closer to
1024.

The second error is your assumption that by going from a 12 bit
file to a 16 bit file, you're getting additional information.

My nums were based on actual counts for 12-bits per channel (36-bit color). I had no choice but to use a 16-bit TIF (48-bit color), as its the only messenger with enough resolution (in the signal carrier sense) to convey the message.

And more importantly, its the only other option in SPP. :^)

Kevin Blair said:
You're not. (Yes, a 16 bit file has the potential to display 64K
unique colors. But remember, you're only starting with somewhere
around 1000 colors from the sensor.) Let's assume that it actually
does give you 10 dB. Those 7 JPEG colors would represent somewhere
around 30 colors in the original file. You might be able to get an
image, but it might not look so beautiful.

Third, if the girl was standing behind you, you probably got your
reflection and not her's. Bummer!

She's obviously leaning over...

...can't you even see that??? :^)

Kevin Blair said:
Kevin

p.s. Very nice picture. Please don't take my discussion about
dynamic range to mean that the SD-9 can't take good pictures.

Thnks, and obviously all DSLR's can take RAW (though I would argue that the SD-9 has a very high quality, very fast, very low bandwith workflow if you intend to excercise that option--so don't make me do it ...so its a generic discussion about the real benefits of JPEG'ing. Its an interesting discussion, I think, because I submit that shooting medium res RAW makes much more sense than in-camera JPEG, purely from a bandwidth/quality POV. Obviously if you need to pop your CF card in your iToaster and view the image, RAW doesn't give you any love.

--
http://www.pbase.com/imageprocessing/sd9

David Hughes · Sep 10, 2003

Lapsing into aesthetics for a moment, the photo of the little girl is wonderful. It doesn't get much better than that.

Ok, back on topic....

sg10 said:
Presto...

http://www.pbase.com/image/19614399/original.jpg

Question: Which rendered image has better dynamic range?

--
http://david.oldcolo.com/gallery/sd9
http://www.pbase.com/sigmasd9

sg10 · Sep 10, 2003

sg10 said:
sg10 said:

p.s. Very nice picture. Please don't take my discussion about
dynamic range to mean that the SD-9 can't take good pictures.

Click to expand...

Thnks, and obviously all DSLR's can take RAW (though I would argue
that the SD-9 has a very high quality, very fast, very low bandwith
workflow if you intend to excercise that option--so don't make me
do it ...so its a generic discussion about the real benefits of
JPEG'ing. Its an interesting discussion, I think, because I submit
that shooting medium res RAW makes much more sense than in-camera
JPEG, purely from a bandwidth/quality POV. Obviously if you need
to pop your CF card in your iToaster and view the image, RAW
doesn't give you any love.

Ooops, that's not exactly what I meant to say.

What I meant was "SD-9 medium res" (6MP-interpolated by SPP to pixels with 25% full color data, but color data with 4096x the resolution) is a much higher end-quality option than the "10D's 6MP JPEG" (which is also interpolated, but to pixels with 33% full color data, and from color data with 1/4096th the color data resolution).

sg10 · Sep 10, 2003

sg10 said:
sg10 said:

sg10 said:

p.s. Very nice picture. Please don't take my discussion about
dynamic range to mean that the SD-9 can't take good pictures.

Click to expand...

Thnks, and obviously all DSLR's can take RAW (though I would argue
that the SD-9 has a very high quality, very fast, very low bandwith
workflow if you intend to excercise that option--so don't make me
do it ...so its a generic discussion about the real benefits of
JPEG'ing. Its an interesting discussion, I think, because I submit
that shooting medium res RAW makes much more sense than in-camera
JPEG, purely from a bandwidth/quality POV. Obviously if you need
to pop your CF card in your iToaster and view the image, RAW
doesn't give you any love.

Click to expand...

Ooops, that's not exactly what I meant to say.

What I meant was "SD-9 medium res" (6MP-interpolated by SPP to
pixels with 25% full color data, but color data with 4096x the
resolution) is a much higher end-quality option than the "10D's 6MP
JPEG" (which is also interpolated, but to pixels with 33% full
color data, and from color data with 1/4096th the color data
resolution).

Ahhh, its "to" color data with 1/4096th the resolution, not "from."

--
http://www.pbase.com/imageprocessing/sd9

Jun2 · Sep 10, 2003

well said. I tryed Color Mechanic pro (demo version) you mentioned in another forum. It is actually not too hard to adjust color much easier than PS. Without that software, I can understand why you like SD9. Just like you said, every camera has its weakness. We need solutions to overcome the weakness. I respect you a lot as you are one of those in the forum come up with solutions instead of endless arrguing which camera is better.

Jun

tammons said:
I dont know about you, but I have owned 5 different cameras at or
over 5mp (+ the SD9) and I have not been 100% satified with any of
them. They all have problems including the SD9.

All I want is a camera that does everything I want, but it does not
exist except for maybe the 14n or 1ds but even those are a bit slow.

I want all the controls and speed of a Canon 1D, 1.3X crop, Canon
or nikon lens mount, true color, not SD9 dull or 10D or S2
overcooked, but actual true color out of the camera, and images as
sharp as the SD9, with a max $2500 pricetag. Maybe one day.

Maybe Sigma will come out with a pro model that will resolve some
of the SD9 issues with more pixels.

Forrest said:

tammons said:

I've read some of these, and even as recently as this month heated
discussions have included claims such as "all the detail in SD9
images is the imaginary result of aliasing", and "the Foveon
phenomenon is just a marketing trick" and so on ad nauseum.

Click to expand...

Someone should do a db analysis on Phil's forums.

Reading the Canon forums, every week someone jumps in from another
forum to tell us all about how superior their camera is, and how we
should all dump ours.

One week it's Sony's upcoming 828. After all, it has eight
megapixels. Everything else be d@mned. Next week it's the Fuji S2
shooters. They have an interopelated twelve megapixels. But
it's a superior form of interopelation, and sharpening artifacts
are acceptable anyway ... the S2 is better than the 1Ds. When they
get tired of it, it's Nikon's turn. Sure, the D100 has problems,
but perceived front/back focus isn't one of them ... so naturally
it's better than the 10D. And of course, there's the SD-9. It has
10.5 million pixels. You can resamle it to the 1Ds image size, and
have more detail. When you run out of cameras to be superior to
the D60, start at the top of the list, and make your way back
through them. The SD-9, S2 Pro, and various Sonys are the most
common.

Of course, you guys get more than your share of trolls, too. I
wonder which forum gets more? Like I said, someone should do a db
analysis. It'd be interesting to see...

Click to expand...

Kevin Blair · Sep 10, 2003

sg10 said:
Ok, there are two misunderstandings, then... :^)

We aren't talking about digital cameras (creator of the message),
we are talking about the same image after file format (the
messenger) conversion, one format with an inherent dynamic range
limit much lower than the other.

But your whole argument up to this point has been that because the SD-9 only saves in RAW that the resulting image will have a greater dynamic range than a camera that saves in JPEG. That just isn't true. The sensor has a much greater influence on dynamic range than the file format. Note, I'm not trying to say any particular camera has better dynamic range than another. Just that your assumption is incorrect.

sg10 said:
But those are real life numbers... try it.

You'll discover what I'm saying is quite true about JPEG. Even you
convert a 24-bit TIF to a "24-bit" (effectively less, I'd argue)
JPEG, you'll lose (merge) 1/2 to 2/3rds of the unique colors in the
image. 36-bit to JPEG is heinous.

I've posted on this before here, with precise color counts.

Unfortunately, I don't have the ability to generate an image with 3M unique colors. I'd be interested in your numbers since I believe it's difficult to take a photograph of nature and get 3M unique colors in a 3M pixel image. I'm assuming that you generated a gradient in Photoshop that resulted in each pixel having a different color value. Correct?

The second point is that it's highly unlikely that any DSLR is going to give you more than 10 bits of dynamic range. It's might be possible in some medium format backs, but then you'll really be out of our price range. Just because you save the file as a 36-bit RAW or 48-bit TIF file doesn't mean that you have that amount of information.

sg10 said:
It is true those two things can be confused, but not in this case.

Here's is how you confuse them, you convert a JPEG to a 48-bit
format, thinking your dynamic range goes back to where it was
orinigally. It doesn't, because now altough many more "place
holders" exist, none in between the lower bit depth will bill
filled with actual information.

Even if you then JPEG the new file, which will have the effect of a
"unique color explosion," like JPEG'ing a JPEG often does--that's
all noise, not optical.

If the original image didn't have a lot of dynamic range, then it doesn't make any difference what you do to the file afterwards or what the original format was. If it didn't have it to begin with, it's not going to have it after you manupulated it.

To your specific point about converting an 8-bit file to 12 bits or 16 bits, that's what I've been trying to get across from the beginning. If the camera only produces an image with 10 bits of information, then it doesn't make any difference whether you save it as a 12-bit RAW file or a 16-bit TIF file. It only has the information that it started with. And going from 10 bits to 8 bits, I believe you don't through away enough information to significantly impact the dynamic range of the image. (You can produce some compression artifacts, but that's a different subject.) More importantly, if the sensor isn't able to distinguish between the information at the extremes, then it really doesn't make ANY difference what format you use.

sg10 said:
My nums were based on actual counts for 12-bits per channel (36-bit
color). I had no choice but to use a 16-bit TIF (48-bit color), as
its the only messenger with enough resolution (in the signal
carrier sense) to convey the message.

And more importantly, its the only other option in SPP. :^)

See above.

sg10 said:
She's obviously leaning over...

...can't you even see that??? :^)

If that was the case, I probably wouldn't have been able to focus on the taking the picture!

What are SD9's shortcomings? Objective comments please.THANKS!

Erik Magnuson

Guest

Forum Enthusiast

Well-known member

Forum Pro

Forum Pro

Forum Enthusiast

Forum Pro

Forum Pro

Forum Enthusiast

Forum Pro

Forum Pro

Veteran Member

Senior Member

Senior Member

Senior Member

Veteran Member

Senior Member

Senior Member

Senior Member

Well-known member

Similar threads

Keyboard shortcuts