Reading mode: Light Dark

Foveon/Bayer comparison

Go to last post
sg10 said:
No, its the fact that JPEG truncates the colorspace to 1/4096th of
its original size, then chops half the remaining unque colors and
tosses them.
Click to expand...
jpeg itself doesn't truncate from 12 bit. The standard supports up to 12 bit color, though this is rarely implemented in practice. SPP apparently chooses to reduce to 8 bpp (= 24 bit) before feeding the results to their jpeg code.

To figure out what's lost, you first need to explain what colorspace the original 12 bpp image is in so we can see how the values map to a standard space. Only then can we decide if the extra bits correspond to color distinctions that are visible, or if they correspond to extra information that is either too dark to be visible under standard conditions or too bright. Note that if Sigma is smart, this (too dark/too bright) will be the case since this extra information is needed for exposure compensation. The other possibility is that the extra bit depth corresponds to distinctions that are too subtle to be detected by the naked eye. This would imply less capacity for exposure compensation, but some extra resilience to extreme curve adjustments.

As has already been explained to you, at low compression levels quantization in jpeg will result in color changes that are difficult or impossible for the human eye to detect.
sg10 said:
So is that a "yes," you understand JPEG destroys the original SD-9
photo's colorspace? Ready to move to the next, equally important
concept?
Click to expand...
You still haven't defined the colorspace or explained what it means to destroy it.

--
Ron Parr
FAQ: http://www.cs.duke.edu/~parr/photography/faq.html
Gallery: http://www.pbase.com/parr/
 
sg10 said:
sg10 said:
There is no dithering. Stop implying that there is. It's just lying.
Click to expand...
You said it yourself. You said JPEG was designed to hide the fact
that half the unique colors have been eliminated (i.e. the
remainder after cutting the color space to 1/4096th of its original
size) .
Click to expand...
Ahhh... Here you again reveal your ignorance about dithering. Even if we accept your imprecise statement above, it implies nothing about dithering.
sg10 said:
sg10 said:
Mot jpeg algorithms don't assume 36 bit input. You do realize that
the human eye can only distinguish about 256 different levels of
brightness per channel?
Click to expand...
All this discussion, and you haven't learned a thing...

http://www.pbase.com/image/21560659/original.jpg

There's your "indistinguishable 256 levels.
Click to expand...
Your image is irrelevant to the discussion and your implication that anybody here has anything to learn from you is laughable.

--
Ron Parr
FAQ: http://www.cs.duke.edu/~parr/photography/faq.html
Gallery: http://www.pbase.com/parr/
 
Ron Parr said:
sg10 said:
No, its the fact that JPEG truncates the colorspace to 1/4096th of
its original size, then chops half the remaining unque colors and
tosses them.
Click to expand...
jpeg itself doesn't truncate from 12 bit. The standard supports up
to 12 bit color, though this is rarely implemented in practice.
Click to expand...
Kodak has a patent on extended dynamic range JPEGs (which are still 8-bit, but with additional color data embedded in the remarks header which is processed during rendering using the exclusively-Kodak ERI engine).

The image that follows represents the end of this discussion.

It shows a pure B&W pattern originally saved as a .png, then JPEG's once, then the JPEG was JPEG'd a second time. Just like the SD-9 crops shown were double-JPEG'd. As you can now see, your "analysis" of supposedly subtle moire using a JPEG of a JPEG is an absurd exercise. Sorry to let you in on it. Obviously if this were an actual photo, there wouldn't be only pure black and pure white in the original, and you'd wind up with a veritable rainbow of color artificats as the direct result of JPEG'ing alone.

There is no point in continuing this discussion...

http://www.pbase.com/image/21586643/original
 
Ron Parr said:
sg10 said:
Member said:
There is no dithering. Stop implying that there is. It's just lying.
Click to expand...
You said it yourself. You said JPEG was designed to hide the fact
that half the unique colors have been eliminated (i.e. the
remainder after cutting the color space to 1/4096th of its original
size) .
Click to expand...
Ahhh... Here you again reveal your ignorance about dithering.
Even if we accept your imprecise statement above, it implies
nothing about dithering.
Member said:
Member said:
Mot jpeg algorithms don't assume 36 bit input. You do realize that
the human eye can only distinguish about 256 different levels of
brightness per channel?
Click to expand...
All this discussion, and you haven't learned a thing...

http://www.pbase.com/image/21560659/original.jpg

There's your "indistinguishable 256 levels.
Click to expand...
Your image is irrelevant to the discussion and your implication
that anybody here has anything to learn from you is laughable.

--
Ron Parr
FAQ: http://www.cs.duke.edu/~parr/photography/faq.html
Gallery: http://www.pbase.com/parr/
Click to expand...
--
http://www.pbase.com/imageprocessing/sd9
 
Ron Parr said:
sg10 said:
Member said:
What exactly do you think these show?
Click to expand...
Is that a yes you can see a difference, or no you cannot?
Click to expand...
They are different images.
Click to expand...
The images in both halves of the crops are utterly identical.

One half is the image file displayed in 36-bit color (SD-9 RAW), the other half is the image file displayed in 24-bit color (same as a JPEG before the unique colors are cut in half).
 
sg10 said:
One half is the image file displayed in 36-bit color (SD-9 RAW),
the other half is the image file displayed in 24-bit color (same as
a JPEG before the unique colors are cut in half).
Click to expand...
It is not, you have no idea what you are talking about and what those samples represent.

--
Regards from Old Europe,

Dominic

http://www.pbase.com/sigmasd9/dominic_gross
 
alienmarmoset said:
I can't believe you dragged out these pictues again. Anybody can
see the difference between the left and the right.
Click to expand...
Ron won't commit to the obvious fact that there is any difference at all.

He has figured out that to do so is to admit that JPEG'ing alone creates an image quality difference that "Anybody can see."
alienmarmoset said:
The real
question is, why do you think these pictures have anything to do
with this discussion,
Click to expand...
He said reducing colorspace to 1/4096th of the original is something "no one can see." As you rightly point out, that's bunk.
alienmarmoset said:
JPEG compression, or digital photography??
Click to expand...
JPEG reduces 36-bit images to 24-bit colorspace, then cuts the unique colors in half and adds very visible artifacts, especially aroung sharp lines. To blame artifacts you see in a JPEG'd JPEG of a long-gone 36-bit SD-9 image is utterly absurd.
 
Did he really say that? Or is this what you WANT to read?
Also why do you respond to alienmarmoset with what Ron said?

Another thing to note is that your posting is really bullsh1t, right as well as left are jpeged so the difference you want to make up here does not really exist. The pictures look different because they were rendered differently and not because one is jpeg and the other is not.
sg10 said:
alienmarmoset said:
I can't believe you dragged out these pictues again. Anybody can
see the difference between the left and the right.
Click to expand...
Ron won't commit to the obvious fact that there is any difference
at all.

He has figured out that to do so is to admit that JPEG'ing alone
creates an image quality difference that "Anybody can see."
sg10 said:
The real
question is, why do you think these pictures have anything to do
with this discussion,
Click to expand...
He said reducing colorspace to 1/4096th of the original is
something "no one can see." As you rightly point out, that's bunk.
sg10 said:
JPEG compression, or digital photography??
Click to expand...
JPEG reduces 36-bit images to 24-bit colorspace, then cuts the
unique colors in half and adds very visible artifacts, especially
aroung sharp lines. To blame artifacts you see in a JPEG'd JPEG of
a long-gone 36-bit SD-9 image is utterly absurd.
Click to expand...
--
Regards from Old Europe,

Dominic

http://www.pbase.com/sigmasd9/dominic_gross
 
sg10 said:
It shows a pure B&W pattern originally saved as a .png, then JPEG's
once, then the JPEG was JPEG'd a second time. Just like the SD-9
crops shown were double-JPEG'd. As you can now see, your
"analysis" of supposedly subtle moire using a JPEG of a JPEG is an
absurd exercise. Sorry to let you in on it. Obviously if this
were an actual photo, there wouldn't be only pure black and pure
white in the original, and you'd wind up with a veritable rainbow
of color artificats as the direct result of JPEG'ing alone.
Click to expand...
The fact that you can abuse an image with excessive jpeg compression proves nothing . If you think this is "moire" then we can add "moire" to the list concepts you don't understand.

--
Ron Parr
FAQ: http://www.cs.duke.edu/~parr/photography/faq.html
Gallery: http://www.pbase.com/parr/
 
sg10 said:
alienmarmoset said:
I can't believe you dragged out these pictues again. Anybody can
see the difference between the left and the right.
Click to expand...
Ron won't commit to the obvious fact that there is any difference
at all.

He has figured out that to do so is to admit that JPEG'ing alone
creates an image quality difference that "Anybody can see."
Click to expand...
I guess we can add "compression levels" to the list of things you don't understand now.
sg10 said:
sg10 said:
JPEG compression, or digital photography??
Click to expand...
JPEG reduces 36-bit images to 24-bit colorspace, then cuts the
unique colors in half and adds very visible artifacts, especially
aroung sharp lines. To blame artifacts you see in a JPEG'd JPEG of
a long-gone 36-bit SD-9 image is utterly absurd.
Click to expand...
Get back to us when you've understood compression...

--
Ron Parr
FAQ: http://www.cs.duke.edu/~parr/photography/faq.html
Gallery: http://www.pbase.com/parr/
 
Don't end it here... If you don't reply, I'll start a new thread to question you.

At what JPEG quality did you save the images? I don't have PS on the machine I'm using now, so I saved the PNG file (resized to 25% so as to match the size of the original file you worked on) three times in MS PhotoEditor at 100% quality, and couldn't see any changes at all. Then at 90% quality three times, again no noticable change. Finally, at 80% quality, the change you observed started to appear, but to a less extent. Will you please do the same test again, using JPEG quality of 12 in PS?

Apparently, a low JPEG quality (high compression) may introduce some moire-like pattern. However, it is against basic logic to claim the other way round that moire in SD9 images are all caused by JPEG compression. If the image is saved at the highest JPEG quality (12 or 100%), then moire patterns, if any, are real luma moire.
sg10 said:
The image that follows represents the end of this discussion.

It shows a pure B&W pattern originally saved as a .png, then JPEG's
once, then the JPEG was JPEG'd a second time. Just like the SD-9
crops shown were double-JPEG'd. As you can now see, your
"analysis" of supposedly subtle moire using a JPEG of a JPEG is an
absurd exercise. Sorry to let you in on it. Obviously if this
were an actual photo, there wouldn't be only pure black and pure
white in the original, and you'd wind up with a veritable rainbow
of color artificats as the direct result of JPEG'ing alone.

There is no point in continuing this discussion...

http://www.pbase.com/image/21586643/original
Click to expand...
 
Ron,

See the message I posted below... I've run a test and sort of proved that sg10 must have used a low JPEG quality factor to produce the moire-like pattern. It seems he never (carefully) read Rick Matthews' article, and we can assume he doesn't understand what moire is.

Yi
Ron Parr said:
sg10 said:
It shows a pure B&W pattern originally saved as a .png, then JPEG's
once, then the JPEG was JPEG'd a second time. Just like the SD-9
crops shown were double-JPEG'd. As you can now see, your
"analysis" of supposedly subtle moire using a JPEG of a JPEG is an
absurd exercise. Sorry to let you in on it. Obviously if this
were an actual photo, there wouldn't be only pure black and pure
white in the original, and you'd wind up with a veritable rainbow
of color artificats as the direct result of JPEG'ing alone.
Click to expand...
The fact that you can abuse an image with excessive jpeg
compression proves nothing . If you think this is "moire" then we
can add "moire" to the list concepts you don't understand.

--
Ron Parr
FAQ: http://www.cs.duke.edu/~parr/photography/faq.html
Gallery: http://www.pbase.com/parr/
Click to expand...
 
sg10 said:
alienmarmoset said:
I can't believe you dragged out these pictues again. Anybody can
see the difference between the left and the right.
Click to expand...
Ron won't commit to the obvious fact that there is any difference
at all.

He has figured out that to do so is to admit that JPEG'ing alone
creates an image quality difference that "Anybody can see."
Click to expand...
Ron can speak for himself, but since these images have nothing to do with JPEGing, admitting a quality difference would say nothing of the sort.
sg10 said:
sg10 said:
The real
question is, why do you think these pictures have anything to do
with this discussion,
Click to expand...
He said reducing colorspace to 1/4096th of the original is
something "no one can see." As you rightly point out, that's bunk.
Click to expand...
How do you figure I pointed that out? I said nothing of the sort. What I was trying to point out is that these images have absolutely nothing to do with this discussion at all. Perhaps I should have stated that more clearly.
sg10 said:
sg10 said:
JPEG compression, or digital photography??
Click to expand...
JPEG reduces 36-bit images to 24-bit colorspace, then cuts the
unique colors in half and adds very visible artifacts, especially
aroung sharp lines. To blame artifacts you see in a JPEG'd JPEG of
a long-gone 36-bit SD-9 image is utterly absurd.
Click to expand...
Even if this were true, ding1 kindly posted a link to the 48-bit TIFF that shows the same moire. So clearly it wasn't caused by JPEG compression.
 
ding1 said:
Don't end it here... If you don't reply, I'll start a new thread to
question you.
Click to expand...
Oh no.... :^)
Yi Ding said:
At what JPEG quality did you save the images? I don't have PS on
the machine I'm using now, so I saved the PNG file (resized to 25%
so as to match the size of the original file you worked on) three
times in MS PhotoEditor at 100% quality,
Click to expand...
You are confusing a compression parameter with a quality %. 100 is a parameter, it is not a quality percentage nor was it ever meant to be associated, or proportional in any way to one.
Yi Ding said:
and couldn't see any
changes at all. Then at 90% quality three times, again no noticable
change. Finally, at 80% quality,
Click to expand...
Same remark as above. This is basic stuff. :^(
Yi Ding said:
the change you observed started to
appear, but to a less extent. Will you please do the same test
again, using JPEG quality of 12 in PS?
Click to expand...
That's because you are working with my 400% magnification of the pattern.

The fact is, JPEG reduces the colorspace so dramatically and adds so many aritifacts that there is no possible way to "analyze" subtleties in any image, let alone supposed moire in an SD-9 image. Most savvy discussions about the effect some say they see under certain rare conditions is nothing more than a manifestation of the display media's raster pattern. In fact, if you look at a "perfect" single-pixel B&W checkboard pattern on your monitor right now, you'll see clear rainbowing thoughout. That's not SD-9 moire either.

There are many reasons why posting a triple-JPEG of a single-pixel res color image pattern and blaming the plethora of artifacts on long-gone 36-bit SD-9 color data is silliness. Just to summarize the ones I can think of off the top of my nugget...
  • 36-bit data is was already wiped out, only 1/4096th of the colorspace is there to "analyze."
  • Of the remaining 400K colors (avg-ish), from the orignal 3M (avg-ish) 36-bit image, about half of those are eliminated and the result is dithered to simulate the residue from of the original image after conversion to down to 16.8M color choices from the original 68 Billion.
  • When JPEG'ing a JPEG (even once), the unique color count usually goes up, not down. A lot. This isn't new optical data, it is statistical noise introduced largely in algorithmic/harmonic color patterns.
  • My B&W image showed this vividly using only 2 colors, a JPEG of an SD-9 image is doing its number on around 200K unique colors of the 400K that remained from the 3M originals. So, about 100,000x the color variation. A JPEG of a JPEG is taking those and adding about double the number of unique colors back again (this is noise, not data), and doing the same with what data remains, plus acting on noise.
  • There are serious JPEG artifacts around sharp lines, even at a compression parameter of 100. The SD-9 suffers more than other digital cameras from JPEG'ing because it can actually produce sharp lines.
  • Raster induced color artifacts make it impossible to asses any and all of the above on a computer monitor. Yes, you can assess Bayer moire on a computer monitor, because it is a dominant pattern in the image, just like you can assess a picture of an apple.
If you don't believe me, that's fine, but don't make up your mind until you print at hat image from a 16-bit TIF output from the original RAW file, using a professional 36-bit copy service, at a (crop) size sufficient to see individal pixels as clear single color rectangles. Until you do that, you have no possible way to know if you are right or wrong.
 
sg10 said:
sg10 said:
At what JPEG quality did you save the images? I don't have PS on
the machine I'm using now, so I saved the PNG file (resized to 25%
so as to match the size of the original file you worked on) three
times in MS PhotoEditor at 100% quality,
Click to expand...
You are confusing a compression parameter with a quality %. 100 is
a parameter, it is not a quality percentage nor was it ever meant
to be associated, or proportional in any way to one.
Click to expand...
Well, it was my mistake that I added a "%" mark to the number, which is called "JPEG quality factor" in MS PhotoEditor that ranges from 1 to 100. However, the fact that I have an error doesn't mean you are correct in your claims. This is again a basic logic.

Now, answer me the question: at quality did you process the JPEG files in PS?

An advice: you don't need waste your own time repeating your arguments.
Otherwise it is alway to waste other people's time.
 
Ron Parr said:
Member said:
The images in both halves of the crops are utterly identical.

One half is the image file displayed in 36-bit color (SD-9 RAW),
the other half is the image file displayed in 24-bit color (same as
a JPEG before the unique colors are cut in half).
Click to expand...
So, are you a liar or an idiot?
Click to expand...
Ron, those two parameters are not mutually exclusive. There is a high probability that he is both.

I also believe he is the reason for the recent falloff in the quality of conversation in this group. It's getting hard to have any type of serious conversation without him popping up. He literally mutilated Cliff Johnson's recent testing thread, screwed up a thread on the Sony 4 color sensor (where he kept refering to it as a "new Bayer" or "better Bayer" without any knowledge of what Bayer's pattern is). He will not accept a correction, no matter how many references one posts.

In short, he's interrupting all the adult conversation, and gradually driving everyone out of the room.

And I bet he's sold quite a few Canons and Nikons to people who stop by this forum looking for useful SD-9 information. My God, if that's the kind of person who uses an SD-9....

--
Ciao!

Joe

http://www.swissarmyfork.com
 
You must log in or register to reply here.

Keyboard shortcuts

f
Forum
m
My threads
Mobile site
About
Editorial content
Cameras & Lenses
Community
All content, design, and layout are Copyright © 1998–2025 Digital Photography Review All Rights Reserved.
Reproduction in whole or part in any form or medium without specific written permission is prohibited.
When you use DPReview links to buy products, the site may earn a commission.
©GPS Media - Guides, Products, Services.
Back
Top