36MB Foveon proved too pricey for Hasselblad

Discussion in 'Digital Photography' started by George Preddy, May 28, 2004.

    1. Advertising

  1. George Preddy

    Lionel Guest

    Kibo informs me that (George Preddy) stated
    that:

    >http://www.hasselblad.se/news/newsItem.asp?secId=312&itemId=249&iShowYear=2000


    "The Hasselblad DFinity changes this situation by integrating Foveon
    prism technology with three of its CMOS image sensors to improve color
    reproduction while providing protecting against moiré and other unwanted
    imaging artifacts found in the singlesensor cameras and backs. "

    I notice that they're talking about one of Foveon's older (late 2000)
    professional prism-based 3 sensor systems, not about consumer-level
    cameras like the SD9.

    --
    W
    . | ,. w , "Some people are alive only because
    \|/ \|/ it is illegal to kill them." Perna condita delenda est
    ---^----^---------------------------------------------------------------
     
    Lionel, May 28, 2004
    #2
    1. Advertising

  2. Re: Reading comprehension proved too difficult for Preddy

    George Preddy wrote:

    ???
     
    Jukka-Pekka Suominen, May 28, 2004
    #3
  3. Anti Spam & Uncle Lart, May 28, 2004
    #4
  4. George Preddy

    Mick Sterbs Guest

    Mick Sterbs, May 28, 2004
    #5
  5. George Preddy

    Tony Spadaro Guest

    Tony Spadaro, May 28, 2004
    #6
  6. George Preddy

    Crownfield Guest

    George Preddy wrote:
    >
    > http://www.hasselblad.se/news/newsItem.asp?secId=312&itemId=249&iShowYear=2000


    try to buy one.

    they could not produce one with acceptable quality or price.

    if they could make a 20mp foveon back for $10,000,
    they could sell all they could make.

    the problem is that a quality 22mp foveon back
    that produces quality images does not exist.

    22 mp backs that do include phase one and imacon.

    imacons ixpress will produce 22x3 megapixel images.
    not interpolated, george, 22mp x 3.
     
    Crownfield, May 28, 2004
    #7
  7. Crownfield <> wrote in message news:<>...
    > George Preddy wrote:
    > >
    > > http://www.hasselblad.se/news/newsItem.asp?secId=312&itemId=249&iShowYear=2000

    >
    > try to buy one.
    >
    > they could not produce one with acceptable quality or price.


    They produced several prototypes, those who used them said they were
    the best digital photographic tools ever created. It is true that
    Hasselblad couldn't afford Foveon, so the deal ultimately fell
    through.
     
    Georgette Preddy, May 29, 2004
    #8
  8. "Georgette Preddy" <> wrote in message
    news:...
    > Crownfield <> wrote in message

    news:<>...
    > > George Preddy wrote:
    > > >
    > > >

    http://www.hasselblad.se/news/newsItem.asp?secId=312&itemId=249&iShowYear=2000
    > >
    > > try to buy one.
    > >
    > > they could not produce one with acceptable quality or price.

    >
    > They produced several prototypes, those who used them said they were
    > the best digital photographic tools ever created. It is true that
    > Hasselblad couldn't afford Foveon, so the deal ultimately fell
    > through.


    Actually truth be known the system was to expensive for end users. It was
    clumsy having to be tethered to the computer. And the system had technical
    problems.
     
    Anti Spam & Uncle Lart, May 29, 2004
    #9
  9. "Tony Spadaro" <> wrote in message
    news:kiKtc.22$...
    > It simply never worked.


    The Hasselblad system may not have worked, but there were working prism
    based Foveon cameras with three 4 megasensor sensors.

    Also, the original Foveon, prism-based, studio cameras didn't suffer from
    the color problems inherent in the X3 sensor. The color separation by prism
    was much better than the color separation by silicon, without the color
    contamination problems.

    The original Foveon system used three sensor elements each with 4
    megasensors. The color information was combined to form an image with 4
    megapixel spatial resolution, with each pixel being full color. At the time
    it was introduced, it was inarguably the highest resolution digital studio
    camera on the market.

    Alas, the prism approach was too expensive, and too heavy, to be usable in a
    non-studio camera, hence the development of the less-capable, but cheaper,
    X3 sensor.

    With the development of low noise, full-frame, Bayer sensors, such as the 11
    megapixel sensor used in the Canon EOS-1Ds, the Foveon approach lost its
    appeal. The Canon sensor, in the EOS-1Ds, is more than 3x the resolution of
    the Foveon X3 sensor, with better color and lower noise, at higher ISOs.

    I don't know where someone came up with 36 megapixels! Even if you counted
    each sensor, in each element, as a pixel (which is not an accurate
    definition of a pixel), it would still only be 12 megasensors (4 megapixel
    spatial). I think that he didn't realize that the Hasselblad site had
    already mis-stated the resolution as being 12 megapixels rather than 12
    megasensors, so he multiplied 12 megasensors by 3. The Foveon sensor gets
    more megapixels every week. Amazing!

    Visit: http://sigmasd10.com. This is the premier site for unbiased
    information about the pros and cons of the Sigma SD10 digital SLR.
     
    Steven M. Scharf, May 29, 2004
    #10
  10. George Preddy

    Crownfield Guest

    Georgette Preddy wrote:
    >
    > Crownfield <> wrote in message news:<>...
    > > George Preddy wrote:
    > > >
    > > > http://www.hasselblad.se/news/newsItem.asp?secId=312&itemId=249&iShowYear=2000

    > >
    > > try to buy one.
    > >
    > > they could not produce one with acceptable quality or price.

    >
    > They produced several prototypes, those who used them said they were
    > the best digital photographic tools ever created. It is true that
    > Hasselblad couldn't afford Foveon, so the deal ultimately fell
    > through.


    right.

    so now the back for the blads are phase one and imacon:
    p20 and ixpress 528c.

    they produce images at 22 mp x 3 colors.
    16 bits worth.
    no interpolation needed.

    and the colors are correct.

    they are worth about 35 sd 10 cameras.
     
    Crownfield, May 29, 2004
    #11
  11. George Preddy

    Crownfield Guest

    Crownfield wrote:
    >
    > Georgette Preddy wrote:
    > >It is true that
    > > Hasselblad couldn't afford Foveon,


    right.

    some lenses for the blad cost more
    than six whole sigma cameras.


    >
    > right.
    >
    > so now the back for the blads are phase one and imacon:
    > p20 and ixpress 528c.
    >
    > they produce images at 22 mp x 3 colors.
    > 16 bits worth.
    > no interpolation needed.
    >
    > and the colors are correct.
    >
    > they are worth about 35 sd 10 cameras.
     
    Crownfield, May 30, 2004
    #12
  12. George Preddy

    Dave Haynie Guest

    On Sat, 29 May 2004 04:38:53 GMT, "Steven M. Scharf"
    <> wrote:

    >
    >"Tony Spadaro" <> wrote in message
    >news:kiKtc.22$...
    >> It simply never worked.


    >The Hasselblad system may not have worked, but there were working prism
    >based Foveon cameras with three 4 megasensor sensors.


    It's a total no-brainer to hook up 3 monochrome sensors, put in a
    diachroic prism, and have a real (eg, not Bayer, not X3) color imaging
    system. EVERY pro-class video camera, and even some prosumer (under
    $1000) do this. Minolta (of all companies) used to sell a DSLR (using
    the lenses from their short-lived APS system, as I recall) with a 3
    CCD system. It's just the issue of size, expense, and lack-of-need
    (eg, humans just aren't that sensitive to color information vs. pixel
    information -- which is why EVERY video system on the planet
    subsamples color in some way or another: in DV, DVD, and HDTV it's 4:1
    subsampled) that prevents it from being more prevalent.

    But it makes sense you'd see the MF and LF folks messing around with
    this. After all, they're all about excess :)

    >Also, the original Foveon, prism-based, studio cameras didn't suffer from
    >the color problems inherent in the X3 sensor. The color separation by prism
    >was much better than the color separation by silicon, without the color
    >contamination problems.


    The prism system may have been bundled that way by Foveon, but that
    was unnecessary -- any camera maker could integrate their own prism
    and sensor set. Foveon was, on the other hand, the first company
    selling large, good quality CMOS sensors commercially (Canon has had
    their own in-house technology, Nikon does too, but these are not
    available on the open market). They did the first 16Mpixel CMOS, too.

    >Alas, the prism approach was too expensive, and too heavy, to be usable in a
    >non-studio camera, hence the development of the less-capable, but cheaper,
    >X3 sensor.


    That sounds like nonsense. Maybe too large for a conventional 35mm
    replacement, but as a medium format replacement? Look up the Sony
    PDX-10 or TRV-950, or the Panasonic DV-953, etc. These are not SLR
    shaped, but they're not out of the ballpark on size/weight relative to
    a MF camera, or even DSLR. Yes, the CCDs are smaller, but it could all
    scale without getting crazy. I think the real reason is the simple
    fact that no one would see any difference, once you got to the
    10Mpixel-or-so level of pro-class single-CCD cameras today.

    I'm not talking about "under the microscope" differences -- they're
    there, they were there before Sony sold the very first electronic
    still camera, there were there back when the world was fixed focal
    lengths and Rangefinders. But in practical terms, when you view a
    printed photo from a normal distance, you'll see a loss of spatial
    frequency quite easily (that, in a nutshell, is also the only
    difference between LF, MF, and 35mm on film). You won't ever see
    interpolated color; beyond a certain resolution, it's simply
    invisible. So, while at the 0.3Mpixel it's an issue (the jury's still
    out on HD; the first HD camcorder under $5K is a single-CCD unit,
    albeit one with a unique color mask), it's not an issue at the high
    end.

    >With the development of low noise, full-frame, Bayer sensors, such as the 11
    >megapixel sensor used in the Canon EOS-1Ds, the Foveon approach lost its
    >appeal. The Canon sensor, in the EOS-1Ds, is more than 3x the resolution of
    >the Foveon X3 sensor, with better color and lower noise, at higher ISOs.


    Yup. As a human, you won't see the effect of the Bayer pattern. The
    loss of spatial resolution, on the other hand, is dramatic between the
    two, and quite visible. As are color sampling and noise issues.
    >
    >I don't know where someone came up with 36 megapixels!


    Foveon numbers are all fictional. Worse yet, the uneducated fools who
    seem to push this technology here (clearly, without understanding it)
    don't understand what a pixel is. Pixels are, by very definition,
    spatial elements (unless you work for IBM, then you have to say
    "PEL"). I started into computing in 1973, and that's what the term
    meant then, as now. Only the marketing people at Foveon and/or Sigma
    seem to believe that a pixel with full RGB information is somehow
    actually three pixels. Worst yet, while accepting this dogmatically,
    the cheerleaders around here want to divide a Bayer (or modified
    Bayer; many camera these days run RGB[something else] or CMYG sensors)
    by 3 or 4. And they can't even understand the obvious contradictions?
    They also simply don't understand the way human color vision works, or
    they would either stop bashing Bayer sensors, or toss out all their
    DVD, VHS, and televisions, which do dramatically worse things to color
    than any Bayer sensor had yet dreamt of.


    Dave Haynie | Chief Toady, Frog Pond Media Consulting
    | Take Back Freedom! Bush no more in 2004!
    "Deathbed Vigil" now on DVD! See http://www.frogpondmedia.com
     
    Dave Haynie, Jun 3, 2004
    #13
  13. George Preddy

    Annika1980 Guest

    >From: Lionel

    >I notice that they're talking about one of Foveon's older (late 2000)
    >professional prism-based 3 sensor systems, not about consumer-level
    >cameras like the SD9.


    Going from Hasselblad to Sigma is kinda like dumping your Bentley for a Ford
    Pinto.
     
    Annika1980, Jun 4, 2004
    #14
  14. George Preddy

    Annika1980 Guest

    >From: Crownfield

    >imacons ixpress will produce 22x3 megapixel images.
    >not interpolated, george, 22mp x 3.


    The Betterlight scanning backs will produce images of 6000x8000 and larger.
    That's 48MP of 100% RGB data uninterpolated.
     
    Annika1980, Jun 4, 2004
    #15
  15. George Preddy

    Crownfield Guest

    Annika1980 wrote:
    >
    > >From: Crownfield

    >
    > >imacons ixpress will produce 22x3 megapixel images.
    > >not interpolated, george, 22mp x 3.

    >
    > The Betterlight scanning backs will produce images of 6000x8000 and larger.
    > That's 48MP of 100% RGB data uninterpolated.


    but try using a scanning back on living moving subjects.

    what makes ixpress interesting is that
    it give you 22mp interpolated for moving subjects,
    and 88 mp non interpolated for still subjects.

    sort of the best single solution to two problems.
     
    Crownfield, Jun 4, 2004
    #16
  16. "Dave Haynie" < <mailto:>> wrote in
    message

    <snip>

    > That sounds like nonsense. Maybe too large for a conventional 35mm
    > replacement, but as a medium format replacement?


    Exactly. Too large for 35mm. Foveon's intial system couldn't easily be
    ported to 35mm or smaller bodies. To get any volume they had to do something
    like X3, especially since they knew that Bayer technology was resulting in
    larger and larger sensors with higher resolution, that would soon make their
    complex, and expensive system, undesirable.

    > Foveon numbers are all fictional. Worse yet, the uneducated fools who
    > seem to push this technology here (clearly, without understanding it)
    > don't understand what a pixel is.


    People like "Preddy" refuse to learn. And when the whole Sigma/Foveon house
    of cards tumbles to the ground, they'll invent a grand conspiracy to explain
    it all.

    > Pixels are, by very definition,
    > spatial elements (unless you work for IBM, then you have to say
    > "PEL").


    Ah, IBM-Speak. Excuse me why I delete some datasets from my hardfile, and
    swap out the planar in my computer.
     
    Steven M. Scharf, Jun 4, 2004
    #17
  17. "Annika1980" <> wrote in message
    news:...
    > >From: Lionel

    >
    > >I notice that they're talking about one of Foveon's older (late 2000)
    > >professional prism-based 3 sensor systems, not about consumer-level
    > >cameras like the SD9.

    >
    > Going from Hasselblad to Sigma is kinda like dumping your Bentley for a

    Ford
    > Pinto.
    >

    worse than that ;) the 3 chip Dfinity aka Foveon II Studio Camera had 3
    22x22mm chips, a prism for a total of 12 Megapixels
     
    Darrell Larose, Jun 4, 2004
    #18
  18. "Darrell Larose" < <mailto:>> wrote in message
    <news:c_0wc.2457$>...

    > worse than that ;) the 3 chip Dfinity aka Foveon II Studio Camera had 3
    > 22x22mm chips, a prism for a total of 12 Megapixels


    Depends on how the sensors and prism were aligned.

    If each sensor was offset so you got 12 megapixels that were distinct in
    their spatial orientation, and then the color was interpolated, then it
    would be a 12 megapixel, 12 megasensor camera.

    If there the sensors were aligned so you got 4 megapixels, with color
    calculated from the three sensors for each pixel, then it was a 4 megapixel
    camera, just as the SD9 and SD10 are 3.43 megapixel cameras.

    Whether you stack the sensors or use a prism, doesn't make any difference.
    If it's 4 megapixels x 3 sensors/pixel then it's a 12 megasensor camera with
    4 million pixels. If it's 12 megapixels with 1 sensor/pixel, it's a 12
    megapixel camera with 12 million sensors.

    Bayer technology is actually more amazing than Foveon technology, because
    Bayer takes the physiology of human vision into account to enable higher
    resolution and more accurate color. I can imagine Foveon getting into
    television transmission and insisting on tripling the frame rate to have one
    frame/color. You'd increase the complexity for no benefit.

    The whole Foveon story is based on false premises. Resolution is not
    improved by measuring all three colors at every pixel location. Even if
    Foveon was able to solve the issues regarding color accuracy, the premise
    that human vision would be able to tell the difference between interpolated
    color on a 6 Mp Bayer versus a 3.4 Mp Foveon sensor, is inaccurate, unless
    you intentionally create situations to trick the Bayer sensor. Now if you
    compared a perfect 3 Megapixel 3 sensors/pixel chip against a perfect 3
    Megapixel 1 sensors/pixel chip, you'd be able to tell the difference. One of
    Foveon's biggest problems is that they're no longer competing against 3
    Mpixel Bayer, they're competing against 5, 6, 8, and 11 Mp Bayer. Also,
    their fabrication process is not ideal for camera sensors.
     
    Steven M. Scharf, Jun 5, 2004
    #19
  19. "Steven M. Scharf" <> wrote in message
    news:2hkwc.22204$...
    > "Darrell Larose" < <mailto:>> wrote in message
    > <news:c_0wc.2457$>...
    >
    > > worse than that ;) the 3 chip Dfinity aka Foveon II Studio Camera had 3
    > > 22x22mm chips, a prism for a total of 12 Megapixels

    >
    > Depends on how the sensors and prism were aligned.
    >
    > If each sensor was offset so you got 12 megapixels that were distinct in
    > their spatial orientation, and then the color was interpolated, then it
    > would be a 12 megapixel, 12 megasensor camera.
    >
    > If there the sensors were aligned so you got 4 megapixels, with color
    > calculated from the three sensors for each pixel, then it was a 4

    megapixel
    > camera, just as the SD9 and SD10 are 3.43 megapixel cameras.
    >
    > Whether you stack the sensors or use a prism, doesn't make any difference.
    > If it's 4 megapixels x 3 sensors/pixel then it's a 12 megasensor camera

    with
    > 4 million pixels. If it's 12 megapixels with 1 sensor/pixel, it's a 12
    > megapixel camera with 12 million sensors.
    >
    > Bayer technology is actually more amazing than Foveon technology, because
    > Bayer takes the physiology of human vision into account to enable higher
    > resolution and more accurate color. I can imagine Foveon getting into
    > television transmission and insisting on tripling the frame rate to have

    one
    > frame/color. You'd increase the complexity for no benefit.
    >
    > The whole Foveon story is based on false premises. Resolution is not
    > improved by measuring all three colors at every pixel location. Even if
    > Foveon was able to solve the issues regarding color accuracy, the premise
    > that human vision would be able to tell the difference between

    interpolated
    > color on a 6 Mp Bayer versus a 3.4 Mp Foveon sensor, is inaccurate, unless
    > you intentionally create situations to trick the Bayer sensor. Now if you
    > compared a perfect 3 Megapixel 3 sensors/pixel chip against a perfect 3
    > Megapixel 1 sensors/pixel chip, you'd be able to tell the difference. One

    of
    > Foveon's biggest problems is that they're no longer competing against 3
    > Mpixel Bayer, they're competing against 5, 6, 8, and 11 Mp Bayer. Also,
    > their fabrication process is not ideal for camera sensors.
    >

    In reading the Dfinity and Foveon II Studio Camera specs, they aligned the
    prism and 2056x2056 pixel sensors to form an interpolated 12 Megapixel
    colour image, that using preddy math would be 4.3 MP.
     
    Darrell Larose, Jun 5, 2004
    #20
    1. Advertising

Want to reply to this thread or ask your own question?

It takes just 2 minutes to sign up (and it's free!). Just click the sign up button to choose a username and then you can ask your own questions on the forum.
Similar Threads
  1. Darrell Larose

    Re: 36MB Foveon proved too pricey for Hasselblad

    Darrell Larose, Jun 4, 2004, in forum: Digital Photography
    Replies:
    7
    Views:
    929
    Steven M. Scharf
    Jun 9, 2004
  2. SoHillsGuy

    "Allo Allo" R1 in January...bit pricey?

    SoHillsGuy, Oct 4, 2003, in forum: DVD Video
    Replies:
    1
    Views:
    435
    Invid Fan
    Oct 4, 2003
  3. Richard

    HBO dvds. Why so pricey?

    Richard, Jun 18, 2004, in forum: DVD Video
    Replies:
    6
    Views:
    653
  4. Grappletech

    BestBuy's PC stuff is pricey

    Grappletech, Feb 6, 2007, in forum: Computer Information
    Replies:
    0
    Views:
    382
    Grappletech
    Feb 6, 2007
  5. Giuen
    Replies:
    0
    Views:
    1,540
    Giuen
    Sep 12, 2008
Loading...

Share This Page