Fuji Anti-Blur

Graham Fountain wrote:
> David J Taylor wrote:
>> I am simply asking that Fuji state how they can get such good
>> performance, in a scientific and quantifiable way. So far, they
>> have not done so, unless you know otherwise. As I already said, if
>> you like the results, fine. I am not claiming that Fuji violate the
>> laws of physics, only that they are subject to the same laws as
>> everyone else.
> Fuji do state how they do it. Their cameras include various revisions
> of their SuperCCD technology. SCCD has an octagonal layout rather
> than the traditional square layout, and it allows them to have a
> greater proportion of the sensor area actually used to receive light.
> Each individual sensel is larger than in traditional CCD sensors.
> Therefore subject to the laws of physics, they can achieve a better
> S/N ratio.
> http://www.fujifilm.co.uk/digital/popups/superccd_product_page.html


Thanks for that pointer. I am quite happy to accept that by increasing
the active percentage of each cell, the sensitivity can be improved, but
the article provides no numbers, so it is impossible to predict just what
sensitivity gain can be obtained in this way. It would also be
interesting to know if they have managed to improve quantum efficiency as
well, or done something with the RGB filter (like use a CMY one).

David
 >> Stay informed about: Fuji Anti-Blur 

John Bean wrote:

> On Sat, 15 Jul 2006 17:13:34 -0400, ASAAR <caught.TakeThisOut@22.com>
> wrote:
>>In an earlier era, that led to engineering proofs that
>>bumblebees shouldn't be able to fly. I think we can all agree at
>>this late date that bumblebees don't violate the laws of physics.
>
> That's a rather good example (that I'd forgotten) of poorly
> understood (pseudo) science leading to completely erroneous
> conclusions.

Actually that's an example of an urban legend. The fact is that the
engineer in question did the calculation on a napkin while attending a
dinner party and may very well have been intoxicated at the time--he was as
disturbed by this as anyone else and later did a more detailed calculation
to find out what he had done wrong. There was never any "engineering proof
that bumblebees can't fly".

> I've been trying unsucessfully to find a
> reference I had from the 1960s concerning AF, when a quite
> reputable scientist said that "true" AF as used by our
> brain/eye would never be possible in a camera, it was
> "impossible" for any mechanism to judge when an image was in
> focus and instead would have to use some sort of
> rangefinder. He was confusing available technology with
> science, and the first AF cameras (Polaroid, etc) did indeed
> use sonic distance measurement. Many later AF P&S cameras
> used active IR rangefinders, but most modern cameras use no
> distance measurement at all - they work in just the way he
> said was "impossible".
>
> I'll try to find the reference, it's a reminder to never use
> pseudo-science to claim that anything is "impossible" -
> you'll always look really stupid when technology catches up
> and somebody goes ahead and does it anyway.
>

--
--John
to email, dial "usenet" and validate
(was jclarke at eye bee em dot net)
 >> Stay informed about: Fuji Anti-Blur 

On Sun, 16 Jul 2006 00:50:22 -0400, ASAAR <caught.TakeThisOut@22.com>
wrote:
>a good enough explanation would be "Fuji Inc's researchers have finally
>divined the proper use of arcane magic spells and potions". <g>

Announcement: new F40 can optionally be supplied with solid
gold case transmuted from the standard alloy for an extra
$10... LOL.

I'm just happy it works to be honest, my F10 is the best "go
everywhere" camera I've ever owned despite its limitations.
The F30 addresses some of these but not enough to attract me
to upgrade. Maybe the gold F40 will do it


--
John Bean
 >> Stay informed about: Fuji Anti-Blur 

"ASAAR" <caught DeleteThis @22.com> wrote in message
news:amgjb21o6sfcop0lnoctn995cims5381a9@4ax.com...
> On Sat, 15 Jul 2006 22:59:14 +0100, John Bean wrote:
>
>> I'll try to find the reference, it's a reminder to never use
>> pseudo-science to claim that anything is "impossible" -
>> you'll always look really stupid when technology catches up
>> and somebody goes ahead and does it anyway.
>
> Another example (from my era) involved those that put very low
> limits on the maximum theoretical data transfer rate that modems
> would be able to achieve. Praise be that they were mistaken, since
> I'm not yet using broadband.
<snip>

What you fail to realize is the phone system changed to digital and that is
what allowed higher transfer rates. Note to get the 53KB (in one direction
only), the receiving modem must be connected digitally to the network. If
this were not the case, today's top speeds would be around 34kB/second on
very short non-loaded phone lines.

David
 >> Stay informed about: Fuji Anti-Blur 

On Sun, 16 Jul 2006 13:43:29 GMT, David wrote:

>> Another example (from my era) involved those that put very low
>> limits on the maximum theoretical data transfer rate that modems
>> would be able to achieve. Praise be that they were mistaken, since
>> I'm not yet using broadband.
> <snip>
>
> What you fail to realize is the phone system changed to digital and that is
> what allowed higher transfer rates. Note to get the 53KB (in one direction
> only), the receiving modem must be connected digitally to the network. If
> this were not the case, today's top speeds would be around 34kB/second on
> very short non-loaded phone lines.

Sorry, but what I spoke of occurred long before the phone system
switched to digital equipment. The theories were from the time when
I personally used 110 and 300 baud modems (the first couple had no
crystal oscillators and occasionally needed to have their RC
oscillator sections manually tuned). Many universities had the high
speed 1,200b modems. It was the change to a more complex modem
design that invalidated the theories, which I believe predicted that
the maximum baud rates would have been approx. 2400b. I may be
mistaken about this, but it was certainly well below even 9,600b.
 >> Stay informed about: Fuji Anti-Blur 

"ASAAR" <caught.DeleteThis@22.com> wrote in message
news:emskb25qutm4gvksutlq5jcocm99japgfb@4ax.com...
> On Sun, 16 Jul 2006 13:43:29 GMT, David wrote:
>
>>> Another example (from my era) involved those that put very low
>>> limits on the maximum theoretical data transfer rate that modems
>>> would be able to achieve. Praise be that they were mistaken, since
>>> I'm not yet using broadband.
>> <snip>
>>
>> What you fail to realize is the phone system changed to digital and that
>> is
>> what allowed higher transfer rates. Note to get the 53KB (in one
>> direction
>> only), the receiving modem must be connected digitally to the network. If
>> this were not the case, today's top speeds would be around 34kB/second on
>> very short non-loaded phone lines.
>
> Sorry, but what I spoke of occurred long before the phone system
> switched to digital equipment. The theories were from the time when
> I personally used 110 and 300 baud modems (the first couple had no
> crystal oscillators and occasionally needed to have their RC
> oscillator sections manually tuned). Many universities had the high
> speed 1,200b modems. It was the change to a more complex modem
> design that invalidated the theories, which I believe predicted that
> the maximum baud rates would have been approx. 2400b. I may be
> mistaken about this, but it was certainly well below even 9,600b.
>
I have no idea where these theories of which you speak come from. Shannon's
information theory, which predates modems, surely does not set such a limit.
The information capacity of an analog channel is determined by bandwidth and
signal to noise ratio and that is as valid today as when it was formulated.
Could you cite such a theory stating otherwise? Perhaps they were
calculating the maximum rate for the absolute worst case transcontinental
telephone connection which was allot worse than the connections of today
which are digital except for the local loop. If you use the worst case all
analog connection of yesteryear, then you can kiss 56K modems goodbye. No
modulation technique can overcome the basic limits on analog channel
capacity.

David
 >> Stay informed about: Fuji Anti-Blur 

In article <56vug.67098$fb2.31821@newssvr27.news.prodigy.net>,
David <someone.DeleteThis@some-where.com> wrote:

>I have no idea where these theories of which you speak come from. Shannon's
>information theory, which predates modems, surely does not set such a limit.
>The information capacity of an analog channel is determined by bandwidth and
>signal to noise ratio and that is as valid today as when it was formulated.
>Could you cite such a theory stating otherwise? Perhaps they were
>calculating the maximum rate for the absolute worst case transcontinental
>telephone connection which was allot worse than the connections of today
>which are digital except for the local loop.

Regardless, wouldn't the local loop be the ceiling?

>If you use the worst case all
>analog connection of yesteryear, then you can kiss 56K modems goodbye. No
>modulation technique can overcome the basic limits on analog channel
>capacity.
>
>David
>
 >> Stay informed about: Fuji Anti-Blur 

On Sun, 16 Jul 2006 17:57:21 GMT, David wrote:

> I have no idea where these theories of which you speak come from.

No doubt.


> Could you cite such a theory stating otherwise? Perhaps they were
> calculating the maximum rate for the absolute worst case transcontinental
> telephone connection which was allot worse than the connections of today
> which are digital except for the local loop.

No. What I stated was from memory going back at least 30 years.
I searched briefly and didn't see anything. If you're really
interested it would be just as easy for you to search as it would be
for me to do it, and if you remain unconvinced, that's fine with me.
Otherwise, the little that I can recall may help (but it's kind of
vague after so many years) and it was that the theoretical maximum
baud rate was valid only until modems went beyond frequency shift
and phase shift keying, which allowed speeds up to about 1200b.
More complex designs (the word "quadrature" comes to mind) was part
of the breakthrough, and soon modem speeds went to 2400b, 9600b and
beyond. These modems all sent analog signals, but the use of
digital lines in the CO (as you mentioned) was later needed to get
the highest speeds, but as I said, what was erroneously thought to
be the maximum possible data rate using modems was from an earlier
(by decades) era, probably dating back to the 60's or earlier.
 >> Stay informed about: Fuji Anti-Blur 

"tjab" <tjab DeleteThis @wam.umd.edu> wrote in message
news:e9e47h$97c@rac3.wam.umd.edu...
> In article <56vug.67098$fb2.31821@newssvr27.news.prodigy.net>,
> David <someone DeleteThis @some-where.com> wrote:
>
>>I have no idea where these theories of which you speak come from.
>>Shannon's
>>information theory, which predates modems, surely does not set such a
>>limit.
>>The information capacity of an analog channel is determined by bandwidth
>>and
>>signal to noise ratio and that is as valid today as when it was
>>formulated.
>>Could you cite such a theory stating otherwise? Perhaps they were
>>calculating the maximum rate for the absolute worst case transcontinental
>>telephone connection which was allot worse than the connections of today
>>which are digital except for the local loop.
>
> Regardless, wouldn't the local loop be the ceiling?

Actually, for local loops under a mile, the performance is quite good. DSL
with several megabits per second illustrates that. The real killer was the
noise of long distance analog carrier circuits where 10s of not 100s of
repeaters sat between you and the other end. They each added a bit of noise,
frequency, and phase distortion. Digital transmission with its ability to
regenerate a fresh copy of the bit stream at repeater points solved all of
those problems. Remember the 'you can hear a pin drop' commercials for
Sprint? Us older folks remember what long distance transmission sounded like
in the days of 110 baud modems.

David

>
>>If you use the worst case all
>>analog connection of yesteryear, then you can kiss 56K modems goodbye. No
>>modulation technique can overcome the basic limits on analog channel
>>capacity.
>>
>>David
>>
 >> Stay informed about: Fuji Anti-Blur 

"ASAAR" <caught.RemoveThis@22.com> wrote in message
news:uv9lb2lfr2d1qfcal79oircd9p0bh2e23e@4ax.com...
> On Sun, 16 Jul 2006 17:57:21 GMT, David wrote:
>
>> I have no idea where these theories of which you speak come from.
>
> No doubt.
>
>
>> Could you cite such a theory stating otherwise? Perhaps they were
>> calculating the maximum rate for the absolute worst case transcontinental
>> telephone connection which was allot worse than the connections of today
>> which are digital except for the local loop.
>
> No. What I stated was from memory going back at least 30 years.
> I searched briefly and didn't see anything. If you're really
> interested it would be just as easy for you to search as it would be
> for me to do it, and if you remain unconvinced, that's fine with me.
> Otherwise, the little that I can recall may help (but it's kind of
> vague after so many years) and it was that the theoretical maximum
> baud rate was valid only until modems went beyond frequency shift
> and phase shift keying, which allowed speeds up to about 1200b.
> More complex designs (the word "quadrature" comes to mind) was part
> of the breakthrough, and soon modem speeds went to 2400b, 9600b and
> beyond. These modems all sent analog signals, but the use of
> digital lines in the CO (as you mentioned) was later needed to get
> the highest speeds, but as I said, what was erroneously thought to
> be the maximum possible data rate using modems was from an earlier
> (by decades) era, probably dating back to the 60's or earlier.

Your recollections are correct, but I was only questioning your assertion of
faulty theories to explain your original statements about maximum modem bit
rates. The theories were not faulty. For many years, engineers knew they
were nowhere near the capacity predicted by Shannon using PSK and FSK. It
took new modulation methods such as QAM, OFDM, and active equalization and
echo canceling to approach the theoretical bit rates. The available
bandwidth was already cut in half to accommodate bi-directional
communication since in those days, dynamic equalization and echo cancelling
would have priced modems out of reach and taken racks of equipment. Also, at
the time, the Bell system was very conservative and the mantra was that the
modem should work on any connection even though faster ones could be built
that worked under more ideal conditions.

David
 >> Stay informed about: Fuji Anti-Blur 

David J Taylor wrote:
> I still believe that their "anti-blur" advertising is designed to confuse
> punters into believing that the Fuji camera has a feature (image
> stabilisation) which it does not.

Yes, I was initially fooled to think that it has actual stabalization,
but on further reading realize it may not be the case, so decided to
ask on here. Based on the comments, I will not be considering the S5200
as I am looking for IS. thanks
 >> Stay informed about: Fuji Anti-Blur 

David J Taylor wrote:
> Graham Fountain wrote:
>> David J Taylor wrote:
>>> I am simply asking that Fuji state how they can get such good
>>> performance, in a scientific and quantifiable way. So far, they
>>> have not done so, unless you know otherwise. As I already said, if
>>> you like the results, fine. I am not claiming that Fuji violate the
>>> laws of physics, only that they are subject to the same laws as
>>> everyone else.
>> Fuji do state how they do it. Their cameras include various revisions
>> of their SuperCCD technology. SCCD has an octagonal layout rather
>> than the traditional square layout, and it allows them to have a
>> greater proportion of the sensor area actually used to receive light.
>> Each individual sensel is larger than in traditional CCD sensors.
>> Therefore subject to the laws of physics, they can achieve a better
>> S/N ratio.
>> http://www.fujifilm.co.uk/digital/popups/superccd_product_page.html
>
>
> Thanks for that pointer. I am quite happy to accept that by increasing
> the active percentage of each cell, the sensitivity can be improved, but
> the article provides no numbers, so it is impossible to predict just what
> sensitivity gain can be obtained in this way. It would also be
> interesting to know if they have managed to improve quantum efficiency as
> well, or done something with the RGB filter (like use a CMY one).
A CMY filter would be an interesting proposition - effectively each cell
becomes two colours, and hence has twice the light capturing capability.
This on it's own would give a 1 stop gain in effective ISO. I'm not sure
how it would affect colour accuracy though. I'd like to see someone
bring out a B&W only camera - without the colour filter a 2 stop gain
could be had, and noise is less obtrusive in B&W anyway, so I'd think
ISO's approaching 10,000 should be very achievable.
>
> David
>
>
 >> Stay informed about: Fuji Anti-Blur