High resolution...through digital interpolation...

Confused writes:

When I saw the first flat panel orange plasma monitor:
"THIS is the future", I remember thinking.

You were right ... but it sure took a long time to get there!

(I don't remember the year... mid to late 70's ??)

CDC began marketing PLATO commercially in the 1970s, but I believe the
project dates from about ten years earlier. And military flat panels
predate that.

Computer technology has always lagged far behind ideas. In most cases,
software lags behind hardware, too, but in some cases (PLATO, Multics,
etc.) hardware lagged behind software.

Virtually none of today's computer software "wonders" is really new. It
was all thought of decades ago; there just wasn't any hardware on which
to run it.

--
Transpose hotmail and mxsmanic in my e-mail address to reach me directly.

On Tue, 12 Apr 2005 05:46:50 GMT, Confused
wrote:

On Thu, 07 Apr 2005 22:36:50 GMT
In message
wrote:

In message ,
Ron Hunter wrote:

I will even give you a name to check out. There should be ample record
of his writings in the area back about 20 years ago. John Navas.
He, and many others explained in great detail why it was physically
impossible to EVER send more than 450 bps over a telephone line. OF
course those who didn't believe in limitations just went right ahead and
DID it. Now John runs a website that tells you how to maximize the
speed. And life goes on.

He was a usenet regular in some of the Windows-related newsgroups a few
years back when I used to participate in them.

Ah, for the good old days with a terminal and a 300 BAUD modem at
home, when every single byte was worth it's weight in time... ;^)

Jeff

You misspelt "wait". :-)

--
Bill Funk
Change "g" to "a"

On Tue, 12 Apr 2005 05:46:50 GMT, Confused
wrote:

On Thu, 07 Apr 2005 22:36:50 GMT
In message
wrote:

In message ,
Ron Hunter wrote:

I will even give you a name to check out. There should be ample record
of his writings in the area back about 20 years ago. John Navas.
He, and many others explained in great detail why it was physically
impossible to EVER send more than 450 bps over a telephone line. OF
course those who didn't believe in limitations just went right ahead and
DID it. Now John runs a website that tells you how to maximize the
speed. And life goes on.

He was a usenet regular in some of the Windows-related newsgroups a few
years back when I used to participate in them.

Ah, for the good old days with a terminal and a 300 BAUD modem at
home, when every single byte was worth it's weight in time... ;^)

Jeff

You misspelt "wait". :-)

--
Bill Funk
Change "g" to "a"

In message ,
Confused wrote:

Ah, for the good old days with a terminal and a 300 BAUD modem at
home, when every single byte was worth it's weight in time... ;^)

I remember the first time I saw a modem in action; I went to a
co-worker's home on out lunch break, and he showed me his computer. He
had a compuserve account, and connected to get his email. The text
appeared on the screen one character at a time, as if someone were
typing it, very fast.
--


John P Sheehy

Bryan Olson writes:
Bart van der Wolf wrote:
Well, if you analyze image content (e.g. Fourier analysis) you can't
help to conclude that a large part of all image structures has a
predominantly horizontal/vertical orientation.

If you can actually cite that experiment, I'd like to see it.

It's pretty simple to do yourself. Take a digital image. (Optional:
convert to monochrome so you only have one channgel to process).
Perform Fourier transform on it using a variant that puts the "DC term"
in the centre. Draw 4 lines at +-22.5 degrees and +-67.5 degrees
through the centre, dividing the output into 8 octants. Note that 4 of
these octants are all the points closer to a horizontal/vertical line,
while the other 4 are all the points closer to a 45 degree diagonal
line.

Then draw a circle that just touches the sides of the image across its
narrower dimension. Discard all points outside the circle. This makes
each octant have the same number of points, representing the same range
of spatial frequencies.

Now calculate the average coefficient value (use the magnitude of the
complex number) in the 4 horizontal/vertical octants, and separately
calculate the average coefficient value in the 4 near-diagonal octants.
Ignore the very centre; it doesn't belong to any octant (and is probably
far larger than all the other coefficients).

Are the two averages about the same? Or is the average for the
horizontal/vertical octants larger than the average for the diagonal
octants?

This tells you the answer for one image. Repeat for many "typical"
images to get some sort of average.

On the other hand, lines that are horizontal in the world often
are not parallel to the horizon in photographs, due to
perspective.

True, but if they are within +- 22.5 degrees of horizontal or vertical,
then (in theory at least) the rotated Fuji sensor has higher resolution
than a row/column grid sensor of the same size and pixel count.

Dave

Bryan Olson writes:
Bart van der Wolf wrote:
Well, if you analyze image content (e.g. Fourier analysis) you can't
help to conclude that a large part of all image structures has a
predominantly horizontal/vertical orientation.

If you can actually cite that experiment, I'd like to see it.

It's pretty simple to do yourself. Take a digital image. (Optional:
convert to monochrome so you only have one channgel to process).
Perform Fourier transform on it using a variant that puts the "DC term"
in the centre. Draw 4 lines at +-22.5 degrees and +-67.5 degrees
through the centre, dividing the output into 8 octants. Note that 4 of
these octants are all the points closer to a horizontal/vertical line,
while the other 4 are all the points closer to a 45 degree diagonal
line.

Then draw a circle that just touches the sides of the image across its
narrower dimension. Discard all points outside the circle. This makes
each octant have the same number of points, representing the same range
of spatial frequencies.

Now calculate the average coefficient value (use the magnitude of the
complex number) in the 4 horizontal/vertical octants, and separately
calculate the average coefficient value in the 4 near-diagonal octants.
Ignore the very centre; it doesn't belong to any octant (and is probably
far larger than all the other coefficients).

Are the two averages about the same? Or is the average for the
horizontal/vertical octants larger than the average for the diagonal
octants?

This tells you the answer for one image. Repeat for many "typical"
images to get some sort of average.

On the other hand, lines that are horizontal in the world often
are not parallel to the horizon in photographs, due to
perspective.

True, but if they are within +- 22.5 degrees of horizontal or vertical,
then (in theory at least) the rotated Fuji sensor has higher resolution
than a row/column grid sensor of the same size and pixel count.

Dave

Confused wrote:
And flat-panel displays predate this by 20 years. Microwave ovens
predate it by 40 years at least, as does digital voice transmission.


When I saw the first flat panel orange plasma monitor:
"THIS is the future", I remember thinking.

(I don't remember the year... mid to late 70's ??)

Jeff

And I remember talking to TI about their digital mini light mirror array
in the 1980s