• Welcome to the Internet Infidels Discussion Board.

The superiority of ternary computers?

One I ran into recently. Camera response to color is not the same as human eye response to color. When dealing with typical things illuminated by light distributed across the spectrum it works well enough. When dealing with light sources which are not proper spectrums it can fail very badly. Nebulae of excited gas emit at specific frequencies depending on what they are made of. Some of those are in the visual spectrum and you can end up with an image that looks very different to the naked eye vs the camera.
those kinds of images, which I myself have taken, are usually done with black/white cameras with color filters. The so-called “false color” is done to emphasize specific physical and/or chemical structures of interest and not to appear as they would to the eye.

Not sure what you mean by “proper spectrums” though.
 
Processors are made using CMOS. It is not really a voltage threshold detection. An input voltage has to be high enough to ensure a trasist9r turns fully on or low enough to ensure a transistor is fully off.

CMOS uses a minimum of transistors. Start adding more transistors to implement three state logic and costs go up. Chips start with a silicon wafer. Cost per chip is proportional to chip area. Increase chip size and buyer costs go up.



Interesting ACM paper. Note that silicon resistors on chips are huge compared to transistors and take up a lot of space.



Practical implementation is economic. Development costs, chip yields, new software.

A lot of digital signal processingg algorithms are based on base 2. Fast multiplication is done with a fast binary shift register. A lot of things work out nicely in base 2. Binary addition.

A lot o factors to consider. In short if there was economic benefit it would be done.

Dial up telephone line modes used multi value logic.

The high data rates of telephone modems were achieved by using multi-valued logic to transmit multiple bits per symbol over a limited-bandwidth analog telephone line
. Traditional digital logic has two values (binary), but multi-valued logic uses more than two discrete levels to represent data
 
As to why a ternary computer might be good, I've seen the argument that base 3 has more optimal use of memory space than base 2.

Here is what seems to be that argument. Consider base b, and consider expressing a number N in that base. Then the number of digits is

log(N)/log(b)

Larger b will mean fewer digits. But one must represent each digit in some way, let's say with b memory slots for each digit. Then the total memory becomes

log(N) * b/log(b)

The second term is independent of N, and an optimal value of b is thus optimal for every value of N.

One can find the best value of b/log(b) by taking d/db and then setting the result to zero. One finds log(b) = 1, giving us b = e ~ 2.7182818... the base of the natural logarithms. This means that base 3 is a little better than base 2.
  • 2: 2.88539
  • 3: 2.73072
  • 4: 2.88539
  • 10: 4.34294
 
One I ran into recently. Camera response to color is not the same as human eye response to color. When dealing with typical things illuminated by light distributed across the spectrum it works well enough. When dealing with light sources which are not proper spectrums it can fail very badly. Nebulae of excited gas emit at specific frequencies depending on what they are made of. Some of those are in the visual spectrum and you can end up with an image that looks very different to the naked eye vs the camera.
those kinds of images, which I myself have taken, are usually done with black/white cameras with color filters. The so-called “false color” is done to emphasize specific physical and/or chemical structures of interest and not to appear as they would to the eye.

Not sure what you mean by “proper spectrums” though.
Light sources not approximating a blackbody emitter. And I'm talking about stuff in space with emission lines at points the camera and the eye will see them differently.
 
One I ran into recently. Camera response to color is not the same as human eye response to color. When dealing with typical things illuminated by light distributed across the spectrum it works well enough. When dealing with light sources which are not proper spectrums it can fail very badly. Nebulae of excited gas emit at specific frequencies depending on what they are made of. Some of those are in the visual spectrum and you can end up with an image that looks very different to the naked eye vs the camera.
those kinds of images, which I myself have taken, are usually done with black/white cameras with color filters. The so-called “false color” is done to emphasize specific physical and/or chemical structures of interest and not to appear as they would to the eye.

Not sure what you mean by “proper spectrums” though.
Light sources not approximating a blackbody emitter. And I'm talking about stuff in space with emission lines at points the camera and the eye will see them differently.
I guess I was just curious as to what you think is “proper” about a blackbody and not for any other kind of spectrum.

I am quite familiar with many kinds of spectra, especially astronomical ones.
 
A digital video camera extends into the near infrared.

RGB and YUV color.



In an RGB led vary the currents in each led and you see a changing color spectrum.


A digital camera is an approximation to how the eye works.
 
Back
Top Bottom