Shannon capacity - for experts only!
Discussion
!! ...and it has a special footnote at the end for the really geeky !simpo two said:
Definitely one for Bacardi and Ed I think!
<a href="http://www.imatest.com/docs/shannon.html">www.imatest.com/docs/shannon.html</a>
I say: 'Wibble'!
I've just run that webpage through the BabelFish translator and it said :-
"What the fu........."
Can someone explain it in a single sentance?
>> Edited by FourWheelDrift on Thursday 9th December 19:47
Ill try to explain it from my understanding:
Hypothetically, the Shannon capacity can be used as a measure of digital camera quality calculated by an equation involving sharpness and noise values recording at different contrast levels.
The different contrast levels are used because:
noise is more of a worry in areas of one colour (low contrast) a plain backgroud for example, and sharpness is more of a concern in areas of detail such as the edges of a subject on a background (high contrast) so it is necessary to get an inclusive set of values.
From what I can tell, this is a scientific or quantitative method of proving that the better a camera is, the sharper and less noisy the images will be. I believe they are trying to provide a scale for quality that is more accurate than just how many megapixels a camera is.
Cheers.
Matt
>> Edited by gravymaster on Thursday 9th December 20:17
Hypothetically, the Shannon capacity can be used as a measure of digital camera quality calculated by an equation involving sharpness and noise values recording at different contrast levels.
The different contrast levels are used because:
noise is more of a worry in areas of one colour (low contrast) a plain backgroud for example, and sharpness is more of a concern in areas of detail such as the edges of a subject on a background (high contrast) so it is necessary to get an inclusive set of values.
From what I can tell, this is a scientific or quantitative method of proving that the better a camera is, the sharper and less noisy the images will be. I believe they are trying to provide a scale for quality that is more accurate than just how many megapixels a camera is.
Cheers.
Matt
>> Edited by gravymaster on Thursday 9th December 20:17
HarryW said:
As it said, old hat, been used in communications for years .
Yes, but it's the first time I've seen the theory in public in image processing (but in truth I've not been looking). Been around under cover (if you know what I mean) for years...
Fatally flawed in a number of ways and easy to get false good/bad numbers out of it in standard photographic testing.
If anyone fuels me with enough beer down the pub I'll rant about why it doesn't really work for...well minutes probably.
The simple answer is images are perception. Want to know if the camera system's good? Take a picture and LOOK at the output.
Strange thing about the article is the 'very geeky' bit is the simplest.
Of no interest to me thanks.
Well said that man......
.......and again
Looking at charts and graphs to determine picture quality is like looking at charts in Hi Fi mag's and understanding sound quality. Forget formulas and graphs and use the best tools for the job...... your eyeballs.
V6GTO said:
BORING!
Well said that man......
ThatPhilBrettGuy said:
Take a picture and LOOK at the output.
.......and again
Looking at charts and graphs to determine picture quality is like looking at charts in Hi Fi mag's and understanding sound quality. Forget formulas and graphs and use the best tools for the job...... your eyeballs.
simpo two said:
And frankly, I'd put histograms in the same bin.
Wouldn't go that far. I rely on histograms far more than some stupid little preview on the back of the camera. I'm just processing to-days shoot and one eye is on the histogram and one on the picture. A useful tool when you know how to use it.
Bacardi said:
Wouldn't go that far. I rely on histograms far more than some stupid little preview on the back of the camera. I'm just processing to-days shoot and one eye is on the histogram and one on the picture.
Heh, a bite! I'll stick with 'use the best tools for the job...... your eyeballs.'
Shannon Capacity is fatally floored as a method of measuring a camaras results. It measures two things and then gives you one number(if I understand correctly) an average, or sum, of the two. It shoot be more like a football score, ie:- 4-3 or similar. A camera could be very good at noise reduction but terrible at sharpe edges and end up with a reasonable score, but the camara would be good for nothing. Or am I missing something?
Martin.
Martin.
simpo two said:
Heh, a bite! I'll stick with 'use the best tools for the job...... your eyeballs.'
Ultimately yes. Depends on the final destination of your picture. If you take a digital pic for a magazine which goes via a computer monitor, are you sure your monitor's set up correctly? Are you sure you're not clipping highlights you might want to retain detail in? Can you really judge the quality of the exposure on the rear screen of your D70 in very bright sunshine? I wouldn't, except to check composition and to see if I caught an expression I wanted. I'd check exposure on the histogram.
Evaluating digital camera quality by not getting the exposure spot on and then resorting to auto levels in PS defeats the object of the exercise. So, yes, your eyeballs are the 'final' judge but the histogram is a useful tool during the process, for reassurance..... eyeball backup.
Gassing Station | Photography & Video | Top of Page | What's New | My Stuff