6th April 2005, 03:33 PM
WAH so many 'Beautiful Minds' here
6th April 2005, 05:54 PM
Originally Posted by Watcher
Don't tell me what I already know.
So from your example, I can capture 10 stops even with 6 bits, or 4 bits, or better still, 1 bit?
0.001 to 0.511 --> bit 0
0.512 to 1.024 --> bit 1
Yes, that will capture 10 stops according to your arguement. But, you can't differenate that a "0" signal is either at 0.001 lux or 0.511 lux. I know that is quantisation error. But that's not the point...
The original question was "Does one bit depth represent one stop difference". And now we're talking about DR & errors. And you still want to sidetrack and talk about sound. Let's get back to the main question shall we?
From post #28 I said, "U can see from there that each bit accounts for every stop difference. Therefore, there is theoratically 11 stops."
Then you replied in post #30, "I can have 2 stops of difference in the amount of exposure latitude and still have all 12 bits or 16 bits used up. They measure different things.".
Let me get this straight before we start to miscommunicate again...
"one stop of light" --> means that the final amount is twice or half the original measured value, with reference to the original value.
"Exposure Value (EV)" --> refers to a number, with reference to a fixed value (EV0 = 1s at f/1.0).
The world can be said using EV to measure light (at least it's true when we use cameras). This is different from using "stops". What I was doing here is using stops to measure the range of the sensor, not EV.
Again, I'm talking about stops, each bit accounts for one stop difference. What you did was to use EV to measure the capability of the sensor. Is that where we miscommunicated?
6th April 2005, 06:42 PM
If my circuits and processors are very low in noiseless, I can have a 7 stop sensor, with 14-16 bit ADC (IIRC D2X has 14 bits). Now, in that case, I can have > 1 bit for 1 stop. From my previous example, which you concurr, I too can have <1 bit for 1 stop. Therefore the original question "Does 1 bit depth represent..." is no. There is no mandatory relationship (or correlationship if you must) between bit depth and stops. Soundwise is not a sidetrack; it is to draw identical models for clarification. We can use 8 bits for concerts that are loud and soft, we can use 16 bits for speech with very little variation on volume. Loudness == magnitude of signal == magnitude of light.
Originally Posted by AReality
When light falls onto the film or sensor, the amount is controlled by the aperture + shutter. The amount (aperture + shutter) = eV, but nowhere is a ratio of light on the film, it is the absolute amount.
Originally Posted by AReality
Now assuming that we have a subject that has exactly 6.8 stops of range. We meter the system A so that the amount of light at a fixed ISO falls into a 7 stop sensor without any loss. We can see that there isn't any loss as the darkest element is above RGB(0,0,0) and the brightest isn't at RGB(255,255,255). Now, I have another system B that has the same ISO but have 11 stops of range and we take this same subject and meters it so the darkest element has the same RGB value (say RGB(dr,dg,db)). This is to ensure a floor reference point.
Now into the scene, we put two objects that are 8 stops and 10 stops brighter than the darkest object. This increases its range of brightness to darkness, and now the DR is 10 stops. What happens? The system A with the same exposure and sensitivity will get RGB (255,255,255) on the new items, right? What happens on system B? The two new items will definately have < RGB(255,255,255).
How does this translate? The amount of light falling on a sensor will trigger the electrical current; never the ratio of light. This electrical current is then eventually, via ADC which translate the analog signal into digital bits. Thus, the sensor is normalized and calibrated to respond btween x1 eV to x2 eV (at a particular gain), this then derives the DR, measured in ratio; not the other way round. The sensor converts the signal on ABSOLUTE measurements of light at a particular gain/ amplification /ISO sensitivity to an electrical, NEVER on the ratio between the brightest and darkest objects.
The dog wags the tail, not the tail wags the dog.
I can have the sensors of systems A & B (assuming they have the same Vmax) connected to the same ADC. I still have the ratio between the brightest and darkest objects recorded for 7 for System A and 11 for system B. It is the responsibility of the AE system to manage the amount of light falling on the sensor or film.
Last edited by Watcher; 6th April 2005 at 06:49 PM.
6th April 2005, 06:58 PM
From your examples in posts #28, #34 and #38, you assumed that because it has 12 bits ADC, you take out 3 bits of noise, therefore DR = 9 stops (#28: "Means 9 stops can be recorded"). What if you change the ADC for a 10D to say 14 bits? Does it mean that, if the noise is the same, you can now have 11 stops recorded?
I had asked you back in post #40: "what can be read in a 16-bit A-D conversion like in a medium format digital back?" Should it be 13 stops?
In post #42, you said that "The 8 bits are actual truncated values of 16 bits. This is because there is no need for so many bits due to noise issues;" Now you said "I know that is quantisation error". So which is it? Quantization error or noise? This is assuming that the sensor and circuit has the same amount of noise.
QED. There is no correlation between 1 bit = 1 stop of light when it comes to DR.
Originally Posted by "AReality, post #42
Last edited by Watcher; 6th April 2005 at 07:08 PM.
6th April 2005, 07:00 PM
6th April 2005, 07:19 PM
I have sent a slide of a sunset shot to print in fuji lab,after the scan and print,lots of detail is gone,but when I print it myself in the darkroom using Ilford P-30 kit,the difference is very great,the tonal range print by fujilab has left out lots of detail that I do in myself in the darkroom.
6th April 2005, 07:33 PM
Originally Posted by jOhO
I think this is not goin' anywhere... It has been going round in circles... I'm sure we've got other better things to do... Let's just leave it as that. Let the individual readers decide (if they ever understand everything)... As the saying goes: "Arguing on the internet is like running in the Special Olympics; Even if you win, you're still retarded." The saying applies to all... I think I'll just stop here. Cheers everyone...
6th April 2005, 08:14 PM
Heh.. The question begs.. What do they use to scan? Almost any slide/ film scanner available has lower resolving power than film/ slides have to offer. Some of the exceptions are the drum scanners like Aztek Premier.
Originally Posted by sabahan
6th April 2005, 08:19 PM
Here's a page dealing with the maths behind calculating CCD gain.
Basics of Bit Depth and image sampling.
Analysis of CMOS sensors for astronomy (If the first 2 articles didnt put you to sleep)
6th April 2005, 09:23 PM
i'm not as familar with light as with sound. i'll use sound as an example
assuming the microphone is perfect linear to loudness of the sound:
u can use your 8 bit to record 0dba to 24dba, with each bit recording a difference of 3dba. that's 24db of range
i would like to use my 8 bit to record from 24dba to 120dba, each bit recording a difference of 9dba. that's 96db of range
of course the mic is not that linear....
6th April 2005, 09:54 PM
7th April 2005, 03:02 AM
Originally Posted by mattlock
You might have misread/misinterpreted my words, or maybe it's a case of me not writing it properly. (In this reply, I'm not using forceful words like "must be" and "will be" but softer words such as "maybe", "could", "seem".)
Anyway, when doing research for academia or any purpose for that matter, it is important to also note the professional qualification of the person/organisation who discovered the facts or proved the theory. The web has full of information obtained from personal experience which could be misleading and less qualified. (No offence and I don't imply that you could be ill-qualified.) Hence, just as how academic research is done, I PREFER to accept information or facts discovered NOT by individuals from personal experience.
Despite your personal experience, I have already quoted in my earlier post that Kodak estimates ISO100 film has at least 22 megapixels of digital camera resolution. And since I don't prefer to accept personal experience (again no offence to you), Popular Photography magazine with whatever lab equipment they used, determined ISO100 film is about 40 megapixels equivalent. For whatever reason they tested film, AFIP (I believe it means Armed Forces Institute of Pathology) published in its Peer Review Science Paper that ISO100 has the equivalency of 54 megapixels. The same paper also said ISO25 and ISO400 film has the equivalency of 124 megapixels and 22 megapixels respectively. Through their tests, it appears that lower ISO films have more resolution.
You said/implied that if film was scanned up to 40 megapixels, most of it will be noise. Any scan at any resolution will surely have noise. The noise comes from the scanner, NOT from the original image on the neg/trans. I'm don't know how those labs tested the films but I'm sure they knew scanning will degrade the film image resulting in an unfair comparison. They probably did their test in some way without scanning the film.
You could be generalising. ISO400 is not all disgusting. Maybe you printed from a Kodak 400? Fuji films appear to be finer in grain. Or, could the grainy appearance be the result of slight underexposure? (You don't have to reply to this since it would be really difficult to determine anything online. My point is that you could be generalising.) Of course, I know, and agree with you, that faster ISO film will be grainier.
I DID NOT say "3 stops of dynamic range for slide film and 7stops for negative film". I believe I said/implied "exposure latitude". I may be wrong but I doubt both terms refer to the same thing. I got my "exposure latitude" information from published books, 2 of which are textbooks used by RMIT University's photography students published by the reputable Focal Press.
I believe the amazing numbers "13" and "11" you mentioned about motion picture films because I found close-enough numbers on the Kodak website on their motion picture films. However, if I recall correctly, the term used was "exposure latittude", not "dynamic range" which you have been using.
Yes, motion picture film is different, but not totally different. Development is different (ECN-2 vs C-41), but it is made using the same principles as still camera film. And motion picture films are also available in 35mm format. Ok, maybe it's not a good comparison with still camera film on my part.
You said, "Another thing, when you project an image on a screen it is not equivalent to printing a picture that big." I agree, but I see it as the screen being similar to the print. Ok, maybe it's not a good analogy on my part.
You said, "And negative film can only take about a stop of underexposure. even with the extra latitude, you'll just get a very grainy image." Yes, I know, and most people know too, and that's why we all try to shoot "correctly" exposed images on film. In the digital, shadow areas are similarly "grained" by electronic noise.
You also said, "Negative film does not necessarily print with much more dynamic range on PAPER than slide film." Yes, true, but since I'm no expert, and I haven't found proof, I just assume that over the decades, the paper may have improved to match negative films. (My assumption could be false.)
With regards to digital backs and medium format, I won't comment since I don't use them. I may have said/implied that my post was with regards to 35mm format.
You said, "Sharpening the file just made the grain more apparent." I assume you meant the film image file. Yes, that's expected. Grains will show. However, if you oversharpen your digital camera image, you also get something right? Pixels don't show like grains do but you get oversharpening halos.
You said, "Digital cameras do NOT reduce a lens' optical quality, more likely they show the lens' defects more clearly because you can zoom in and take a look. you don't notice it so much with film because film tends to look softer than digital (thanks in part to the grain in film)." Hmm... shows defects? Never mind. Ever heard of anti-aliasing filters? All digital cameras (maybe except lousiest webcams) have them. You should find out what these filters do, and how they so badly reduce lens resolution!
These filters are the reason why out-of-camera digital images usually look soft, and often benefit from some software sharpening, whether it's in-camera sharpening or image editing program sharpening. (You got to admit that almost everybody who shoots digital does some software sharpening before it's printed but the film image seldom requires sharpening.) Since you say film looks soft, then it may be because your digital image has been sharpened (as compared to the film one).
My apologies for the near heart attack but I suspect that you misread or I wrote badly.
7th April 2005, 07:56 AM
I am fascinated! And I think many would like to know more.
But I suspect that many reading this post are as confused as me! (Seriously! Not meant as a joke!)
Is it possible for some knowledgeable blokes to summarise all these for those of us who are less technically minded?
7th April 2005, 02:51 PM
7th April 2005, 03:37 PM
thumbs up for Jemapela
I respect you Jemapela. Thumbs up for you.. good, good fighter.
7th April 2005, 05:27 PM
Ok guys listen:
1. Put on your cloths
2. Get out away from your computer
3. Go get some drinks or a puff of the ciga
Comeon, just shoot with what you like. No matter who is right or wrong, Canon, Nikon, Sony, Kodak, Matsushita etc etc are not going to see this and build us a better CCD/CMOS sensor. The sun is good today.
Off for my night class at SIM
7th April 2005, 05:35 PM
From norman koren:
"With the new Imatest program, digital camera tonal response and dynamic range to be measured easily and accurately using a transmission step wedge-- a piece of film with zones of incresing density. The details of the measurement are in the Imatest Q-13 tour.
Here are the results for the Canon EOS-10D at ISO 400, converted from RAW format with Capture One LE. The upper left plot is the density response curve. The lower left plot is noise measured in f-stops-- a relative measure that corrresponds to the eye's response. Each step represents 1/2 f-stop (a density step of 0.15)."
Since we can't hotlink pics here, here's the conclusion:
The total dynamic range of the EOS-10D is 8.5 f-stops
8th April 2005, 07:41 PM
In all likelyhood, this is not a technical "digital vs. analogue" issue. Digital image processing is vastly more powerful than wet lab methods, but that convenience makes it also prone to abuse. Photofinishers are known to "optimize" prints (read: make them colourful and boost contrast). Also, thanks to modern technology, nowadays a monkey can operate a minilab, and the lowered qualifications of the operators may be reflected in the end result.
Originally Posted by sabahan
9th April 2005, 09:10 PM
Originally Posted by theITguy
While everybody needs to read from somewhere and learn, you're also quite correct.
Staying away from the computer to shoot more images is actually more beneficial.
The late martial arts expert Bruce Lee once said (I can't remember exactly but something like), "Knowing is not enough, we must do. Learning is not enough, we must apply."
9th April 2005, 09:34 PM
Read but learn right
While everyone needs to read and learn, it is important to read and learn from more credible and qualified sources. This is practised in all good universities and institutions of higher learning worldwide.
Originally Posted by don juan
Personal experiences from Toms, Dicks, and Harrys (I'm not refering to anyone in particular) splattered all over the Internet have little value. Anyone, Michael Jackson or Lee Kuan Yew, can post his personal experience/feelings of his new EOS-1D Mk3 on the Internet.
For similar reasons, you wouldn't buy an instructional photography book authored by the comical Iraqi Information Minister or Mr Bean (if they actually ever get published).
Hence, in my post, I quoted information coming from more credible and more trustworthy sources such as Kodak, Popular Photography, and AFIP.
As human beings, we mix facts, personal experiences and feelings into a forum. Filter out the facts and store them safely in high quality brain cells. Retain the others in lower grade memory space.