Save 10% On Pictar Home Studio Pro Photography Kits With Code: EPHOTOZINE10

Need a new camera

Attention!

This topic is locked.

Reason : op questioned answered


Steppenwolf 8 1.2k
17 Jan 2014 11:23AM

Quote: My gut feeling is that is that when dealing with extremely sparse and sporadic photon collection it is better to have larger individual sites. You on the other hand seem to be saying that it is the increased circuitry of a larger number of sites which causes significant enough interference to cloud the lower level of signal present.


The reason that it's better to have fewer sensels in very low light is because of the "standing noise" of each sensel - if you don't like the term, substitute any you like. I don't know why you find this so hard to understand. Imagine, for example, that the sensels were perfect - i.e. generated no noise of their own at all. In this case it would make no difference whether you had 12Mp or 112Mp on the sensor in terms of noise. You could either look at the high resolution image and see all the shot noise, or you could down-res the image and end up with EXACTLY the same image as the 12Mp sensor. The fact that the down-res'ed image looks much noisier than the 12Mp image is because of the noise generated by the imperfect sensels. It's nothing to do with shot noise.


Quote:Good to see you are now accepting that shot noise cannot be got rid of anyway


I've always accepted that. I spent 4 years studying this stuff at Oxford - so I have a very basic grasp of the fundamentals. Quantum theory is mind-blowing but I'm not sure that anyone can claim to understand it. It's just one of those off-the-wall theories that accurately predict what happens - for the time being at least. As my tutor used to tell me (often), there's one thing that all theories have in common - they're all wrong.

Join ePHOTOzine for free and remove these adverts.

ChrisV Plus
13 2.3k 26 United Kingdom
17 Jan 2014 12:01PM

Quote:
-Good to see you are now accepting that shot noise cannot be got rid of anyway-

I've always accepted that. I spent 4 years studying this stuff at Oxford - so I have a very basic grasp of the fundamentals. Quantum theory is mind-blowing but I'm not sure that anyone can claim to understand it. It's just one of those off-the-wall theories that accurately predict what happens - for the time being at least. As my tutor used to tell me (often), there's one thing that all theories have in common - they're all wrong.



What's geography got to do with your grasp of the fundamentals? And were you ever charged for it?

I think if we went into an argument about falsifiability and the ultimate reliability of any scientific theory we would go several steps backward. There is general scientific consensus about the existence and nature of shot noise.


Quote:The reason that it's better to have fewer sensels in very low light is because of the "standing noise" of each sensel - if you don't like the term, substitute any you like. I don't know why you find this so hard to understand. Imagine, for example, that the sensels were perfect - i.e. generated no noise of their own at all. In this case it would make no difference whether you had 12Mp or 112Mp on the sensor in terms of noise. You could either look at the high resolution image and see all the shot noise, or you could down-res the image and end up with EXACTLY the same image as the 12Mp sensor. The fact that the down-res'ed image looks much noisier than the 12Mp image is because of the noise generated by the imperfect sensels. It's nothing to do with shot noise.


You will insist on using terms that only have meaning to yourself - how can we have a reasonable discussion on those grounds?

OK - whether this noise is 'standing noise' or 'goony bird' noise it doesn't matter - making the sensels larger for any given area makes them more efficient at gathering data from weak signals than a greater number of smaller pixels. If you had one giant sensel you'd be able to gauge the amount of light pretty accurately, but do bugger all else in terms of generating an image with complex detail. The whole point of having lots of sensels is that you are measuring light at specific locations. Now lets just say that you were able to make your raindrop gathering buckets square and perfectly abut each other. You're covering the same area so it doesn't matter how many buckets you have. Firstly all buckets have rims - anything which is discrete has a boundary. Once your sampling rate drops beyond a certain level, the data becomes totally unreliable for that individual site. Moreover the rim means it's never going to theoretically average flow collected by say 100 buckets as reliably as one giant bucket covering the same area. Flow will be scattered across and beyond boundaries - not to mention the fact that you of course lose data from specific location. The lower the flow, the more this matters. I would have thought that was self evident in a universe where [theoretically, admittedly] you can't have your cake and eat it.

It's not me that is telling Canon and Nikon to reduce the number of photosites in their high-sensitivity sensors. Honest.
Steppenwolf 8 1.2k
17 Jan 2014 1:59PM

Quote:It's not me that is telling Canon and Nikon to reduce the number of photosites in their high-sensitivity sensors. Honest.


They do it for the reasons that I've stated - several times. Firstly because the more sensels you have the more "standing noise" (i.e. noise due to imperfect technology) is generated and secondly (as I stated earlier) because you end up with more dead space on the sensor. When light levels get lower these effects becomes more significant.

It's absolutely stuff-all to do with shot noise. If the sensels were perfect (i.e. they basically just counted photons accurately and there were no gaps between them) there would be no need for Canikon to make low Mp sensors - for the very simple reason that the output from a higher Mp camera could be down-sized to achieve the same image as a lower Mp camera. So you could actually have the "free lunch" in that you could use the high resolution image if it was acceptable or you could have a low resolution image just as good as the low Mp camera produced.

Maybe some of your confusion is with the term "shot noise". It's not really "noise" in the standard meaning of the term. It's a faithful rendition of reality - not some distortion of it. It's just that reality has a quantum nature if you cut it up small enough.

Anyway, we'll have to agree to differ as this is getting tedious. There are much more interesting things to think about.
17 Jan 2014 2:01PM
Giving to distinguished members of scientific community all respect they may ask for I need to share my observation that in real life it better pays to listen to a practical specialist and this is what they have to say. Please notice how contribution from different noise sources changes under different conditions. Interesting reading overall for the people who wish to look behind the technical scene of photography.
The point is that there always can be designed a set of circumstances where anyone's wildest theory finds it's confirmation. The only thing, one may have no power over those circumstances and therefore needs to go for a compromise. That's where theory fails and practice takes over.
ChrisV Plus
13 2.3k 26 United Kingdom
17 Jan 2014 2:22PM

Quote:

Maybe some of your confusion is with the term "shot noise". It's not really "noise" in the standard meaning of the term. It's a faithful rendition of reality - not some distortion of it. It's just that reality has a quantum nature if you cut it up small enough.



It is confusing isn't it? But the result in terms of the image produced is noise. I think if you read closely there may be a clue to that in the term.

However I'll agree to disagree because clearly no matter how I word it you won't accept that decreasing sensor sites results in recording a better low threshold signal. Even though you appear to accept that manufacturers decrease sensor sites to achieve a better low threshold signal.

There are some things I will clearly never understand.
lemmy 13 2.8k United Kingdom
17 Jan 2014 2:26PM
Does anyone have any idea how many angels could dance on a quantum senselated sensor and would this affect the amount of money I get paid for my pictures? Tongue
ChrisV Plus
13 2.3k 26 United Kingdom
17 Jan 2014 3:30PM
It's 274,983.7.

Angels can be subdivided into 10ths. But God gets very angry if they pose for commercial purposes. Any fule no that.
lemmy 13 2.8k United Kingdom
17 Jan 2014 3:59PM

Quote:Angels can be subdivided into 10ths.


I meant arc-angels. An arc-angel is made up of smooth curves and dividing one into tenths would inevitably involve quantisation errors. An angel with errors would be a demon. Demons, unlike angels, do not require a model release for commercial use. I am surprised that, as a professional photographer, you did not know that.

However, I do agree with you that decreasing sensor sites results in recording a better low threshold signal and I certainly do not accept that manufacturers decrease sensor sites to achieve a better low threshold signal. Nicéphore Niépce said as much himself in one of his blogs of 1827 "Si les sensels étaient parfaits (c'est à dire qu'ils fondamentalement juste comptés photons avec précision et il y avait pas d'espace entre eux), il n'y aurait pas besoin de Camera Obscura faire capteurs bas Mp".

Wise words uttered some 187 years ago and as true today as they were then.
ChrisV Plus
13 2.3k 26 United Kingdom
17 Jan 2014 4:32PM
If you're going to alter the goal posts midway through the fishing expedition, it's no wonder people are going to be thrown off the scent and thus find it a hard nut to crack.

I think in this particular innings you may well have scored an own-goal.
peterjones 18 5.0k 1 United Kingdom
17 Jan 2014 6:54PM
What a thread zzzzzzzzzzzzzzzzzzz, no wonder the OP has vanished; go ye to the venerable yet redoubtable Ffordes; they have s/h D7000s from £449 and s/h 18/105 lenses from £129, both well inside your budget, you won't be disappointed with either the camera, lens or Ffordes service.

G' luck, Peter.
Steppenwolf 8 1.2k
18 Jan 2014 8:58AM

Quote:
However, I do agree with you that decreasing sensor sites results in recording a better low threshold signal and I certainly do not accept that manufacturers decrease sensor sites to achieve a better low threshold signal. Nicéphore Niépce said as much himself in one of his blogs of 1827 "Si les sensels étaient parfaits (c'est à dire qu'ils fondamentalement juste comptés photons avec précision et il y avait pas d'espace entre eux), il n'y aurait pas besoin de Camera Obscura faire capteurs bas Mp".

Wise words uttered some 187 years ago and as true today as they were then.



If Niépce ever had said that, Lemmy, he would have been dead right.

There's no disagreement that decreasing the number of sensor sites results in lower noise. The question is whether this is because of "shot noise" or not. Good to see that Niépce agrees with me.
KevSB 16 1.5k 5 United Kingdom
18 Jan 2014 9:04AM

Quote:My nikon D90 broke (just befor Christmas) amazon have refunded me so now I am in need of a new DSLR and would love some advice/recommendations/names of cameras to look at.
I'm not very word savvy when it comes to cameras so if I could ask that if you need to use fancy camera related words or sentences can you explain what you mean? I like recommended cameras rather than just being told to look at what the camera can do - mainly because I do not have a clue.

I am off to Florida in June so would like a camera that would give me amazing photos out there especially at night time and for fast action shots and firework displays.



The original post. the poor guy just wanted a recommendation in simple terms.
keithh 16 25.6k 33 Wallis And Futuna
18 Jan 2014 9:21AM
.....yeah but we all know that photography is rocket science.