Simple terms explanation please

[steelfixer]

Hi all.

I need something explained but in simple terms as I am in no way photographicaly minded.

In my simple understanding of things the ISO number related to the amount of light allowed into the camera.

Now, as there are no lenses used for DSO work so there for no means for the cemera to reduce the amount of light coming in does the ISO setting matter.

I am using a canon 350 D.

Please remember to keep it simple.

Thanks

Graham

[IanC11]

If I'm correct the ISO rating isn't how much light is allowed through the lens ( that's the " f " number ) the ISO number was how light sensative the camera " FILM " was, now on a digital camera I assume its how the ccd chip reacts to the light reaching it.

I'm no expert so I stand to be corrected.

Ian

[swag72]

With regards to ISO, DSLR's and AP, the general thought is that you get your ISO as high as you can without detriment to your images. You want to catch as much light as possible in AP as this will mean that the more light you get on your sensor, the more detail you will have in your image.

I used 800 ISO on my 1000D, if I went higher, it would introduce a lot of noise that I couldn't deal with, and so it's a bit of a compromise really.

Have a play with your camera, take an image at 100 ISO, 400 ISO and 800 ISO and you will be able to see the effect of additional light hitting the sensor as well as the additional noise that creaps in at higher ISO's. You can then decide which is best for you.

You may find for your 350D that 400 ISO is best, as I understand that the older camera's are not good at the higher ISO's.

Hope that helps.

[Cornelius Varley]

If you are shooting through a telescope (which now becomes the cameras' lens) you have a fixed aperture (f number). To control the amount of light reaching the sensor you can adjust the shutter speed or alter the ISO. Changing the shutter speed is a better option than altering the ISO.

Peter

Peter

[steelfixer]

Many thanks for the explanation.

Peter you say it is better to change the shutter speed which is good as I have a problem with my camera.

I cannot change the ISO it is stuck on 1600.

I will be using it through 'back yard' so I can control the exposure using the bulb mode.

Any thoughts on time of exposure or is it just a matter of trial and many errors.

[TonyM]

The ISO setting changes the sensitivity of the CCD chip.

There is a trade-off between ISO and picture quality in that the lower ISO numbers need more light to acheive correct exposure but result in less noise in the image. A higher ISO needs less light to achieve correct exposure but results in more noise in the image.

With a lower ISO setting and fixed aperture you will need a slower shutter speed. With a higher ISO setting and fixed aperture you will be able to use higher shutter speeds to get a correctly exposed image.

[swag72]

Shutter speed will be determined by a number of factors, including light pollution and whether you are guiding.

With a light pollution filter in place and guiding, I was regularly using 8 min subs on my DSLR

[steelfixer]

Thanks Sara.

Light pollution is not too bad and I will be guiding.

[swag72]

It really is trial and error for your particular location. You don't want to fully saturate stars for example, other wise you will lose all star colour and likewise galaxy core's too. It may take a night to really work out your optimum exposure length, but once you've got there, you'll probably not deviate from that too much, so it's worth spending time on in my opinion.

[steelfixer]

Thanks Sara.

All I need now is for it to stop raining and we will give it a go.

I did have some clear skies here last night for the first time in weeks.

At 6 oclock some ****** tried to steal some cables and blacked out my part of the town, left me with no power until this morning.

I just cannot win. lol

[themos]

Simple terms, then:

1. Light comes in packets of energy called photons. All green photons have the same energy (red photons have slightly less energy and blue photons have slightly more but let's think of green ones for now).

2. The atmosphere will send green photons to your camera (that's light pollution and skyglow) and so will the DSO you are pointing at. They all look the same.

3. Every time a green photon reaches your 350D sensor, there's a 26.9% chance that an electron will be kicked out of an atom and become a free charge.

4. If, after a certain time of exposure, 1000 photos arrive at one of your sensor's pixels, then you'll have 269 electrons in that pixel.

5. At the end of the exposure, the sensor electronics will read that charge and produce a voltage proportional to that charge.

6. The reading process introduces an error, called the "read noise". That's typically a few electrons' worth: sometimes the 269 will be counted as 265, sometimes as 275.

7. The voltage will be digitised (in 12 bits), that is, it will be converted to an integer number between 0 and 4095. This is where the ISO setting comes in.

8. If no charge was found in a pixel, you'll still get a non-zero number out of the electronic process and that is called the "bias", probably something like 128. Not all pixels would produce 128, some a bit higher and some a bit lower as the read noise is still there.

9. When you double the ISO, the electronics assign a twice as large number to the pixel (after removing the bias). So, let's say that at ISO 400 the 269 electrons are read as 269 (we got lucky) and then the bias (128) is added to give a number = 269+128 = 397. At double the ISO, ISO 800, we'll see the number = 269*2+128 = 666 (if we get lucky with the read noise again).

10. Remember, we can't make numbers greater than 4095, so the brightest pixels will all show 4095 and we'll have lost information about their relative brightness. That's another way of saying that doubling the ISO reduces your dynamic range. It's as if you're "zooming in" to a small range of brigtnesses.

11. The read noise tends to be smaller at higher ISO settings.

12. Ta cap it all off, due to the quantum nature of light, a region in the sky that delivers, on average, 1000 photons to the pixel over a certain time, will not produce them at a steady rate: the next time you expose for the same amount of time you might get 1025 or 975. The range of values you will see is plus or minus the square root of the average count. Square root of 1000 is about 32 so that gives you a sense of the variability.

13. If you get 25 photons, the variation is 5 and their ratio is 20%. If you get 15625 photons the variation is 125 and the ratio is 0.8%. Well lit regions will show less noise. That is why we take such long exposures.

14. In astrophotography, we expose long enough so that the variation from photon variability is much bigger than the effect of the read noise. The first is purely sky (and optics) dependent, the second purely sensor dependent.

15. This is as simple as it can get, sorry :-)

PS: please point out any mistakes, thanks.

[robbieince]

There may be some that think ISO allows more light on the sensor - that isn't true. That is fixed by the aperture and the focal length of the scope/lens on front of the sensor.

Think of ISO as the volume control on your stereo. It amplifies more when turned up. It amplifies the background noise too though. So, at higher settings, what is normally quiet (faint) is easier to hear (see) but the background noise is more obvious too.

ISO can't make something visible that isn't already there but it can be used to make exposures shorter so other errors like bad drive tracking aren't a cause of problems.

Regards

Rob

[Shibby]

So, if the A/D convertor was 13bit, would this offer all the advantages of doubling the ISO but without losing dynamic range?

Sorry if that's going a bit off topic steelfixer, I think you already have your answer.

[steelfixer]

Themos - Many thanks for the detailed explanation it gets my vote to become a sticky.

For a complete camera novice like me that information is pure gold.

It must have taken you ages to type that lot out.

Many many thanks for taking the trouble.

Rob - Thanks for the advice I will put that into pratice when ever this rain decides to stop.

Lewis - No problem I find this stuff facenating

[themos]

Shibby, the "width" of the Analog-to-Digital Converter is dictated by the physical property of how much charge can be held in a pixel, what is known as the "full well capacity". That is given in numbers of electrons. That is to be compared with the read noise, also in numbers of electrons. The ratio is usually called the dynamic range.

Think of lowest ISO setting first and use the numbers from the table in the link below: it takes 8 electrons at ISO 100 to move from one number to the next in the value recorded for each pixel and you can't store more than 32680 electrons in the pixel while the read noise is 10 electrons. That gives a dynamic range of 32680/10 = 3268. That is the number of distinct states it makes sense to record. As the number of states doubles with each added bit, it is possible to encode these states in 12 bits (numbers of 0 - 4095). We even have room for a small bias voltage corresponding to a value 128 or so. There would be no advantage is making the ADC more precise.

Now look at the highest ISO setting (1600) : each electron now will produce a voltage 16 times as big so to move from one number to the next we only need "half an electron". That's a bit of a waste as there never is "half an electron" but let that go for now. The read noise is less, 7.4 electrons and the pixel can still hold 32680 electrons maximum. The dynamic range would be the ratio: 4416, which is better than the low ISO case. BUT now we are limited by the 12-bit ADC which will be reading and converting voltages 16 times higher. Many more pixels will have voltages that saturate the 12 bit output and you are then limited by the width of the ADC. That's how you get the effective full well capacity of 2100 electrons at ISO1600. And the resulting effective dynamic range of 2100/7.4 = 284, which is worse that the low ISO case. So you can say that DSLR systems were designed to optimise well-lit scenarios and their ADC cannot cope with high ISO settings. I think astro-CCDs have 16-bit ADC for that reason. My 450D has a 14-bit ADC and should be slightly better, for astrophotography where we are photon-starved, than ones that have 12-bit ADC.

Clarkvision.com: Nikon D200 Noise and Full Well Analysis

[ollypenrice]

Another vote for stickying Themos!

Olly

[steelfixer]

Themos - have you ever thought about teaching this stuff, Even I can understand that last post.

[SamAndrew]

So what have we all learnt from todays lesson ?

You'd better start revising because I hear there's going to be a test

[themos]

Question 1: Why are the "gains" of astro cameras (that's the number of electrons needed for each increase in the readout by 1) usually close to 1.0?

QSI 600 Series - Cooled CCD Cameras

[themos]

Question 2: Google the answer to "which two ISO settings for your camera bracket the gain=1.0 electron/unit-recorded condition?"

[ollypenrice]

I was going to ask if 'gain' and ISO were effectively the same thing? I don't (seriously) use webcams and nor do I image with DSLRs but they seem to be much the same thing, Gain and ISO. Is this correct?

Olly

[themos]

I think so, Olly, yes.