What I cant understand is......everyone tells me to use like 800 ISO or it is favored even 400 is very popular....what i would like to know is what's the advantages and disadvantages of using high or low ISO's.
I have found that the very high ISO's like 1600 or even 3200 enables me to get much more detail then at 800 for my 30 second exposures....is this because my exposure's are short....and if this is the case whats the difference doing a 2 minute exposure say at ISO 800 and doing a 30 second exposure at 3200 ISO.
I did some 90 second exposures at 400 ISO and also some 30 second exposure's at 3200...I thought the 30 second exposure gave me more detail....obviously if I went much longer at 3200 it would be all washed out in a ball of light....so at the moment I am failing to see the full benefit too long exposure's in suburban area's....I must be missing something.
One of the problems of using a higher ISO number Tony is that you can get more "grain"/"noise" in your image that then needs to be processed out if possible. Lower ISOs are usually "smoother"/"quieter" but your exposure needs to be longer.
I see Paul....then considering Grain/Noise can easily be removed through software why would someone do a 2 minute image when you you can do the same in 30 seconds.????????
Grain/noise can't always be removed.. not all of it. Stacking multiple shots will help reduce grain and noise, so many short exposures with a high ISO may be the same as a long exposure at a low ISO.
Noise results in loss of information that cannot be recovered with filtering, no matter how clever the algorithms used. Filtering only improves the appearance of the image, does not recover the information.
What I would like to know is how come digital cameras have adjustable ISO ratings? Isn't the photomultiplier just counting photons that fall in a tiny bucket (pixel)? All you want from the camera is to tell you the number of photons that fell in each of so many million red, blue and green buckets. Where does ISO fit in this? This is the way I see it:
With film, higher ISO means increased grain size, so each grain catches more light. You are sacrificing resolution for improved light gathering ability. The digital equivalent I suppose is combining (adding) the intensities from several neighbouring pixels to get more photon counts, at the expense of reduced resolution. But this could all be done and fine tuned with software (the digital equivalent of the chemistry used to adjust the grain size in film during the development process). In which case you would be best off using the lowest ISO setting all the time to give you maximum flexibility, because you can always increase the ISO (to improve the signal) in post-processing but never decrease it (to improve spatial resolution).
Or is there more to it than that? I have no experience with digital photography but I am very curious about these things because some of the digital SLRs seem to be getting up to 35mm film quality, and prices are coming down too...
It's much more complicated than we even imagine.The relationship of noise vs ISO depends on many variables but at a given temperature is the same for any sensor and consistent.
Never work with JPG's as the compression introduces noise that is even worse than the real noise.So take home message is start with raw frames, convert to tiff and also do this for dark and flat frames.
Two good resources for info on sensors and noise and comparison to film is
Tony, the 20D has very low noise and in fact the noise performance, at even the highest ISO, is probably far better than that of a prosumer digicam, so the shorter you can make your exposures (less tracking error), or the more light you can gather in your current 30 second exposure the better. If you're doing really long guided exposures then the lower the ISO the better. Unless the noise is noticeable don't worry about it. BTW many sharpening algorithms actually add noise to achieve their effect.
Thanks Guys.....I know its a tuff subject....I am just kind of wondering what benifit I would get by lowering my ISO and taking longer exposures....
If you can see it from my point.......the shorter exposures at higher ISO was giving me the same results in a 1/4 of the time....why would I think about taking longer exposures at lower ISO's for the same result...and with this comes the need of very accurate guiding.
Dont get me wrong I know their must be some benifit to long exposures with lower ISO as everyone does them..I just cant see it in the shots I have taken.
Maybe the Loss of rezolution from high ISO is something that I dont see considering your dealing with 8 Megapixels of rezolution....considering we are reducing it under 1 megapixel to post on the forums.
I know I have lots to learn and this why I posted this thread....I just need some convincing about low ISO and longer exposures.
This is an interesting thread for me, even though I'm only using a cheap DSLR camera (kodak easyshare). I particularly pricked my ears up at your comment avandonk..I think my camera actually shoots in JPG? (for want of a better term, sorry) at least, that's how they get saved to the computer by default?
Hmmm. I have no other options, it only takes in JPG. Is it a waste of time converting to BMP after the fact? I would have thought so. My pics have usually got a hell of a lot of 'grain' in them..
Last edited by asimov; 09-09-2005 at 09:36 PM.
Reason: stupid typo!
Thanks cventer, I have not digested it yet but it looks like it explains some
of the difficulties and misconceptions I had.Not to mention the conundrums!
Again thanks.
the higher the ISO the shallower the photon bucket is before it registers as 'full' and shuffles a signal off the bus recording a 'hit' against that pixel.
Wow that sounds exciting....I have a photon bucket.....lol
Thanks guys.....the links have helped.
Chris..I have just read the link you posted with "Terry Lovejoy"
What I do find surprising is his comparison shots at different ISO's with same time exposures...their hardly any difference....If I took say 1 minute exposure at ISO 200 then the same at 1600 the difference is incredible but his shots are hardly noticeable....??????
Below are 2 x 45 second exposures of the same target M17 on the same night but with 2 different ISO's both unproccessed images resized and compressed...the brighter one is at ISO 3200 with the duller at ISO 400 I prefer the 3200 image.
Could you provide links to the originals (uncompressed raw RGB), Tony?
Edit: Had a good look at and tweak of the jpegs. The ISO 3200 definitely has more detail, and is all-round better, than the ISO 400. But jpeg compression can really kill fine/soft detail.
If you haven't used the raw setting Tony, it's in the menu's and can be used in any of the creative modes. You can set the 20D to take both raw and JPG together and have the best of both worlds. BTW You'll need to go to photoshop's web site for the 20D raw plugin for photoshop.
Thanks Phil....yeah I know how to set for Raw but havn't used it as the programs I have been using dont support it....so I have just been using the best quality JPEG setting for the time being.
Steve I will upload the original Jpeg for both pictures...just give me 5 minutes from the time of this post.
Why do the ISO200-ISO1600 images look the same on my webpage - even though they are the same exposure? The reason is that I have boosted the luminance on each ISO so that it matches the ISO1600 image. That means the luminance in ISO200 image was multiplied by 8x and so forth. These tests must be done in raw mode.
What I was attempting to show was that there is no more information in the image from ISO200-1600 for a reasonable exposure. On my 300D for exposures of several minutes or more ISO200 seems to record as much detail as ISO1600, although background banding looks a bit ugly and I tend to use ISO400 as a result.
Your 20Da will have different characteristics as it seems ISO800 and 1600 have very low noise relative to 200/400.