ISO explained
After aperture settings, possibly the most misunderstood setting on modern cameras is the ISO setting. There's a lot of confusion regarding ISO and how it effects your images. Hopefully, you're confusion will be gone after reading this article.
I suppose the first thing to explain is what ISO actually is. Housed in Geneva, the International Organization for Standardization - or ISO - is, as the name would imply, an organization founded to create international standards for various industries.
When film is manufactured it can be made in various levels of sensitivity to light. Low light conditions required film that is more sensitive etc. In order to ensure proper exposure, it is vital a photographer knows exactly how sensitive any particular type of film is. That sensitivity rating is, in our industry, simply known as ISO. The higher the number, the higher the sensitivity of the film.
Traditional film has been replaced with a light sensitive chip in digital cameras but the ISO standards still apply ensuring the sensitivity ratings of the digital sensor, regardless of manufacturer, match.
Unlike film that is manufactured at one specific sensitivity, digital cameras have the ability to vary the sensor's sensitivity using electronic amplification. It's not unlike turning up the volume on your radio. You might even think of ISO values as a simple way of reporting the level of amplification.
You might be inclined to believe that just keeping the sensitivity turned all the way up seems like a good idea. That's not really the best idea. There may be times you just don't want all that sensitivity. In bright light your images are likely to be overexposed, especially if you want to use long exposures such as the flowing water shown in the previous post about shutter speed.
When slow shutter speeds are used to create images like this, its necessary to reduce the sensitivity of your sensor as much as possible and also stop the aperture down to reduce the amount of light entering the camera to avoid overexposure - or an image that's too bright. In fact, sometimes if the light is very bright or you want a longer shutter speed than the low ISO/small aperture combo can provide, it may be necessary to add a dark, neutral density filter to the lens to limit light further.
So, you can see why you might want lower sensitivity at times.
However, at other times there may be conditions that support more sensitivity. For example, the photograph of the dragster from the previous post that was shot at 1/8000 of a second.
The long focal length lenses sport photographers use have smaller maximum apertures than most shorter lenses. When combined with the need to photograph at high shutter speeds to stop action, increasing the cameras sensitivity is the only available option.
So, is that the end of the story? High ISO for fast shutter speeds and low ISO for low shutter speeds?
I'm sorry to deliver such a sad message, but no. In fact, if that were the case, the manufacturers could simply link the sensitivity of the sensor directly to the shutter speed so that it would change automatically for us as we changed shutter speed. Oh what a simple life that would be.
Life isn't simple friends.
ISO has a much different trade off than just the shutter speeds it provides for, although that's definitely one of the primary considerations when choosing ISO.
The bigger picture is actually how ISO effects your image. Like everything else, there is a compromise that comes with ISO. Just like turning up the volume of your radio, quality begins to suffer as you increase the ISO. And, just like your radio volume, that loss of quality may be minimal at moderate increases but when pushed to its limits you can't fail to notice it.
The quality degradation I'm talking about is, of course, noise. Electronic noise is technically defined as an unwanted disturbance in an electrical circuit. In photographic terms, it means your picture becomes grainier and less sharp.
Low ISO |
Low ISO magnified |
High ISO |
High ISO magnified |
The generally accepted belief is to always best to capture images at the lowest ISO possible. That is naturally pretty sound advice. Provided we properly define "possible." Because all images can be captured at ISO 100. They might be terrible, dark, blurry, etc. But they can be captured at 100 ISO. So possible may be a poor choice of adjective in this case.
I think a better way of thinking might be to avoid sacrificing quality needlessly. Quality is not defined solely by lack of noise. A bit of noise in an otherwise tack sharp image is my preference over a noiseless image rife with camera movement for example.
Yes, even pros get sloppy at times. |
I got sloppy and shot this session at 160 ISO which meant my shutter speed ended up a bit lower than the long lens I was using should dictate. I got by with most of the images, but a few, such as this one, had a bit of camera movement in them.
A higher ISO wouldn't have made a significant difference in the noticeable noise. However, a sharper image would definitely increase the quality. So in this case, using a lower ISO meant I actually lost quality rather than gained it through less noise.
So please don't fall into the trap of believing that noise level is solely responsible for quality in your photographs. It is simply one of several factors that work together to, collectively, create quality.
So please don't fall into the trap of believing that noise level is solely responsible for quality in your photographs. It is simply one of several factors that work together to, collectively, create quality.
So exactly how much noise is too much noise? If you ask 10 photographers that question, you'll receive no fewer than 10 answers. Noise tolerance varies from person to person. My experience is that younger photographers tend to be less tolerant than seasoned photographers. Perhaps we've just accepted it as part of the process. Or, maybe its due to our experience with high ISO film which was far grainier than any camera today produces.
Another distinct possibility is that seasoned photographers have learned to view images at reasonable magnification rather than the extreme "pixel peeping" I witness constantly. Yes, at times, I'm guilty of some degree of it myself but I can recognize when I'm being unrealistic.
Let's use the gazebo image as an example. I didn't do any fancy math. I'm no mathematician. But I think, when viewing the entire image with the enlarged portion highlighted, you'll agree the enlarged portion is certainly less than 10% of the entire image. So if we fill our screen with only a portion of the total image, how large have we actually made the entire image?
Let's use the gazebo image as an example. I didn't do any fancy math. I'm no mathematician. But I think, when viewing the entire image with the enlarged portion highlighted, you'll agree the enlarged portion is certainly less than 10% of the entire image. So if we fill our screen with only a portion of the total image, how large have we actually made the entire image?
I find, due to our monitor sizes and ability to magnify, we often view image far larger than realistic. Most of us don't produce billboards and yet we look at images magnified to that level. And then view them from only a foot or two away. Would you ever view a billboard from that distance? Of course not. So be reasonable with magnification, especially when you're judging your image regarding noise.
Here's how I suggest your camera. Go outside on a good, bright, day and take a photograph of the same subject at each ISO from low to high. You must do this when the light is decent because if you record lousy light, you'll get a lousy recording no matter what ISO you're using. That's not a fair test.
Probably the easiest way to perform this test is to set your camera to Program Auto or P on the mode dial. That will allow you to change the ISO and the camera will automatically determine aperture and shutter speed. Since this is a test of noise and nothing else, those settings aren't terribly relevant as they do little to effect noise.
Once you have your images, view them at whatever level of magnification you feel is reasonable. You'll learn there is very subtle, if even visible at all, differences between small ISO changes. But somewhere between your lowest and your highest ISO you'll begin to hit the edge of your comfort zone. The place where images at higher ISO are noisier than you'd like for your day to day shooting, but ISOs lower than those seem to show little improvement.
When you find that middle ground you're ok with, you've found your base ISO that I suggest using for your standard work. Armed with the information of how your camera performs at each ISO you will be able to decisions for other needs when they arise. I have a few zones of ISO if you will. My midrange, normal use ISO, my low ISO for lowest noise or long exposures, and my high ISO range. I rarely if ever use the highest ISO my cameras provide but I do push my comfort boundaries as needed.
So why can't or won't I just suggest what ISO you should use? Why am I telling you to test your ISO? Two reasons really. First of all, I have no idea of what you consider acceptable and what you don't. Secondly, your camera and my camera won't perform the same way. In fact, none of my cameras perform exactly the same as the others. What? Why not? Isn't ISO standardized?
Yes, ISO is a standard. Every camera will create essentially the same exposure in the same light and same settings although even that can vary from model to model. But what ISO has not standardized is how noisy the image is at any given ISO. Different cameras simply perform differently. There may be a noise level standard at some point but to my knowledge it doesn't currently exist. And in any case, ISO is simply sensitivity. Period.
So what makes one camera or another camera more or less noisy at any given ISO? Fortunately there's a relatively simple answer to that question. The age of the camera and the sensor size will both effect the level of noise.
Each generation of camera performs a little bit better than its predecessor when it comes to higher ISO performance. This is due in large part to improved in-camera image processing and noice reduction. My older and beloved broken Canon 7D, for example maxes out at 6400 ISO whereas I can push my new 90D body to 25,600 ISO.
Yes, I said broken 7D. I don't believe you always need the latest and greatest gear and have dedicated another blog to that very point. Its about using older, less than perfect and downright broken gear. If you're interested, you can view that here: https://functionallyunfunctional.blogspot.com
Yes, I said broken 7D. I don't believe you always need the latest and greatest gear and have dedicated another blog to that very point. Its about using older, less than perfect and downright broken gear. If you're interested, you can view that here: https://functionallyunfunctional.blogspot.com
If you're using a little bit older, or even broken, camera that doesn't provide high ISO, don't fret. At lower to midrange ISO the differences between an older body and newer body are quite subtle. These two images, both at 800 ISO show that neither camera has a huge advantage over the other.
7D 800 ISO |
90D 800 ISO |
As you can see, at this ISO these two cameras are performing quite acceptably, at least by my definition of what's acceptable.
A bigger leap in low ISO noise performance comes with sensor size. Larger sensors have larger pixels. Small pixels are noisier than large pixels.
I have both the EOS R and the 90D. The EOS R is a 30MP full frame sensor. The 90D is a 32MP APS-C (crop sensor). So the resolution of both cameras is quite fairly matched meaning the primary difference between the two ends up being pixel size.
90D, ISO 6400 |
EOS R, 6400 ISO |
To around this point the difference is subtle. The 90 D is holding its own but the R has an advantage.
However, at higher ISO it becomes a no contest battle.
However, at higher ISO it becomes a no contest battle.
90D, 25,600 ISO |
EOS R, 25,600 ISO |
Now before all of my readers rush out to buy the EOS R or some other full frame body because its clear that full frame cameras are better....
Let's look at the difference at 800 ISO which is likely a setting you'll be using far more than 25,600 or even 6400.
FYI... I didn't originally photograph the setup with the EOS R set to 800. I was originally only trying to demonstrate the high ISO differences. When I set it back up I didn't perfectly nail the original lighting, angle, etc. Try to look past those differences and only look at the amount of noise in the three images as you compare them.
EOS R, ISO 800 |
90D, ISO 800 |
7D, ISO 800 |
Once you've done that, you'll be ready to start your juggling act with the three settings that control your exposure - aperture, shutter speed, and ISO.
I know you may be wondering so on most of my cameras I'm routinely have my ISO set between 800 and 1600. That generally allows me to keep my shutter speed high enough to prevent camera movement. I drop or raise it as needed but shutter speed is generally my priority so I raise the ISO (within reason) to afford me the shutter speed I desire. And, of course, as mentioned last time, I use camera supports a lot!
Comments
Post a Comment