In this series of articles, I will describe an example of where photography is reaching the limits of physics, namely, when light is scarce. Many photographers expect even in this case that their camera delivers perfect photos. Here, I will try to explain why this is not possible. And I would also like to address the differences between large and small sensors, because I have the impression that many owners of compact cameras expect too much from their small-sensor cameras, among others, because they ignore or do not know what the difference implies.
This first part in the series serves as an introduction to the topic, asks for reasons for unsatisfactory, particularly blurry, photos, and discusses possible "countermeasures". In the second and third parts, which are supplemented by additional pages with test images, I compare cameras with sensors of different sizes with respect to their performance at different ISO values. In the fourth part, I present the concept of an "exposure value" (EV) as a measure of the amount of light that is available for a photo and present sample photos for illustration. In the fifth and sixth parts, I present examples of situations in which only little light is available. For illustration, I provide the exposure values and other photographic data. In the seventh and last part, I try to summarize the results.
In the age of digital photography, and even more so in the age of cell phone photography, many owners expect their cameras and mobile phone cameras to do miracles. Above all, they are supposed to produce photos that are always perfectly exposed and sharp - even when the light is scarce. But unfortunately that's not possible! Photography could be described as "painting with light." Not a brush does the painting, but an electronic device that has to obey certain physical laws. One of these is that a camera requires sufficient light to be able "paint." I do not want to dig deeply into physics in the following. Instead, I will after, an introduction to the topic, demonstrate using practical examples that virtually all cameras get into problems with image quality when light is scarce. However, there are differences between different types of digital cameras, which are basically caused by the size of the built-in sensor. As a simple rule we can remember that the larger the sensor, the more light falls onto it, and the better the image quality. Also follows from this that the larger the sensor of a camera, the longer it can maintain some minimum image quality when light is scarce.
Interestingly, not necessarily night scenes where bright lights stand out from a dark background or where a brilliant evening sky is seen present a problem. Here you usually know that you have to do something in order to obtain usable photos, for example, support the camera on a tripod or otherwise. More likely images of celestial bodies provide a problem (except for the Moon), because some cameras do provide the required long exposure times or because the results are unusable (in simple terms: the sensor is getting hot during long exposure times, and the image gets noisy). Here, I will not bother with this application because it is very special and concerns only a small group of people.
The problem that I would like to address typically affects indoor shots, as well as outdoors in the morning or evening for example, I am too late in the evening with my photos...), or on cloudy winter days, where you would like to get results like during a summer day. With respect to interior shots, we are often disappointed that the images are too dark and / or blurry - and even with the aforementioned types of outdoor scenes we quickly achieve (long) exposure times that are prone to camera shake. If we look at the images more accurately, they can also look "mushy," like painted, or totally look (or all at once), because we have increased the ISO value strongly to achieve acceptable shutter speeds. Who then scolds his or her camera, probably asks too much of it or uses the wrong camera for the intended purpose. But even my best cameras reach their limits with such shooting conditions - admittedly, I had not necessarily own cameras that do miracles in low light... But I owned and possessed cameras with sensors of different sizes, so that I can appreciate the behavior of cameras with large and small sensors at least a little - and I will write about this here.
Let me start with the problem that photos are out of focus. Photos may be blurred for various reasons. In the following, I will give a few reasons for this and discuss them briefly.
With automatic focusing (AF), it may happen that the focus simply "goes astray." This happens particularly when there is little light available. The reasons why an auto-focus "goes wrong," can be varied and would fill a separate article. Here I simply want to state that this can happen, and indeed it is more likely the less light is available. Therefore, most cameras offer an autofocus assist lamp, but this does not solve all the problems (for example, because it does not reach far enough) - or you do not want to use it in the given situation, because it is too bright or may draw too much attention.
Manual focus can, of course, go astray even more than autofocus, for example, because you miscalculated the distance, or cannot recognize the target in the viewfinder sufficiently to focus on it (that happened to me, for example, at a wedding - luckily I was not the "official "photographer).
In low light, you can quickly get into the range of longer exposure times. Some photographers may still hold 1/30 s out of hand, some even a 1/15 s or 1/8 s, but at some point there is an end to this. Some cameras have devices to prevent camera shake, or at least reduce it (some companies advertise with up to 5 f-stops profit). In my experience, it is rather a decrease than a prevention and therefore not always leads to satisfactory results - but that could be due to the cameras that I have used and still use. Of course, you can place the camera on a tripod or place it on a firm surface to avoid camera shake, but both measure are only suitable for static subjects (or if you want to capture movement traces). And not always you have a tripod with you or can find a suitable surface...
Anti-shake methods cannot prevent motion blur (for example, people turn their heads, move their arms, children run around ...) if the exposure times are too long. But not always such blur is perceived as bothersome.
So our aim should be to achieve the shortest possible exposure times, times that can still be hand-held so that possible blur is kept low. In the following, I would like to mention three methods that are ment to serve this purpose and that are most certainly familiar to photographers.
Many cameras have a built-in (mostly) small and weak flash. This flash is sufficient to adequately illuminate a few people or objects nearby. It neither suffices to take pictures in a stadium (although there is always a flurry of flash lights from the stands - but actually there is usually enough light to take photos of the arena), to illuminate large interior spaces such as churches or halls, or to photograph a larger group of people. The light falls off very rapidly, so that the front is lit, often even overshadowed, and in the back everything looks dark.
The "flash" topic fills myriads of books, and I therefore would like to keep silent on this - except for mentioning that I only use a flash when there is no other way. I simply do not like how images that are illuminated with a simple, in-built flash look, and I also do not want to invest more effort in this. But the fact remains that a flash can "freeze" movements and can also largely prevent camera shake. By the way, most digital cameras offer a "slow sync" functionality for the flash where an object in the foreground is illuminated and "frozen" by the flash, while the background is illuminated by ambient light, resulting in a long shutter speed (probably requiring a tripod). The problematic thing that I found is that some cameras use "slow sync" as their default flash mode in their so-called "intelligent automatic" mode that many users activate by default. So they wonder afterwards why their flash pictures are blurred. It would certainly be better to offer a normal flash mode in such a case.
This is certainly the approach to shorten exposure time with the least "side effects" - except for that the depth of field at large apertures can be very narrow (for details see ...) and that fast lenses often large, heavy, and expensive.
In addition, your own camera may not have a fast lens built in, and that's unfortunately the one that is available to you... In my example, my Leica Vario X has a maximum aperture of f/3.5 at 28mm and at telephoto range (70 mm) even drops to f/6.4. This makes it hard for me to shoot indoors when there is little light available, for example, when taking photos of distant details in churches...
Now I have come to the "lever" that is probably most closely associated with the topic of "sensor size." Virtually all cameras offer an "auto" setting for the light sensitivity at which the camera selects the ISO number automatically (sometimes, you can specify lower and upper limits for the ISO range). Many amateur photographers only use this setting (I did this in the past, too), and if you also use certain automatic settings of your camera (for example, "intelligent" automatics), you cannot take any influence on this. At best, you can choose a different automatic mode, which provides you with greater control of the ISO values.
If the camera, however, allowed you to change the ISO setting in a particular mode, this is, of course, THE screw to turn first (with a DSLR or system camera, you can, of course, also change the lens first, if a fast one is available).
Unfortunately, to this "screw, " too, the old dogma applies: "No effect without side effects!" I would like to clarify this in the following. Here I would like to conclude with stating that, within a certain range, increasing the ISO number is a good way to get to shorter exposure times and thus, effectively prevent - or at least reduce - camera shake and motion blur.
What other reasons may there be for unsatisfactory photos? Once we assume that the front element of the lens was clean, the following two reasons come to my mind. Further below, I will, however, only deal with the second.
Lenses of simple cameras are generally less sharp than expensive interchangeable lenses or fixed lenses on expensive cameras. For example, on the computer screen, I can clearly recognize the difference between photos made with the Ricoh CX4 with a small sensor and my APS-C cameras, although the screen has fewer pixels than the photos that all these cameras produce. However, the difference between photos made with my APS-C cameras and those made with my wife's Sony RX100 with 1 "sensor is barely visible on the screen. Other factors also contribute to these differences. For example, my wife's Ricoh compact cameras sometimes showed blurred areas on the sides or in the corners that looked disturbing (the lenses were poorly centered; therefore, the cameras went into the repair). This is an example of lack of quality control.
Top cameras and lenses are not necessarily sharper than others in the center, but they are evenly sharp over the entire image field without too much sharpness drop in the corners (all lenses are less sharp in the corners than at the center). The sharpness of the lens of my wife's Sony RX100, for example, decreases in the corners significantly; but I am satisfied with my Leica X Vario's lens, even though it is a little weaker in the corners. Meanwhile, many camera manufacturers correct lens weaknesses in software - not always for the better ...
Lenses are huge topic to discuss. Here, I just wanted to suggest that the lens of a camera can also be the reason for unsatisfactory photos - and, as described, for a number of reasons. When fixed lenses you unfortunately have only the option to change a whole camera and not just the lens ...
Now I a finally have arrived at the topic of "side effects of higher ISO values". There used to be film emulsions with different ISO (ASA/DIN) values to cover the various needs of photographers. Already then we knew that high-speed films could not keep up with the quality of low-sensitivity films, they were grainier and more contrasty.
The ISO setting is basically a "gain control" setting: If you have less light at your disposal, you increase he gain. For audio amplifiers, we know that this will lead to increased noise - and cameras are no different: you get more luminosity and color noise. Ultimately, the photos are thereby more "gritty" (luminosity noise) and "colorful" (color noise). How do the camera manufacturers address this issue?
Well, they answer with sophisticated algorithms (calculation methods) that are probably almost as secret as the recipes of the alchemists in the Middle Ages. And in fact, there are considerable differences in the way how different camera manufacturers and publishers of photo editing programs handle this. This is evident, for example, in comparative tests such as camera tests on dpreview.com.
Camera manufacturers use two basic approaches: Some reduce the noise only weakly (and are often criticized in camera tests...), but more details remain in the photos - and photographers can process the photos with better success. Others "iron out " noise generously and provide the photos with a watercolor-like appearance, so they look like "painted" when they are enlarged to 100%. This is done sometimes to a larger, sometimes to a lesser degree, and some companies start this at lower ISO settings than others. Such photos cannot really be "saved" in post-processing...
From which point on we can say that we do not like a photo due to its noise or the actions against the noise is certainly a very personal thing. Often you can read that the photos that a camera XYZ produces at high ISO values are suitable for Web images, but not for print use. This is one of many examples that show that the question of whether the image quality of a photo is still acceptable, is associated with its intended use. When I took pictures at conferences, image quality was of secondary importance, because I made the photos primarily for use on Web pages. The situation is different, for example, when taking photos of family gatherings, which you want to print and give away later.
Typically, cameras use sensors with a so-called AA filter (or anti-aliasing filter). This is meant to prevent the formation of moiré in certain regular patterns (for example, nylon stockings), but it makes the picture a bit less sharp than when no filter is used. In recent years, manufacturers have increasingly made the move to offer cameras without such a filter (or to offer even two versions, one with and one without a filter). The differences can be recognized, but only at the pixel level, that is, if you look at pictures in 100% magnification. I do not want to discuss issue further, but wanted to mention it briefly. In the next two parts of this article series, I will show photos made by a camera (Ricoh GR) without such a filter.
If the light is scarce, the limitations of digital cameras show. You can indeed take measures to counter fuzzy and blurred images, but only within certain limits - the use of a flash, the use of fast lenses, and increasing the ISO value are possible measures. However, the first two measures are not always applicable. Thus, increasing the ISO value seems to provide a "silver bullet." However, if you increase the ISO value, you arrive, sooner or later, at image results that are no longer acceptable. As a general rule of thumb applies (no rule without an exception!): The smaller the sensor, the sooner you will arrive at this limit. And, as I have shown above, also on the subsequent use of a photo depends, whether its technical quality can be considered as acceptable or not.
In the second and third parts of the series, I compare cameras with sensors of different sizes at different ISO values. Low light typically means that you have to use higher ISO values. As already described, this "silver bullet" runs against its limits, especially for cameras with small sensors. Here, I want to demonstrate this with vivid examples.
In the fourth part, I try to be define more precisely, what low light means, and introduce the concept of an "exposure value" (EV) as a measure of brightness. This approach will accompany us in the subsequent articles.
In the fifth and sixth installment of the series, I introduce examples of situations where there is typically low light - and where I have experienced this and can demonstrate it with photo examples. For your convenience, I provide the exposure values and other photographic data for the photos. Sometimes, I used cameras of different sensor sizes, so that I can directly show the differences between them.
In the seventh and last part, I try to summarize my findings and to reach conclusions.