I’m hoping to get a relatively detailed, but, plain language description about the difference between your High School graduation photo, and the screen you’re looking at right now.
How grainy is that vs. this?
I’m hoping to get a relatively detailed, but, plain language description about the difference between your High School graduation photo, and the screen you’re looking at right now.
How grainy is that vs. this?
Short version: You don’t convert ISO to megapixels; they measure completely different things. ISO measures sensitivity to light (in both film and digital). Megapixels measures spatial resolution, the same way film size (“35mm, medium format, large format, IMAX”) does.
Long version:
ISO measures the digital equivalent of “film speed” and is a more or less a direct conversion from the old ASA system of film speed numbers (100 is smooth but not very sensitive, 800 is sensitive and grainy, etc.). If your high school graduation was taken with 200 speed film, a digital camera set to the same ISO would look roughly similar in terms of grain. This is complicated by the differences in film vs digital light sensitivity. In film grain, crystals react to light differently. In digicams, noise is caused mainly by electrical fluctuations in the circuitry and not necessarily by stray photons hitting the sensor (although that happens too).
Megapixels, on the other hand, measures resolution (how much detail an image holds). Your standard 35mm film, according to Wikipedia,can hold as much detail as approximately 4 to 16 megapixels but that is an imperfect comparison at best and sort of useless at worst, because a lot of factors determine image quality, especially when you’re converting between film and digital, that the catch-all term “resolution” doesn’t quite describe it all. For example, with film the limiting factor isn’t usually the film itself (meaning the negative) but the processing equipment and subsequent prints. With digital, the megapixels don’t fully describe the amount of detail because it doesn’t take into account sensor size and circuitry design (both of which contribute to image noise in a similar way that a high ISO setting would). And most digital cameras, especially consumer-oriented ones, include software that artificially sharpens and/or smooths pictures as soon as they’re taken, erasing raw detail but hopefully making the picture look better to the untrained eye. (The alternative is to turn off this in-camera post-processing and to do it yourself with special software, but the results aren’t necessarily better there either.)
Aaaaaaand… to further complicate matters, how sharp a digital picture looks also depends on your screen, not just the photo itself.
Pixel density of a screen is correlated to but not the same thing as screen size. For example, a 50" screen can have anywhere from less than 1 megapixel (720p) to more than 8 megapixels (4K), and the 4K screen will look a lot sharper even though the screens are the same physical size.
Think of this way: You can build a face out of, say, 50 normal Lego blocks. Then someone asks you to make it bigger, but instead of using more Legos, you instead buy 50 giant Lego blocks and recreate the same face. But because you still only used 50 blocks, you haven’t added any DETAIL to the face, you just scaled everything up.
If, on the other hand, you can 500 mini Lego blocks, you could create a much more detailed face in the same physical area.
The analogy is that each Lego block is a pixel, and to add more detail you need to add more pixels, not just make each pixel bigger.
To see this phenomenon in action, go to an Apple store sometime and compare a same sized Macbook Pro with a Retina display to one without it. You can also sometimes see this effect if you compare a 720p TV to a 1080p one, but more often than not, stores like Target and Best Buy use bad video compression so the demo pictures look like shit on either set anyway.
I do not think this is true. Some digital camera’s photographs at ISO 3200 will look very grainy (noisy) and another (more expensive) camera’s photograph at ISO 3200 will look very clean.
You might want to take a look at this. The only problem is that, like a lot of Wikipedia articles, it does not seem to be very layman friendly. It seems that digital manufacturers can use several different methods of determining how to assign ISO values, so looking for some kind of easy one-to-one correspondence between film and digital ISO might be difficult.
High-ISO pictures from digital cameras, if they look “smooth” not grainy, have probably had a computer-applied processing (by a hip in the camera) to identify and eliminate “noise”. This can result in a very “plastic” look or smearing, where there is no tiny detail because the processing removed all tiny detail.
The difference between digital and film - the longer film is exposed, the more light that hits the film, the more film exposure happens. Sometimes, like night landscapes or astronomy pictures, you can have exposures into minutes and hours. Digital suffers from electronic “noise”. Random effects - like the “snow” on an old TV screen - becomes amplified and more and more prominent the longer the exposure. Long digital exposures are usually a series of superimposed short exposures.
ISO, like old film ASA, measures how much light is needed to get the ideal picture - not too dark or too washed out. Rule of thumb, ISO 100 at 1/125 sec and f8 (aperture of lens, with/focal length) was usually good for sunlight scenes. ISO 400 and 1/30 f2.8 might be good for decently lit indoor shots. Digital cameras adopted this numbering because it was familiar. Really, the camera takes whatever exposure it has and if possible, makes a passable picture out of it.
The other thing is resolution vs. ISO. In film, light exposure impinged on silver halide; when processed, the areas with the most light “reduced” (developed) into the biggest silver particles. Higher ISO meant the silver when developed, produced bigger “grains”, hence grainier and less detail photos. Sometimes this is an artsy effect, but generally, the smooth look of tiny grained low ISO photos looked more realistic.
With digital, as the exposure involves less light, the longer exposure meant more “noise” in the picture. The advantage of more megapixels may be that with a huge photo, there may be decent detail in spite of the amount of noise. (not to mention blur from movement during long exposures, and electronic focus not working well in low light) The sensitivity of modern digital cameras means that you can get passable pictures even incredibly low light conditions. However, for nice pretty realistic pictures, nothing beats decent lighting.
the trouble with a flash is that the light intensity is 1/4 as bright twice as far from the camera. If you have a variety of distances in the photo, either the near ones are overexposed or the far ones are underexposed,
Excellent reply, Reply. There is one more factor that can contribute significantly to image quality – compression settings. While professional cameras can all save data without any, most consumer cams that I have seen store images as compressed, lossy JPG, and they rarely give the user options to set the compression ratio, or specify it in vague terms like low or high. While it looks terrific in snapshot size, enlarging a typical image from a consumer grade cam will reveal many artifacts imposed by the compression scheme.
This situation is at its worst in cameras like the GoPro, primarily oriented towards video, but can take stills. In my Hero 2, there is no setting for compression ratio, and it is much too high for my taste. They get by with this in video mode, since any artifacts introduced into one frame are negated by different artifacts in the next one, and it just looks like video noise.
You don’t really get “grain” out of digital photographs the way you do with film. Since film uses silver halide crystals and various dyes, smaller crystals = more detail = longer time (i.e. lower ISO value- less sensitive). This is because a larger crystal has a bigger chance of having photons strike the crystal and initiate the chemical change, and smaller crystals are smaller targets. So higher ISO films like ISO 3200 have drastically larger crystal sizes versus say… ISO 50.
Image sensors don’t work the same way- my Canon T2i has a 18 megapixel sensor, regardless of the ISO I’m shooting at. So if I set it at ISO 100 or 12800, my individual pixels are the same size.
The difference is that at higher ISO values and/or exposure times is that “noise” increases. It’s kind of like grain, but it’s not quite the same. It’s basically what happens when you’re operating near the extreme edges of what your sensor can handle. It’s also affected strongly by sensor size- i.e. a larger sensor may have less noise at a higher ISO than a smaller one at a lower ISO (the Wikipedia example is that a DSLR sensor shooting at 400 ISO may have less noise than a point-and-shoot sensor at 100 ISO).
In practical terms, what this means is that back in film days, it was a combination of the actual film type and the actual exposure that determined grain- higher ISO film and/or underexposure tended to make grain prominent.
On digital cameras, it’s mostly a consequence of long exposure time and high ISO settings, BUT… a good digital camera isn’t going to look grainy at say… 400 ISO no matter what you do to it, save post-processing with something like DXO Filmpack to artificially introduce film-style camera grain.
If you’re shooting say… Tri-X 400 in a film camera, it’ll be grainy even on the correct exposure.
That’s actually a bit overexposed. The rule of thumb is known as “the sunny 16” rule. When it’s sunny (full sun), at f/16 your exposure is approximately 1/ISO film. So, for ISO 100, you’re looking at 1/100 sec, or 1/125 if that’s the closest you got. And if you’re shooting positive film, you might want to slightly underexpose anyway. Shooting neg, I might do 1/125 @ f/11.
But it can vary a bit from lens to lens. In theory, they should all produce the same exposure at the same settings, but in my experience, I’ve seen variance up to about 2/3 of a stop.
I think it’s safe to say that most, if not all (at least all current models) digital SLRs, even consumer-grade ones, give you the option to save images in RAW format. Which doesn’t really conflict with what you say, since most digital cameras sold are NOT SLRs, but I just wanted to put that out there.
I don’t doubt that’s the case. But there are many good cameras out there in the under-SLR price range that take amazing pictures for only $100 or so, but I’ve never seen those able to store uncompressed images, nor do they allow much – if any – adjustment for JPG compression. And the lack of adjustment seems to be a new development in the marketplace; I’ve owned older cams in the same market niche that did provide what they often called “fine” or “super-fine” modes for compression settings.
Probably because memory cards have gotten so big and cheap. In the past 10 years or so, memory card size has literally become 1000 times larger. (I still have a 16-megabyte card from about 10 years ago; now you can get a 16-gigabyte card for $12.) In an age when a $12 memory card can store 1000 images at 20 megapixel and reasonably low compression, do you really need an option for storing 2000 images at higher compression?
My guess is he’s saying the opposite: that he wants options for less compression, right?
Yeah, I’m a bit surprised. I don’t deal a lot with P&S digitial cameras, but the ones I’ve dealt with have had variable JPEG settings. They did not have options for raw files, though, that’s for sure. But the P&Ses I’ve used have also been 5-10 years old, so I’m wondering if this is a new thing, after all. (But they also have been a little higher in price–more like the $300-$500 range.)
With modern memory card sizes and write speeds, there’s almost no reason to shoot in anything but the largest image size/lowest compression mode.
And you can use hacked firmware to get RAW output out of pretty much every Canon P&S on the market for the past 5 years - CHDK Wiki | Fandom
digital sensors now have such good noise performance that pictures taken at ISO 1600+ on prosumer-enthusiast cameras can still be used in a professional setting, the equivalent in film would be to use either a high speed (high ISO) film and risk the grain or a lower ISO film and push the exposure up
No, but I wish I had an option for storing 1000 images at lower compression, or none at all.
Honestly, it’s easiest to do so in the computer once you download the images. A simple batch job can do so in most of the commonly used tools. Most consumers wouldn’t care about the feature and most advanced users wouldn’t use it. If you want no compression you probably want RAW mode anyway.
I’ve always assumed ISO settings on digital cameras were mostly a courtesy add-on for the folks who were migrating from film.
I do hope we’re reaching the end of the marketing emphasis on megapixels. At this point, only someone who wants to print gigantic posters needs more pixels than my Canon (EOS5D) provides. Souped up low light functionality and better registration in high contrast situations is a whole 'nuther ballgame. What I would like to see, and probably never will, alas, is some industry wide key or simple rating scheme for comparing sensors from one camera, or camera system, to the next, at a glance.
The holy grail of improvements, for me at least, would be a stabilization system that lets me stop the aperture waaaay down and still take pix without a tripod!
You have that in a lot of lenses, where you can go about 2-3 stops slower than what you can hand-hold at. And I do seem to remember there was at least one camera system that had camera-based vs lens-based stabilization, but I’m blanking on that now. Do keep in mind though that this doesn’t help when your subject is moving, though. (Well, to be exact, it can help in panning types of situations where the subject is moving consistently in one direction.)
It’s anecdotal, but my Canon 28-105 USM f3.5-4.5 lens was roughly equivalent to my 50mm 1.8 lens wide open (was slightly slower/darker) when taking pictures in and around St. Peter’s in Rome at dusk and just after.
So 2 stops for stabilization is just about right- 3.5 to 1.8 is somewhere near 2 stops.
Image stabilization is mainly useful if your subject isn’t moving. Fancy IS allows you to keep the camera steady so you keep the shutter open longer. If you’re shooting a moving subject you need high ISO to give you fast shutter speeds, not a longer steady exposure.