Hold on, what about 2 bits, 4 bits, 6 bits a dollar . . . ?
Rich
Hold on, what about 2 bits, 4 bits, 6 bits a dollar . . . ?
Rich
I started with programming 16-bit systems in the Seventies. I am still comfortable with 64-bit. No big deal.Originally Posted by Corran
But Jac, 8-bit, 16-bit, or 24-bit? Or perhaps 32-bit floating point??
I say the above after having shot digital for almost 14 years now. Before that roll film and medium format once. Even though I learned on film, I never took the time to understand film and light. It was all by trial and error. Going back to film after 14 years digital, I can comfortable say film is a lot more technical and even though both digital and film photographers should understand light, etc, digital makes it easy to not understand. With a little bit of reading, and understanding what the histogram tells you, you can rattle of a half dozen shots and one will be good exposure wise. Then go to photoshop and further "fix" mistakes or oversights due to a lack of basics and unless your totally not artistic in the least, you can get a workable image. Therefore, I say digital is a whole lot easier than film.
Where digital is forgiving 99.9% of the time, film is not. You have to get everything right the first time, 2nd at most. Even if your not wanting to learn that much the economics of it all forces you to learn all you can. Plus, film still has better resolution. As an example on my 4x5 with a 75mm lens I get an 80.5 degree field of view on the long side, for 36mm sensor I get 27 degree field of view.
As you know, to get the same field of view requires a 21mm lens, but at the expense of resolution and more distortion to deal with post. Even with a tilt shift lens, which I have 2 of for my 5DMKIII, the 17mm and the 45mm, now three new tilt/shifts to get, but I digress, you get only a limited subset of movements, and they are a little more awkward to use, but same principal.
The 4x5 has total control and you can do things with it that are almost impossible to do with a full frame digital or almost any other digital. But even doing so, you are left with a much smaller field of view for a given focal length, to keep resolution leading to panoramic shooting, or sacrifice resolution for getting the same field of view. Don't get me wrong, I love digital, but it is a tool like any other tool. Right place, right time.
I'm impressed with all the workarounds that digital astrophotographers now use. For just one example, thermal noise can be dealt with fairly easily by taking a series of dark frames that record the noise with the lens cap on, at the same settings as the 'light frames'. The noise can then be subtracted using software.
People are taking fantastic images even from highly light-polluted areas by using filters that block all but a very narrow set of color wavelengths given off by objects in the deep sky, thus blocking the pollution. This necessitates extremely long total exposure over, often, many nights. Hours and hours of exposure.
Stitching together digital image 'tiles' has also taken care of the resolution edge that film has. I've done some of these things, but am not at all skilled.
DSLRs also do that, even entry level, for example Nikon D3200 does it, just set Shooting Menu -> Noise reduction option to ON, it takes a second exposure of same duration with the shutter closed and firmware substracs that dark shot. It is done in camera.
https://forums.cameratips.com/discus...me-photography
Right, but some of us do it out of sequence in order to take advantage of getting photons on the sensor instead of having to wait through equal amounts of light and dark while shooting. The ambient temperature of the surroundings can effect things, though, so if the temperature is falling as the night goes by, as it typically does, you cannot wait too long to take your dark frames. You also do not necessarily need a dark frame for every light one. It's been awhile, but I would often have the camera take dark frames while I was slewing my setup to a new target. If I had taken, say 20 two-minute light exposures, I might do eight darks with the lens cap on.
Some folks do not even do this anymore because newer cameras have gotten so good at controlling noise in many ways. Higher ISOs can be used, so shorter exposures are possible, so thermal noise (which is already much lower anyway) is avoided.
Here's an image I made a few years ago, of M27, the Dumbbell nebula. It's a combination just like I described: about 20 30-second exposures through a 10-inch f/4.7 Newtonian reflector telescope, and ten or so dark frames at the same settings, combined in a freeware program called DeepSkyStacker. It is nothing like as nice as others out there, but I like it because I did it myself, on my driveway, through a telescope I built in my garage.
And, to keep this post on-topic, here is a film image. If I recall correctly, it was 20 minutes on Fuji Provia 400f through a Nikon F with a 50mm lens, 'piggy-backed' on an inexpensive telescope that had motorized tracking to counteract the spinning of the earth.
Each was challenging in its own way, but the film version was not entirely done by me, since I did not develop the slide myself. The waiting for the film to be returned, while hoping nothing went wrong with the tracking, was part of the 'thrill', if you want to call it that. Notice the definite magenta cast to the image. This is after trying to tone it down digitally. The biggest challenge for the second shot was twenty minutes of tracking, using an eyepiece in the telescope with cross-hairs to keep things perfectly centered.
There is something about astrophotos on film that I like better than digital. At least on the ones I've taken myself. The stars themselves seem more real, for one thing. On the higher resolution version of the film image, I can zoom way in and still see distinct pinpoint stars, very dim, but there. Digital images (and not just mine) seem to often overblow brighter stars before registering the 'tiny' dim ones (which are not actually tinier than bright stars, since all stars are just points of light through even the biggest scopes; these seemingly smaller stars do not look nearly as much that way except in photos.)
M27, you were capturing light born some 1360 years ago, the roman empire fell two centuries before, Visigothic Kingdom was rulling where I'm living... then those photons you catched started a trip from the all nebula places to your camera...Here's an image I made a few years ago, of M27, the Dumbbell nebula. It's a combination just like I described: about 20 30-second exposures through a 10-inch f/4.7 Newtonian reflector telescope, and ten or so dark frames at the same settings, combined in a freeware program called DeepSkyStacker. It is nothing like as nice as others out there, but I like it because I did it myself, on my driveway, through a telescope I built in my garage.
And, to keep this post on-topic, here is a film image. If I recall correctly, it was 20 minutes on Fuji Provia 400f through a Nikon F with a 50mm lens, 'piggy-backed' on an inexpensive telescope that had motorized tracking to counteract the spinning of the earth.
Each was challenging in its own way, but the film version was not entirely done by me, since I did not develop the slide myself. The waiting for the film to be returned, while hoping nothing went wrong with the tracking, was part of the 'thrill', if you want to call it that. Notice the definite magenta cast to the image. This is after trying to tone it down digitally. The biggest challenge for the second shot was twenty minutes of tracking, using an eyepiece in the telescope with cross-hairs to keep things perfectly centered.
There is something about astrophotos on film that I like better than digital. At least on the ones I've taken myself. The stars themselves seem more real, for one thing. On the higher resolution version of the film image, I can zoom way in and still see distinct pinpoint stars, very dim, but there. Digital images (and not just mine) seem to often overblow brighter stars before registering the 'tiny' dim ones (which are not actually tinier than bright stars, since all stars are just points of light through even the biggest scopes; these seemingly smaller stars do not look nearly as much that way except in photos.)
...just entering through the 10" hole of the 1200mm...
M27, you were capturing light born some 1360 years ago, the roman empire fell two centuries before, Visigothic Kingdom was rulling where I'm living... then those photons you catched started a trip from the all nebula places to your camera...
...just entering through the 10" hole of the 1200mm...
There are a number of points regarding digital noise which are being conflated here.
If you add multiple images together, the light-generated signal, which is (nominally) constant increases linearly; the random electronic noise (thermal and shot noise) are added as the sum of their square root values - so when the final image is scaled, the noise is reduced with respect to the desired light. This increases the signal to noise ratio (higher is better) by 3dB for each doubling of the number of images being stacked.
Sensor noise - pixels with a fixed non-zero dark value - require a black frame which can be subtracted from the finished image since they will act as point souces. Since they are also subject to thermal noise, you should also use multiple dark frames to average that noise down.
Neither of these are issues with an emulsion (though it has other issues which I discussed earlier).
Neil
Posting Permissions
Reply With Quote
Bookmarks