Hello all. Long-time visual observer and absolute rock-bottom beginner for DSLR photography / photometery here. I have a simple question. Do any of you have suggestions for a relatively inexpensive camera that could be used for both nature photography and eventually for photometry? There seems to be a bewildering amount of choices out there. Thanks in advance!
Paul Zeller (ZPA)
Indianapolis, IN, USA
I am not at all sorry about my choice of a Canon EOS Rebel, which is a mid-range camera. I use it because my "good" camera is also a Canon; there probably is no significant difference between Canon and Nikon equipment. That said, you should look carefully consider all the possible applications you will have before buying anything, because, for example, some lenses don't work with all the cameras in a given brand's line.
Whatever you do, do not fall for the more pixels is better mystique. While it is true that full frame cameras give you a proportionately larger FOV (for a given focal length) than do ones with APS-C sensors, you pay for it in terms of processing time. This is not necessarily so trivial as simple CPU time; for instance Canon changed the format of their raw file format on some of their high end cameras, which means that even general purpose photo processing software like Photoshop and GIMP require the files to be converted from .CR2 to .DNG before they will process them. If that were not enough, you might find it necessary to upgrade your computer if using a high end full frame camera.
It is not necessary to buy the very latest of a given model; in fact aside from saving money by purchasing last year's camera body (or the year before's) you may find that it is easier to find compatible control software.
So, to summarize: Buy a quality DSLR that uses an APS-C focal plane, has manual mode and the ability to save raw image files.
Minus a few detours and dead-ends, this was my path to DSLR happiness which I can recommend to beginners: Buy an up-to-date, entry level DSLR kit, and put a fast, bargain priced, used legacy lens on it.
Sensors improved quite a bit during the last (say) 10 year, so when faced with the choice of buying a used high end DSLR from a few years ago, or alternatively buy a new, entry level DSLR for the same price, I'd always go for the newer model. The high end cameras mostly have additional features that you won't need for photometry anyway.
It's kind of the reversed situation with lenses: on ebay etc. you can get very good and "fast" lenses from the chemical-film era at a fraction of the price you would pay for a similarly fast modern lens. Not all legacy lenses can be adapted (with a simple adaptor ring) to every DSLR body, but the Canon cameras for example can be adapted to Pentax "K-mount" lenses, for which there is a decent supply of old lenses on the market, (e.g. you can sometimes find an smc Pentax 135mm f/2.5 for around 100 bucks if lucky).
You can then use the DSLR and a fast lens (say up to 200mm focal length) even on a static tripod and then do very solid photometry on bright objects (down to 8ish magnitude).
Of course you might still want to buy one of those special offer kits that consist of a camera body plus a zoom lens, and use the zoom lens for nature photography, but those zoom lenses are not optimal for photometry (aperture / focal length ratios is 1:4 to 1:6 usually, where you want to have something more like 1:2)
And if you want to take it further you can then put the DSLR on a driven mount, or eventually use it on a telescope on an equatorial mount.
Fully support Heinz-Bernd's recommendation ! I would just note that there are problems with the new RAW image file coding from Canon, CR2 moving to CR3. Most of our softwares (maybe 100% !) working with EOS RAW image files use dcraw.exe from Dave Coffin for the CR2 decoding. The new CR3 file type from Canon includes new technique for compressing the RAW images. Apparently Dave has not transitionned to that new file type. It's a serious problem for us. Make sure the new camera you are looking for is not CR3, we should also make sure any firmware upgrade doesn't bring the RAW output CR3 !
If someone has news about that I am interested.
I would also mention there are many old Nikkor F lens, very good, at low price, that are easy to use on EOS camera (many adaptator rings available from China for a few box).
Clear Skies !
Thanks for that info Roger, I wasn't aware of that.
It seems that the latest version of Adobe's DNG image converter can handle converting CR3 files to DNG format (which can then be read by DCRAW-based software), potentially with a bit of loss of meta-data (bug?) https://forums.adobe.com/thread/2507182
I would like to add a detail ...
If you control your DSLR by software like Maxim DL or APT, your scientific file will be of the "fits" type. Which resolve a lot of problem for the treatements. Maxim can even separate the RGB spectra!
But that involves having a computer or a laptop to wear.
I don't think so. Maybe I am wrong but I don't see any way to extract RAW image data from a DSLR. The only available from the existing EOS firmware is the CR2 (now CR3) compressed format. To access the true RAW data you should make change to the DSLR firmware itself. The most probable is that Maxim DL converts the CR2 format into computer RAW, then FITS, very probably integrating the dcraw code in its own code. We all do that, I do it in my own software. Dcraw is distributed by Dave only as a C source code, by the way many link it to their own source code and compile the whole.
Heinz-Bernd, I know this idea to convert CR3 to DNG but I had problems to make it work (I made no effort to...). Making such cascade conversions is not something I would do. Such conversion is what takes far the most of our processing time.
Clear Skies !
I have direct experience using the Adobe DNG converter. I bought a 5DS R and found to my immediate chagrin that its ".CR2" files did not open in Photoshop or GIMP. A visit in panic to the Canon web site led me to Adobe, and now I am able to convert to .DNG that pretty much everything can use. I guess Canon finally cleared up the ambiguity by relabeling the file format "CR3".
Adobe apparently never got past drag and drop, because the only way I have found to make the DNG converter work is to select the files and drag and drop them on the DNG converter icon on my desktop. (Windows 7 Pro 64bit w/ SP1). It works fine then, taking less than 10 seconds to convert a 52 Mpx file (4th Gen Core 7 processor). Hence my comment in my original post.
I also got some problem with recent camera CR2 files. Was apparently a problem of compatibility with Windows 7 , now with Windows 10 all work properly with recent CR2 camera. There was no change in the way data were compressed (a classical Huffman technique).
Then CR3 is very different, Canon clearly said they have developped a new compression technique replacing the Huffman one. It provides a 30% compression improvement. They said this was possible thanks to their new Digic 8 processor. From the info available here from Canon (different in US ?) the only camera involved at time being is the EOS M50, but the move to CR3 should rapidely extend, maybe through firmware update.
We have no info on this new compression technique, maybe proprietary...
Of course we can ...
I attached a "fits header" to this post to prove it. I've been doing it for some years ...
And just to be sure, tonight, I redo the test and send a print screen.
Maxim DL, (with the good driver), takes control of the CanonEOS II and the images go directly into the computer without storage on the camera and it records it in fits on the hard drive, as with a CCD camera. Just converting RAW images in fit type. After, you can do a "Extract Bayer Plane" of 2 and 3 to separate the green spectre and treat it. Or if you take color pictures, you can do a "Split Trocolor" to separate the green. (But red and blue can serve too... Ask to Mark Blackford!)
If the images first pass through the camera and are stored ... There, there is formatting in CR2 or CR3.
As reference, you can read the "AAVSO_DSLR_Observing_Manual_V1.4" (https://aavso.org/dslr-observing-manual) at the chapter 3.1.3 "Extraction of individual color channels."
Yeah it's always the same...<sarcasm on>hopefully the previous camera hardware is too slow or something to support the new compression thingy so that even firmaware updates won't migrate old hardware to it, that would be consistent with the comments you quoted.<sarcasm off>
At the time I got my EOS, the CR2s of it would not yet work in DCRAW, so I went to DNG out of necessity, but ever since then I just use it routinely because I like the idea of a more portable format for archiving. I never had a problem with it's batch mode: just specify a folder and let it convert all files in it while I'm having a coffee. Or two.
Heinz-Bernd, what type of recent EOS/CR2 is incompatible with dcraw, I am curious ? I also got some problem with my EOS M3, but minimal, I could correct it easily in my software (test/reference pixels wrongly selected). The very last 2000D / 4000D (in EU) are compatible.
Canon said the new CR3 was linked to the new Digic 8 processor (maybe some algorithm implemented in hardware ? we did that too in my past job...). Maybe that would avoid firmware update to older Digic ? Who knows...
>Heinz-Bernd, what type of recent EOS/CR2 is incompatible with dcraw, I am curious ?
It must have been fixed many, many years ago, but when I got my (now not so recent), EOS 1100D (oh well...introduced 2011!!!), the CR2 files produced by it were not correctly processed by older DCRAW versions used by some software (it produced some image but not even the x-y dimensions were right). A truely historic issue, nothing that should worry people now.
Let me put my two cents in. During last months I’m developing my own set of utilities to help working with FITS files (to facilitate processing with IRIS). Utilities are for Windows 64 command-line, they also work under WINE/Linux. One of them is “iconvraw.exe” batch raw converter, it uses LibRaw code (https://www.libraw.org/, free library based on modified and optimized dcraw sources). You can convert many RAWs (including CR2 format) at once quickly with different options into set of FITS files (running iconvraw.exe without parameters prints help, see also M57.bat batch in sample directory to see how to convert a set of RAW files of a sessions into sets of offsets/dars/flats/lights FITSes). Unfortunately currently CR3 is not supported by Libraw, i will keep an eye on it and update the utility in case new Libraw version support CR3, I hope they will made it).
installation package and sources (Object Pascal, compiled with FreePascal plus LibRawWrapper compiled with MSVC free community edition) can be found here https://sourceforge.net/projects/fits-command-line-utilities/
Iconvraw.exe light_SS_Cyg\*.cr2 /g=light /o=c:\SKY2 /TS-10800 /$OBJECT="SS Cyg"
will convert all CR2 in “light_SS_Cyg” directory into numbered light*.fit files, place them into output directory c:\SKY2, timestamps from RAW EXIF will be shifted 3 hours backwards (/TS-10800 option, time shift minus 10800 seconds to correct timezone), a keyword OBJECT in FITS header will be updated to “SS Cyg” after conversion (multiply keywords could be used). The utility generates descriptive FITS header, supports different RAW formats (thanks to LibRaw) and works much quicker than DCRAW.
(I will make better description of utilities on SourceForge, I only would mention here that the set also includes configurable CFA->RGB converter so it is possible to work with IRIS avoiding its obsolete RAW converter and hence use modern RAW formats not supported by IRIS, rgb channel splitter, parser of AstroImageJ output files (fluxes to instrumental magnitudes), FITS header viewer/editor ...).
Realizing that this thread is now far afield, I nonetheless add two final comments; first that the Adobe DNG converter is available without charge, and secondly that the version of .CR2 used by the 5DR S is truly different from that used by for example the 3Ti (aka D600).