ProfHankD

Thorium-doped lens elements (up to 25% Thorium or more) can have sufficient quantities of daughter products ("impurities") so that it isn't just the Alphas coming directly from Thorium-232 decay. There are several Beta emitters in the thorium series. I regularly use a few lenses with radioactive elements because they have unique optical properties (e.g., Takumar 50mm f/1.4), but I handle them with some respect. One idiot "fixed" radioactive yellowing using a sledge hammer, http://web.aanet.com.au/bayling/repair.html -- THAT is truly dangerous, because Thorium particles could be inhaled, etc.

It's not a coating, but a major component of the glass itself. Also, it is worth noting that it isn't just Thorium (and its often hotter daughters) in some old military optics, but any of various heavy elements including depleted Uranium. Dangerous? Think red fiestaware or living in Denver. In other words, much higher than average background levels close-up, but not particularly dangerous under normal circumstances.

GIMP actually is more "powerful" than Photoshop, but it is 8-bit/color channel before 3.0 and the user interfaces on some plugins feel like research prototypes (often, because they actually implement new research). CinePaint, which was a GIMP fork, has done 32-bit floats/color channel forever (including native HDR) -- it was developed largely for touch-up of digital masters of cinematic releases. Ok, it currently takes multiple FREE tools to blow away one expensive, soon-to-be-monthly-licensed, Photoshop. Not much of a contest as I see it....

A lot of folks use 3D printing for camera parts. In fact, my MakerGear M2 is primarily for building custom camera parts, and I own & use over 100 non-native-mount lenses. I can't believe anyone would take this hack seriously. It's cute in a "steampunk" sort of way, but photographically a bad joke.

Incidentally, the adapters for getting shallow DoF on camcorders work by photographing a ground glass image... but they keep the glass moving so that the surface roughness of the glass "averages out." A speaker driver could probably vibrate it enough in a slotted holder; some adapters simply spin the glass.

One word: GIMP -- http://www.gimp.org/
Ok, maybe one more: CinePaint -- http://www.cinepaint.org/
Both are fully free and work very well.

The discussion here seems to have been largely features of Nikon/DSLR vs. Sony/mirrorless -- which isn't my point.

The issue is that the DPReview summary scores say NEX-7 IQ is markedly inferior although the test results in the reviews do not support that conclusion. This type of bias casts doubt upon all DPReview's normally fine work.

It will be interesting to see what the review of the 7N says. ;-)

Interesting. Go to the studio scene comparison tool. Add the NEX-7. Compare. Up to ISO 6400, the NEX-7 (out since 2011) clearly wins on both raw and JPEG. Above, the D7100 may be a bit better, although both look nasty and the NEX-7 seems to have both more noise and more detail.

Here's the point: the NEX-7 raw and JPEG image quality scores in the summary at the end of the review are POORER than those of the D7100! Huh?

The images are pretty clean, but not very crisp even at low ISO with the lens at f/8-11. Ok, I'm spoiled by using a Sony NEX-7 for the past year, which is generally crisper per pixel than Canon's latest FF cameras, but the IQ here certainly isn't compelling against the latest crop of mirrorless cameras in the same price range... which is presumably what this "mini" DSLR is trying to compete with (since the EOS-M hasn't done very well).

Actually, the images here look very processed to me... I wonder what flaws Canon is repairing in processing? Perhaps it is deliberate product differentiation, ensuring a per-pixel IQ gap between this and higher models?

Really a pretty nice ap -- the movement between shots seems to be handled fairly smoothly, which is the hardest part. With the huge DoF of an iPhone, I'm sure there is a little problem with doing this for more normal focus distances, but overall a very nice job.

Using the image comparison tool, I think this is notably below the NEX-6 in still IQ -- for JPEG more than for raw. Not really a surprise... except it is slightly worse than I expected (a hair worse than the EM-5) and the review keeps saying how great the JPEGs are.

I suppose if you like a DSLR form factor for shooting video, this is a great option. However, the level of remote control looks more interesting to me... provided that they provide a programming API, not just some magic app. Is there an API for wireless control by a program?

Very retro, very Minox. There are apparently wide and tele converters, a side-mounted flash, etc. See: http://www.minox.com/index.php?id=8523&L=1

Of course, I don't see any sample images. Then again, the old Minox cameras were not very impressive in terms of IQ anyway. Actually, Minox doesn't even show side, top, or rear views of the camera. They do quote dimensions: 82 mm(L) x 67 mm(W) x 46 mm(H); for comparison, the much more capable Pentax Q is 98x57.5x31mm -- about 2/3 the volume of the Minox.

In sum, it's a conversation piece that can take snapshots....

No, no. The idea was to get rid of the salt. ;-)
(Oops... this was supposed to go on Dutta's comment below.)

They didn't just throw the lens in boiling water. They boiled some of the parts in a chemical cleaning bath using water-soluble chemicals. Not really all that new an idea, neither is it all that scary; dissolving the sea salt deposits is pretty much the only way to get rid of them... and I hope they changed the bath & rinsed the parts well to remove all traces of the salt.

For both these "cameras," the ratio of case material to internal camera structure is roughly an order of magnitude off norms and maybe 3X the closest Sony NEX. I assume they're powered by radio waves because there's no space for batteries. ;-)

Anyway, the not-affiliated-with-Leica X3 seems like a violation of the Leica trademark and if I were Leica, the letter demanding damages would be in the mail by now.

For what it's worth, the nanocontroller architecture I've been working on for a decade has less logic than a 4-bit counter (see aggregate.org/KYARCH/20070914/), but one needs more than 4 bits local to each pixel to do this right and memory cells are somewhat problematic. My reference sensel size is 5um, and between 100-300 transistors fit under that. With the smaller sensel pitch they seem to be targeting, I doubt they're really doing much more than simple threshold output & reset with per-sensel logic.

This has some of the properties of a new sensor technology I've been working on in my research for several years, and I think it's a great general approach.

However, they don't give quite enough info to distinguish what they're doing from the sensors that have been made primarily for the auto industry to use in rear-view cameras. You may have noticed that some of those have far greater dynamic range than normal cameras, and this type of sampling logic is the reason. I don't know who's making which versions now; a little company making such sensors was acquired by Cypress Semiconductor some years ago and they were producing these sensors, but I believe Cypress stopped doing that in the past year or two....

Optically reducing focal length (and coverage) theoretically should work better and it pretty clearly does. I've tested 8 teleconverters with various base lenses and not one delivers IQ comparable to the base lens wide open -- which the SB apparently often does. In case you're wondering, the best teleconverters I've tested were a 6-element Tamron SP 2X 01F and the Vivitar macro focusing 2X, both of which are well respected and do deliver quite usable IQ when paired with an appropriate base lens (stopped down a couple of stops).

I think what he meant was "Casio doesn't make lenses." ;-)

Actually, I still have fond memories of my QV100. It had no optical finder, but live view and a pivoting lens. Let's hope Casio has some similarly clever (well, they were then) ideas brewing now....

Nikon and Canon have not been doing serious mirrorless yet, and it isn't feasible on DSLRs. As for Leica, they've traditionally gone more with "convertible" lenses, not converters... which makes more sense when you think about the mechanical complexity of rangefinder coupling. Those three are also very conservative (slow to innovate) companies.

Historically, Minolta or maybe Pentax would have been the most likely to innovate in this way. However, Pentax doesn't have an appropriate mirrorless and Sony is probably busy designing native E lenses.

In any case, there's nothing "dubious" about this -- it makes much more sense than (and gives better IQ than) a teleconverter, and both Nikon and Canon make those. The truth is that the folks behind the Metabones converter have excellent credentials as lens designers and they did something really smart. I'd bet that a lot of camera companies are taking a hard look at the patents now to figure-out how they can make their own versions....

I actually have been looking at producing a glassless autofocus adapter for manual lenses on E-mount, initially as a university research project and then as a product. Getting interface design constraints to enable fast AF is critical to this. Let's just say that Sony thus far has not been very helpful. :-(