If you were ever curious about the detailed specs of that so called “translucent mirror” used for SLT cameras than just take a look a Patent Number 8313254 (via freepatentsonline). It explains the material used and the reflective characteristics of the mirror: “the light-beam splitting mirror has spectral characteristics including a reflectivity of 25% or more and 35% or less at all wavelengths from 400 to 650 nm and a reflectivity of 60% or more at a wavelength of 700 nm”
Have fun reading all these tech!
34 Responses
This is some really amazing scientific stuff. Yawn.
You must be lost, child. The Jersey Shore message boards are that’a’way >>>>>
You must have no sense of humor. The MENSA wann-a-be trolling boards are this way ^
Sounds like the mirror is specially tuned so that the reflectivity of infrared is higher (60%), and reflectivity of visible light is lower and flatter (25%). So, if the AF sensor is sensitive to infrared, AF capability is enhanced without (by sending the AF sensor more than half of the infrared light) losing too much light needed for the sensor (by sending the imaging sensor almost all the visible light). Clever. This is consistent with the report that the translucent mirror only loses half a stop, while maintaining pretty darn good AF sensitivity.
This behavior is achieved by alternating layers of high refraction and low refraction materials, rather than a reflective material like silver, so the reflection and transmission is uniform rather than “spotted”, like it is with half-silvered mirrors.
I’m not entirely sure about the physics of this, but this might result in some ghosting, which might account for some of the reports of blurring caused by the half-mirror. So, it’s the multiple layers of reflective surfaces, rather than the PET plastic substrate, that causes blurring?
I’d imagine the purpose of multiple layers is to minimise the amount of total internal reflection that goes on, similar to the purpose of multilayer coatings on lenses. Ghosting is always going to be an issue with a semireflective mirror, but they way they’ve done it will avoid anything defined.
The silica layer is, naturally, optical glass, and the niobium pentoxide layer is the mirror. I expect the choice of niobium rather than silver or aluminium is due to niobium having a favourably non-linear reflectivity.
If indeed that is true, and that the AF sensor is more tuned to infrared wavelengths (as implied) than visible wavelengths, then we can expect the Sony AF system to be more responsive to heat maps/thermal images (eg edges between a human and the air surrounding the human).
Thus, keeping in mind about phase-detection AF, most obviously this is great for sports, wildlife, & portraiture.
On non-organic objects then the amount of thermal radiation depends more on the color of the object, white (yellows, greens, or colors with a red component) would reflect more infrared rays from the sun (or a light source) vs black (or any colors having blue components). So in this sense, if there’s a strong difference between colors, then the Infrared profile of the colors will show a strong contrast (think of it converted into black and white).
Smart move by Sony to redirect more of the wavelengths that are otherwise already filtered out before the main sensor!
It’s what is called a “multi-layer thin film dichroic” of the long-pass, wide-band, partially reflective variety. With enough layers, they can be designed to control the reflectance and transmittance across various bandpasses. A recent design I worked on has 17 long-pass filters (which reflect wavelengths A-B and transmit wavelengths B-C) , 2 polarizers , and 16 different narrow bandpass filters.
So, this is common technology. But I don’t know what the patent claim here is.
“a reflectivity of 60% or more at a wavelength of 700 nm”, so the red color will be affected?
No, it just reflects IR so it wont reach sensor (just AF). Red is under 700nm, beyond is IR. Pretty much it works like 60% IR filter.
> so the red color will be affected?
Yes, and this should cause higher chroma noise
The red shouldn’t be affected and it most definitely won’t cause higher chroma noise.
The hotmirror in front of the sensor cuts out that part of the spectrum already.
Transmittance for the hotmirror is only around 20% at 650nm and around 1-2% at 700%
(http://www.astrodon.com/custom/_2e2a/content/images/AstrodonInsideSpectra.jpg)
Currently, PET alone. You can see the pictures.
Would be cool to make AF purely infrared and reduce visible light loss to something like 5% …
It’d be all fun and games until you tried to focus on something that wasn’t very reflective in the infrared spectrum.
IR focus lamp could be handy.
Yawn… an SLT mirror patent. No competitor in their right mind would want to use an SLT mirror anyway, so that patent is irrelevant. Sony should invest the time and money needed to file these patents in R&D to get rid of the SLT mirror.
Well, how else can Sony compete with Nikon and Canon besides using the SLT mirror and looking unique?
Sony certainly can’t go head to head with Nikon/Canon on their own turf. With regards to autofocus capabilities and moving mirror technology, Sony is far behind. Sony’s OVF display technology is also outdated – even the Nikon D3000 has more comprehensive viewfinder info displays than the a900, although the D3000 finder is a little cramped. Sony DSLRs also have never been known for exceptional build quality or weather sealing. Nikon/Canon pro cameras, along with high end Pentax cameras, dominate in that area.
SLT technology doesn’t work for everyone (including me – that’s I have a Nikon D600 now), but it’s really the only way Sony can compete. Sony can’t take Nikon/Canon head on in a DSLR tech battle, so they’re offering something different.
Uhm, some facts here:
the A900 OVF is known to be one of the best and brightest OVF.
The A900 was the only camera not making any problems on a trip of 30 photographers to the Antartica using all kind of professional equipment, the Canon and Nikon equipment had problems after some hours with the freezing temperature.
Sony is far superior in one non-DSLR-market and that is mirrorless cameras.
So some of your sayings just are not right ;)…
Pentax K-5 too
True, A900 was a great camera but the sales didn’t quite agree with that. But with A55, sales was better than what was expected. Sony needs money after all(actually quite desperately).
A900 has already become a collectors item, get while you can.
The A99 hasn’t sold great so far, either.
@Carl- That’s why they changed their full frame roadmap. FF SLT won’t sell well. Well, maybe whatever type of full frame camera, by Sony, will never sell no matter what.
The a900 finder is one of the biggest and brightest, but I’d appreciate it if you’d read my post, because what I’m saying is correct. I was criticizing Sony’s information display in the finder, which is really bare bones compared to even entry level DSLRs from other manufacturers. It’s clear that Sony really put zero effort into improving the information display.
The a900 OVF lacks a full time ISO display, and can’t show the number of frames remaining. Compare this to any mid-range or higher DSLR from any other manufacturer. The Nikon D300, for example, has a full time ISO display along with a remaining frames display in the finder. In addition, it has on-demand grid lines (no need to carry extra focusing screens), focus direction indicators, and warning icons. The Canon 7D is similar.
Sony’s OVF info display tech really dates back to 2002, and matches cameras like the Canon D60 (not 60D, the old D60). Canon has advanced the OVF info display since then. Sony hasn’t.
Yep… SLT and EVF sucks. That damn EVF is annoying with color shift and blur. I have to keep my eye dead steady to look through it. Ticks me off. Also, the lag time from view to capture sucks. I much prefer an optical through the lens viewfinder.
That’s where Sony screwed the pooch with my NEX 7. Wish it had phase detect focus, too. Otherwise, great camera. Oh, one more thing…
What the hell is up with it’s crappy jaggy lines and weird black dots along sharp highlight edges?
Interesting – I wonder if this could be retroffited to the A77 / A99. The mirror itself could be exchanged quite easily, however I doubt that the AF sensor is sensitive to that particular wavelength. But it sounds very promissing in actually turning many disadvantages of the SLT design into a major advantage in AF and IQ.
I’m hoping that on-chip focussing sensors will end the need for a transclucent mirror.
I think CaNikon hope that doesn’t come too soon.
You buy the best glass you can afford only to pass all your light through this layered filter.
The gain in AF is just not worth the loss to any company other than Sony.
You can say the same thing about AA filter.
“You could say the same about the AA filter”
That’s why some photographers are getting cameras with no AA filter, or prefer cameras with weaker AA filters. They care a lot about image quality.
just came runner up in the November thatnikonguy furry Friends competition on youtube with photo taking on my a77 works for me !!
So it uses IR to AF…
Since shift between visible and IR can be easily calculated I guess you might be right. 🙂
And with AF in video (anybody use it? I didnt) we got huge troubles with shooting in opposite light – this A77
http://photo.qip.ru/users/ozerovpaparacci/4087643/101188737/
http://photo.qip.ru/users/ozerovpaparacci/4087643/101188788/
This 99 –
http://photo.qip.ru/users/ozerovpaparacci/4158628/101188715/
http://photo.qip.ru/users/ozerovpaparacci/4158628/101188722/
And then brighter light source, then longer that effect