PDA

View Full Version : Took 6 years - new sensor demos released



KenCooper
01-27-2016, 05:27 AM
Here is one paragraph taken from the article >

...

We got a chance to see the QuantumFilm technology in action at a private demo, plus learn a little more about how it works. InVisage’s QuantumFilm sensor replaces the traditional silicon CMOS sensor used in most cameras today, including those in phones. It’s thinner (0.5 micron versus the 3 micron of a conventional back-illuminated CMOS sensor), and manages light more effectively. InVisage claims it absorbs 100 percent of light, versus silicon’s 70 percent, or less “cross talk” or light leakage. It results in a higher dynamic range image that’s closer to film, with greater details, and it does everything without any sneaky software tweaks.

...

http://www.digitaltrends.com/mobile/invisage-quantumfilm-camera-news/

kc

rick.lang
01-27-2016, 09:32 AM
Wikipedia describes this as a layer of quantum dot crystals that sit atop a sensor, so the phrase “replace the sensor” may not be accurate. The reviews I read are praising something based on the company’s promotional maryerials but the company doesn’t seem to be very transparent about how it demonstrates the claims or substantiates some pretty silly claims. Implying 100% of light is captured with no spill or light leakage might be hyperbole. Stating that ever higher resolution sensors are not impacting dynamic range is ludicrous because they imply they/ve removed that barrier. Just not possible to imply there’s no inverse relationship at all between resolution and dynamic range. Everyone go back to sleep until this company decides to be as transparent as their quantum dots crystals.

razz16mm
01-27-2016, 01:31 PM
The quantum dot advantage is separating the light capture and voltage accumulation layer from the silicon layer.
Characteristics of quantum dots includes:
1. color transmission determined by dot size, not color dyes. No loss of light transmission
2. In a practical sense infinite resolution. Pixel size is not defined by the quantum film layer.
3. 100% fill factor. Compared to about 80% for CCD and less than 30% for typcial CMOS designs.
4. High conversion efficiency. High sensitivity, high dynamic range.

The silicon layer design defines pixel size but this can be dynamic. You can have very high resolution at the expense of higher read noise, or aggregate many small pixels into large ones for lower noise. This is a dynamic function. Whatever software defined pixel matrix is programmed it still sees the full 100% quantum film layer, no scaling involved in resolution changes. Since the silcon is just manipulating electrical charges from the quantum layer traditional silicon sensor pixel performance limits don't apply. So a tiny 4k cell phone camera sensor is potentially capable of matching current gen S35 silicon sensor performance for DR with global shutter and color gamut defined prismatically like a 3 chip camera.

DPStewart
01-27-2016, 01:40 PM
Science!......and stuff.

rick.lang
01-27-2016, 02:02 PM
Razz16man, calmly separate their marketing babble from reality: "practical sense infinite resolution is gibberish. And if "you can have very high resolution at the expense of higher read noise," ergo, dynamic range is inversely related to resolution. This was news years ago, in a few more years it may be news again. Once Sony has all the rights!

rick.lang
01-27-2016, 02:03 PM
Science!......and stuff.

Science and science fiction.

combatentropy
01-27-2016, 09:19 PM
I'm most excited by global shutter, finally, again. I've been following this for five years. I started a thread about it on the other forum (http://www.dvxuser.com/V6/showthread.php?225559-Quantum-film-start-up-promises-global-shutter/).

The only sensor model so far is the Quantum13, announced last November. With 13 megapixels, 1.1 micron each, it's probably around the size of 8mm film, between a 1/4 and 1/3 video sensor. While I'll gladly take global shutter in my smartphone, I'm waiting for them to expand to DSLRs. According to the article at the beginning of this thread, "InVisage said two of the three major DSLR camera manufacturers have also chosen to use the QuantumFilm sensor in future hardware." That's exciting.

In each pixel of a traditional sensor, some circuitry takes up real estate. The photosensitive surface can't take up the whole area. That's another way of saying that the fill factor is less than 100%. In a QuantumFilm sensor, however, it still has the circuitry, but the photosensitive surface is spray-painted on top of it all. That's why it's accurate to say that its fill factor is 100%.

17025

I don't know about infinite resolution. Invisage says that it can group pixels into larger ones (http://www.invisage.com/technology/), though.

rick.lang
01-27-2016, 11:45 PM
Thanks for that diagram. With a CFA on top, you have the definition of a photosite in terms of the area of the CFA's individual filters and you have photosites also defined by the matrix of electronics below the quantum dot 'film' layer. Apologies if I was a bit testy in my earlier posts, but don't imply to an amateur cosmologist about infinity contained in a smartphone's camera sensor. When you have photons of light passing through a CFA and interacting with molecular structures and the results detected by circuits, you do have a real number of photosites that determine your resolution.

And the selling point of 100% light gathering is yet again an oversimplification as I doubt 100% of light reaches the target. So okay, now we have a photosite 1.1 microns and okay, I'll give you your 100% transmission for your miniscule photosite. But even 30% of a large 6.5 micron square photosite like the BMCC sensor, is an awful lot more light! 35 times more surface area and at least 10x more light.