"To achieve optimal imaging performance, Sony’s special “analog memory” perfectly stores and shields the light collected in the photodiode. This technology minimizes common global shutter CMOS artifacts such as fixed pattern noise and dramatically lowers read noise levels to just 7 electrons (e-). The IMX174 sensor further boasts saturation (full well) capacity of 32,000 electrons (e-), dynamic range of 73 dB, and peak quantum efficiency of 76% at 525 nm."
Compare the others, FWC of Sony's sensor is extremely high,and noise is small. want to konw more detailes
ReplyDeleteWhat about parasitic light sensitivity ? How efficient are all these global shutters ?
ReplyDeleteit has been stated in a paper that PLS = -110dB.
ReplyDelete-yang ni
Excuse my ignorance but I thought that PLS is not measured in dB units. It is not that similar to smear. I have always believed (and maybe I am wrong) that PLS is defined as the ratio between the signal colected by the photosensitive node and the one in the storage node in the same amout of time. Anyhow is there any place where I can find of definition of this parameter?
DeleteSmear in CCD is caused by the parasite light sensitivity and directional blooming. You can always define this parasite signal as compared to the main signal by using dB. SONY uses always their own definition on smear, so they continue to use this in their CMOS sensor.
Delete-yang ni
I am not sure where the term "parasitic light sensitivity" came from (I am sure someone here knows) but it is certainly incorrect English. For shuttered pixels there is no "parasitic" light. There is light signal leakage. I always have preferred shutter "efficiency" as a metric. Looks like Sony's shutter efficiency is about 99.9997% which is very good. For constant illumination, only 3 electrons out of a million signal electrons are from leakage. At least that is what they say. In principle, the efficiency should be wavelength dependent. I wonder how they measure it?
Deleteyou could be a dictator :)- why every one has to use the term that you defined ? We can invent other terms, since we have nothing to invent now in pixel haha !!
Delete-yang ni
Yang Ni, you are funny. First, there is plenty to do in pixels but getting industry leaders to switch will be the challenge. Anyway, you have your own special pixel to be happy about. Second, I am just giving my technical opinion. Just because I don't like the words and units of PLS, doesn't mean you can't enjoy it. Using a "power" unit like dB (20 log A/B) always seems weird to me when we are just talking about the ratio of non-power counted things, like photoelectrons to noise electrons. But it just hangs on, year after year, whether it really makes sense or not. Kind of like miles and degrees F.
DeleteISSCC paper and commercial products are different stories. If it's -110dB than it's THE winner
ReplyDeleteI have some dark frames (and comments) of that IMX on my website: http://www.rkblog.rk.edu.pl/w/p/taking-quick-look-sony-imx174-cmos-sensor/ :) Looks good, although I didn't saw any real pics taken with it.
ReplyDeleteEric,
ReplyDeleteA few years back, "PLS" was most commonly found in datasheets and papers from FillFactory/Cypress/ONsemi, later CMOSIS joined the gang. These days it's proliferated quite a bit, but this gives some indication that the term probably came from a small region with 6 million inhabitants and a relatively high amount of CMOS imaging experts.
In this region people speak "Dutch" (although people may mistakingly call it "Flemish") ... in Dutch, "lichtgevoeligheid" means "sensitivity to light", so...in Dutch, "parasitaire lichtgevoeligheid" really means "parasitic sensitivity to light" ... so ... you're right, there is no parasitic light, but there is a parasitic sensitivity to it and that is what the people who coined the term intended.
This got literally translated and ended up in literature. Before you know, a decade later, the industry is picking up the lingo and it starts living its own life.
Of course, this is just a theory...
The problem with shutter efficiency is that it doesn't align with how we like to interpret it. The first thing you do after mentioning the 99.9997% is calculating how many electrons are erroneously collected in the storage area, because that is something we all naturally relate to. 3 out of a million...so you're realy translating shutter efficiency to "PLS" to make it into something one can grasp.
Seems a lot easier to calculate 3 out of a million from 99.9997% (or use shutter inefficiency, 1-E) then to calculate 3 out of a million from -110 dB.
DeleteIMX174 also shows much higher QE than the other 3 in the table. I wonder if there are any tricks. I suppose it is not BSI, or could it be?
DeleteI wonder what is the price Sony quotes for this type of ICs?
ReplyDeleteAny other camera maker using IMX174?
ReplyDeletehttp://ocularcomponentmv.blogspot.sg/2014/01/imx174-sonys-ground-breaking-global.html