Talk:Image sensor/Archives/2014

CCD vs CMOS

The distinction between CCD and CMOS and the divergence of these technologies has occurred for a number of reasons. This article could be greatly enhanced if a comparison between the two are compared. Is there anyone who would like to provide a detailed comparison? PD 11:17, 7 July 2007 (UTC)

This part is confusing: "Extra circuitry next to each photo sensor converts the light energy to a voltage." A CMOS sensor requires no "extra" circuitry compared to a CCD sensor to convert light to voltage, since this is a task performed by the photodiode, which is basically common to both technologies. The extra per-pixel circuitry in CMOS sensors generally performs other tasks, such as amplification.

You're right, it's not very clear or correct. Both CMOS and CCD collect charge on the photodiode and there's an associated voltage; but in the CMOS sensors, that charge is read out as a voltage via an emitter-follower amplifier transistor, while in CCD the charge itself is shifted out. I recommend you look for a book that compares them and try to find some more accurate terminology to put instead. here is a place to start. Dicklyon 17:25, 3 September 2007 (UTC)

This is incidentally a major source of difference between CCD and CMOS sensors which isn't addressed in the article. The main disadvantage of a CMOS sensor is that they have less space reserved for the photodiodes (the fill factor), due to this extra circuitry. This is largely mitigated today by the use of microlenses, giving both CCD and CMOS sensors similar fill factors. A major advantage of CMOS is that the per-pixel circuitry also allows for per-pixel noise suppression, since the amplifiers are close to the photodiodes. This means CMOS sensors can have better performance in terms of noise, and have been replacing CCDs at both the low (cheaper fabrication) and high (better circuitry) end of digital camera lines. The article citing similar performance dates from 2001, and has been superceded by new technology. --69.227.70.158 16:28, 3 September 2007 (UTC)

It's not clear what you mean here by per-pixel noise reduction. Noise and noise reduction have actually been big problems in CMOS, solved not by the amplifier but by going to CCD-like fully-depleted pinned photodiodes, and/or CDS amplifiers in the column circuits. Dicklyon 17:25, 3 September 2007 (UTC)

The following link: http://www.asmag.com/showpost/12152.aspx?mv=&pages=1 could serve as an impartial source for CMOS/CCD comparison. Epitalon (talk) 13:55, 31 July 2013 (UTC)

Disambiguation

Please could we add this disambig to the top of this page? :

My reasoning is that it would simplify a whole lot of photography articles if we could use "image sensor" as a generic term ... IMO "image sensor" is the best candidate - eg easy distinction from "light sensor", etc - but I'd be delighted if someone can suggest a better one ... any thoughts ? --Redbobblehat (talk) 19:00, 13 February 2009 (UTC)

• "An image sensor intercepts the radiant energy propagating from the scene and transforms it to produce an intensity image. Two different technologies, viz., photochemical and photoelectronic are most common." (p.6) from (Googlebook[1]) Chanda, B. & Dutta Majumder, D. "Digital Image Processing and Analysis", 2005, ISBN 8120316185, 9788120316188. --Redbobblehat (talk) 20:00, 13 February 2009 (UTC)

OK, unless anyone has any objections I will add the disambig to the article in a couple of days. --Redbobblehat (talk) 19:23, 15 February 2009 (UTC)

I don't think "Retina" is appropriate here, but the others are OK. Dicklyon (talk) 19:33, 15 February 2009 (UTC)

Thanks for your support :) Do you think a "biological reference" is inappropriate, or "retina" is not the best one to choose ? ... biology isn't my thing, but a (limited) search around didn't turn up any alternative terms for octopi, etc. Eye obviously includes lenses and stuff, so it's not directly analogous to the "image plane sensor" we're talking about. --Redbobblehat (talk) 23:02, 15 February 2009 (UTC)

Are you thinking that someone is going to end up at this page when looking for info on biological image sensors? I doubt it. Probably also not really an issue to film and video sensors. Maybe it would be better just to put them into the "See also" section. Dicklyon (talk) 00:43, 16 February 2009 (UTC)

Yes, I am. That's exactly what I am saying :-). My point is that in optics the term "image sensor" is not restricted to a particular (CCD, CMOS, etc) technology; "image sensor" may be legitimately used to refer to a retina and/or all the other examples. The current article only refers to CMOS and CCD sensors. By disambiguating here, other wikipedia articles will be able to use image sensor in its full sense without having to waste space and clarity locally by listing the various non-CCD/CMOS types of image sensor as well.

Do you have sources that back up your assertion that the term image sensor is used inclusive of retina and such? Dicklyon (talk) 02:50, 25 February 2009 (UTC)

In addition to the googlebook reference given above, I can offer the exception that proves the rule from Jacobson's chapter "Sensitive materials and image sensors" in the Manual of Photography 2000. Jacobson meticulously uses "electronic image sensor" when referring to CCD, CMOS, etc. In the one instance I found - other than the chapter title itself - of the wider usage, he uses "image sensors size" in the context of APS film. The quote below also clearly demonstrates the specific use of "electronic sensors" where disambiguation is necessary ...

• "The film format and the cassette is smaller than 35 mm, which has allowed the design of more compact cameras with a very much simpler loading system (see Figure 12.14). Generally, for image sensors size is related to quality; the larger the area the better the quality. Manufactures of photographic sensitive materials have improved the quality of their products over a long period of time such that smaller format materials can be introduced which are of acceptable quality to the consumer, the most recent one being the APS technology mentioned above. Electronic sensors are a much more recent innovation and have had a shorter period of evolution of less than 30 years compared with 150 years for photographic materials, but are advancing rapidly." (Jacobson Manual of Photography 2000 p.201).

I can find no reference at all for "image sensor" in an authoritative dictionary or other source that might indicate any semantic change from a common noun phrase composed of "image" and "sensor". This simple semantic construction would suggest a generic term to indicate the photo-sensitive element of an optical system (ie. a "light sensor" would not necessarily require image-forming optics). It follows logically that the retina is the "image sensor" in a biological optical system. Unless someone can find a reliable source to disprove this, I suggest we stick to WP:UCS and give "retina" the benefit of the doubt. --Redbobblehat (talk) 01:06, 26 February 2009 (UTC)

The problem with "See also" sections is they are low visibility - not much more than footnotes and afterthoughts, whereas disambig is prioritised to help navigation around wikipedia. More laborious and verbose solutions might be to include the links in an article introduction, add extra sections (containing not much more than a link or two) or marking this article as a stub that needs substantial improvement to meet wiki standards... I'll add the disambig for now, and let other contributors elaborate at their leisure. --Redbobblehat (talk) 16:49, 24 February 2009 (UTC)

I reverted the triple hatnote; these usages are too unusual to rate such prominent notes. Dicklyon (talk) 02:54, 25 February 2009 (UTC)

Dick, IMO you have again [2] been too hasty to dismiss terminology because it is unfamiliar to you. You may wish to re-read WP:GAME. Having provided you - yet again - with a tiresome justification for a reasonable contribution, I trust you will have enough Wikipedia:Civility to undo your revert. --Redbobblehat (talk) 01:06, 26 February 2009 (UTC)

I did consider incorporating the "disambig" into the article's introduction, but I suspected that this would only cause more antagonism? A more top-down NPOV could actually be more interesting, and useful in reducing the proliferation of POV forks that plague wikipedia's optics, photography and cinematography articles? For example - a very sketchy example:

• "In an optical system, the function of the "image sensor" is to encode the luminous image produced by the lens.

• Image sensor characteristics include:
  • Physical size and shape: image circle/Coverage (lens), ... image plane/film plane, ... field curvature, etc.
  • Responsivity : sensitometry/photosensitivity!, useful exposure range (minimum and maximum sensitivity), spectral sensitivity, lux seconds, etc.
  • Acuity : photon detector/grain density, Optical resolution, Image resolution, Visual acuity, Optical transfer function, Modulation transfer function (infrared imaging), spatial frequency, line pairs per millimeter, pixel pairs per mm, ... etc. "

• Image sensor devices include:
  • In biology this function is performed by an array of photoreceptors called the retina.
  • In traditional photography ... layers of photochemical emulsion ... silver halide crystals (see photo detector) supported by photographic film/film base or photographic plate, ... photographic paper? ...
  • In electronic photography... photoelectrical: video camera tube ... CCD ... CMOS ... etc ... array of photo detectors (photodiodes, ... etc) --Redbobblehat (talk) 01:06, 26 February 2009 (UTC)

Yet, it remains highly unusual to refer to video tubes, film, and retinas as "image sensors". Not common enough that a disambig hatnote is needed, certainly. That would be the test. Lacking that, a see also or a mention in the article is fine. In fact, none of the 4 articles you linked even contains the phrase image sensor at all; if they are said "sometimes referred to as an image sensor", or "the image sensor in ..." or something like that, you would have the beginnings of a (weak) case, but there's not even that. Dicklyon (talk) 04:09, 26 February 2009 (UTC)

Technically, anything that converts an image to another form could be referred to as an "image sensor". It's even more general than your definition above - there is no reason to restrict yourself to a lens. It's often mirrors, or arrays of lenses (such as insect eyes), or direct phase/intensity measurements (radio astronomy images), magnetic field gradients (MRI images), or mechanical scanning, etc. All of these are used to sense images, but as Dick pointed out, no-one would first check the "image sensor" page to find information on film, etc. Perhaps we could add a sentence such as "This article is about CMOS and CCD sensors that convert images to electrical form. For other systems of sensing images, see the see also section", then add the "see also" section below. LouScheffer (talk) 05:04, 26 February 2009 (UTC)

I suppose any "sensor array" can be used to compile data into a human-visible image, but I was thinking of the more precise use of "image" in optics... something to do with focusing the light ... and AFAIK a given lens may be either dioptric (refractive) or catoptric (reflective) ? --Redbobblehat (talk) 18:03, 26 February 2009 (UTC)

OK, ignoring the 2 WP:verifiable references and semantic rationale in support of the wider meaning, giving this kind of priority to the informal popular use of "image sensor" makes it unusable as the specifically generic term I'm looking for :-(. Referring back to my starting point, can anyone think of another handy, concise term which we can use to refer to "the device that encodes the optical image at the image plane" ? ie. emphasising its function in the optical system without pinning it to a particular technology. It just strikes me as silly that every optics article has to use something irritatingly verbose like "the film/sensor/retina" rather than just one definitive catch-all term ... I liked "image sensor" because it works nicely alongside "image circle" and "image plane" (which AFAIK are the preferred terms in optics).

"optical sensor" seems a bit clumsy and less descriptive. AFAIK it could describe a simple light meter sensor, where the "optic" is a diffuser ...?

"photosensitive surface" is clumsy and could refer to a simple light detector ...?

"image encoding device" or "image encoder" I just made up ... ;-P

"focal plane array (FPA)" (Ray 2000 p.120) AFAIK is strictly solid state electronic image sensors, ie CCD & CMOS as opposed to video tubes, photochemical sensors, etc...

"focal plane sensor" or "focal plane surface" would inconveniently yield the acronym FPS ... :-(

"photographic plate" may sound antiquated, but I've seen it used generically and is still fairly descriptive?

"the film" (as in film plane, film format) is probably the most widely used and recognised term in photography, but this seems a bit colloquial and doesn't disambiguate easily ? --Redbobblehat (talk) 18:03, 26 February 2009 (UTC)

R, I can see why you'd want such a generic term, but I am not aware of any in use. I think we have to live with that reality, and not invent new terminology, or new application of existing terminology. The issue of comparing film, eyes, and electronic sensors goes back at least to Albert Rose in 1946, and I don't think he make up any such unifying terms. This guy uses "imaging devices" in reference to Rose's work. Dicklyon (talk) 04:23, 27 February 2009 (UTC)

I have to weigh in with Dicklyon here. Film is more properly regarded as a recording medium. Having worked extensively with both film and electronic sensors, I do not recall ever having heard film called a sensor. The term didn't come up except in reference to electronic sensors. I can see where eyes might be so regarded, but it seems non-standard. YMMV Ehusman (talk) 02:56, 16 February 2013 (UTC)

Area

Chart should include area. For recent sensor's the size is more important since most camera systems have sensor resolution much higher than what their optics can support. —Preceding unsigned comment added by 69.125.146.118 (talk) 18:43, 22 March 2010 (UTC)

I second this opinion. In the world of photography, quality is central and most talked about. The current technology is such that area has the most effect on quality when given two sensors of same resolution but different sizes. Anyone got the stats ? It's just another column (Preroll (talk) 02:18, 9 September 2010 (UTC)).

Pixel count <> resolution. Active pixel size ~ resolution, with caveats (such as fixed pattern noise, fill factor, etc.). See for example Optical resolution. Please stop confusing them. If you want raw pixel counts, please call them "pixel count" not "resolution". I believe that confusion on this point was originally caused by computer monitor manufacturers because old school photographers inherently avoided the issue by using lppmm when discussing resolution. Ehusman (talk) 03:05, 16 February 2013 (UTC)

Companies

In the list of companies that manufacture image sensors - do all of the companied listed actually manufacture them? For instance I thought Nikon only designs their sensors (or modifies other designs) and then gets other companies to do the actual manufacturing. This needs clarifying. — Preceding unsigned comment added by 202.89.25.245 (talk) 03:06, 26 September 2013 (UTC) I don't think that the companies that manufacture image sensors should be included in the article at all, unless it has a strong relation with the actual description of an image sensor. — Preceding unsigned comment added by 194.98.70.135 (talk) 14:33, 12 August 2014 (UTC)