# Talk:Image resolution

Now that I have boldly added sections on bit depth and linked to the chroma subsampling stub, does this need disambiguation, or are they subtopics? Ehusman 22:36, 4 December 2005 (UTC)

Given that

I say YES! but I have no idea how to create a disambiguarion page. Please do it ASAP! --Onejaguar 00:05, 7 December 2005 (UTC)

## Horizontal v. Vertical pixels

Am I misinterpeting which is the long axis on a televison? Isn't a TV longer in the horizontal axis (from left to right) and shorter in the vertical axis (top to bottom) as seen on my TV, and this monitor I am writing this on, and this diagram: Standard-definition television?

The resolution in various media section just went from overly simple to incomprehensible. It reads like the manual to a TV. I haven't a clue what the technical terms mean. Can someone make it readable for someone who doesn't know what all that technical jargon means? Or can the information be simple on this page and lead to more complex information in a link? --Richard Arthur Norton (1958- ) 07:19, 22 February 2006 (UTC)

I see what you mean, Richard. This is a problem with anything potentially technical like this: how do you make give it graduated levels of complexity? I'll give it a try this weekend. Ehusman 20:59, 24 February 2006 (UTC)
I think we should go back to the simple format of a paragraph giving the number of pixels and megabytes in each format then the reader can click on 780i and read all the technical specifications about various pixel shapes in that format. We can have a new header in the 780i article and call it "resolution" and move the tech stuff there. Its all useful, but I prefer simple articles leading to more in depth and complex articles a click away. It should be readable by someone not familiar with the information so they can learn it. Also the tech terms need to be wikified and have simple explanations. I wont be able to help, since its not my field. I just wrote the simple stuff so I could understand for myself. Also, someone keeps changing the number of pixels in the horizontal vs vertical. The large number is always the horizontal, but three times someone changed it. Can you confirm I have them right? --Richard Arthur Norton (1958- ) 21:35, 24 February 2006 (UTC)

It also seems that in the "resolution in various media section", pixel resolutions are conflated with analog line resolution measures. For example, an NTSC laserdisc doesn't have a "resolution" of 420x480. The 420 references TVL (television lines), which is a measure more akin to DPI, except scaled. The section (http://en.wikipedia.org/wiki/Television_lines) describes this very well. So 420x480 for laserdisc is really a senseless measure of resolution (more accurately it would be 560x480 for a 4:3 display screen, meaning that a 4:3 display screen resolves a total of 560 vertical lines with a horizontal scanline count of 480 for NTSC). Another example of misleading information is giving 720x480 as the resolution of anamorphic widescreen DVD. Obviously 720x480 is the frame resolution for ALL NTSC DVD, not just the anamorphic ones (plus, the frame resolution of actual image information is only 704x480). This section (http://en.wikipedia.org/wiki/Pixel_Aspect_Ratio) and this section (http://en.wikipedia.org/wiki/Aspect_ratio_(image)) both provide excellent information about all this. Should someone make a note to this effect in this section, or should the list of "resolutions" be scrapped altogether, providing links to other relevant, and more importantly, accurate descriptions? Tamaskovacs (talk) 23:06, 24 March 2011 (UTC)

It seems to me that this started out with a discussion of image resolution and is gradually getting loaded with the various types of capture, recording, and display media. The image resolution article should stick just with image-related properties, and then link off to the media- and standards-related issues. For example, under "Image capture", the statement says that "35 mm ... film ... has about 6,000 lines ... when using the best lenses and the finest grain film." That's three different statements about film (in general), lenses, and fine grain film rolled into one potentially misleading statement. The film resolution is independent of the lens resolution, but the system resolution depends on both; the sampling mechanism has its own resolution, and is included in the entire system resolution of the final product; but the initial statement says that 35 mm film has 6,000 line resolution. All of this stuff is part of an overall project, but I don't know how to address image resolution separately.
I started the Optical resolution entry specifically to deal with high-level physics and mathematics of system resolution. There are already several pages dedicated to analog and digital video display resolution. I think that the image resolution page should deal specifically with two issues: spatial sampling or quantization error (of which chroma sampling is a part), and dynamic range. The initial image at the top of the page is a good start for the former, but the latter needs some work. For people looking for the resolution of various media formats, perhaps a table and a set of "See also" links would suffice, and remove all of the extraneous discussion. That way, someone wanting to know just a little would be able to get the gist, but the links to dig deeper would be readily available. The column headings would be "Media type" (film, SDTV, HDTV, computer-related ?GA), "specific format" (35 mm, NTSC, PAL, 1080i), "horizontal x vertical pixels", "Pixel aspect" (square, etc.), "Pixel depth", "Color sampling", "See also" links (e.g. to CCD, Photographic film, Display resolution, Video tape recorder, Hdtv, NTSC, PAL, SECAM.

## Banding.. is this the right term?

In the dynamic range section, it currently states, "Bit depth can impact dynamic range capability if you include the caveat of banding or contouring issues. If you have an insufficient bit-depth for a given dynamic range, the discrete steps between one level and the next can become distinctly visible, and gradations in the image will not appear smooth; instead they will have visible banding problems.". Surely it is refering to posterization, not banding? From my experience (specifically with digital cameras), banding has been vertical or horizontal lines relating to the reading of CCD/CMOS lines from the sensor. So, should this term be replaced? Posterization has a relevent article, and banding does not. Diliff | (Talk) (Contribs) 09:13, 23 May 2006 (UTC)

## NTSC DVD's have 480 lines of resolution

The article states that "DVDs have roughly 500 lines" but is an approximation necessary? NTSC DVD's have exactly 480 and PAL exactly 576. Why not just state that? I would make the edit, but then I'm not certain if there's some technical reason for the approximation. Also the article refers to common frame rates of 15 or 30 frames per second. These may be common for NTSC but as I recall 24fps is common at the cinema and 25fps common in Europe (specifically the uk). 92.40.253.169 (talk) 22:03, 14 March 2012 (UTC) stephen

I agree, this would make more sense. Do you have a source? I know you're right about the figures, but this article has been slack with sourcing things. I'll look for one if I have a chance, but I don't right now. Micahmedia (talk) 03:07, 8 April 2012 (UTC)

## Confusion

The article used to list the number of lines of resolution horizontally and and the number of lines vertically and then multiply them for the number of megapixels. Now "columns" and "lines" are used incorrectly. When you talk about lines of resolution you have to mention whether they are in the vertical axis or the horizontal axis

I'm not familiar with lines being used other than in the sense of how many horizontal lines can be resolved. If you have a different interpretation, find a source, and put it in the article; after that the examples can say which method or interpretation they are using. Dicklyon 17:01, 3 September 2006 (UTC)
"Line" just means a line in any direction (including non-right angles, like on a vector CRT,) a vertical line of pixels is called a column and a horizontal line of pixels is called a row. As for references, look at the article; second subhead, about halfway down the first paragraph. You can't miss it. :-) 208.54.14.81 18:30, 3 September 2006 (UTC)
Feel free to think of me as slow, but I'm not finding what you're referring to as a reference. And your interpretation of "line" is not one that I am familiar with in the context of resolution. How about a citation? Here are some examples of bookw with lines as used in this article: [1] [2] [3] [4] That last one has "horizontal lines of resolution" for what you're calling columns, I think, but none cite image resolution in terms of rows and columns; that's more of a "pixel resolution" thing which is not an actual resolution; unfortunately, it's what a lot of the examples are about, like the 35mm movie film scanners. This could use some clarification, for sure, but simply swapping out lines is not the answer. Dicklyon 21:48, 3 September 2006 (UTC)

## article is confusive

Hello, colleagues! I agree, the article is confusive. It is not because you did not edit it well, it is because the term has too many meanings. For example, what does mean each of the sentences below?

(A) The resolution of this microscope is 1 micron.
(B) We need to increase the resolution.
(C) At $~\rm NA=0.2~$, the model predicts the resolution $r = 300 \rm \frac{lines}{mm}$.

First at all, should we treate resolution as some physical quantity like length of frequency, or it is just fussy term like beauty or generosity?

(Perhaps, beauty is not a good example, this term may be in use in quantum chromodynamics, where physicists have to give names to new fundamental quantities.)

If we treate the resolution as physical quantity, we should state that in the case (A) and (C), pretty different (and, perhapse, inverse) quantities are called "resolution".

This confusion sould be mentioned at the preamble of article resolution.

then, for each meaning, the definition should be provided.

Then, different quantities called "resolution" should be related with mathematical expressions, for example,

${\rm resoluiton}_{\rm A} = \frac{1}{ {\rm resoluiton}_{\rm C} }$

Then, if one needs to increase the resolution (example B above), then, perhaps, one means the resoluitonA and an adverticement of some high-resoluiton optics may mean the resoluitonC.

Without such a specification, the term "resolution" has no more meaning than a $~\sqrt{\rm generocity}~$ or $~\log({\rm stupidity})~$.

(Similar problem takes place with som eother terms, for example, Einstein coefficients; not only value, but even dimension of these coefficients depends on the system of notations.)

Either we should declare one of meanings as true, and other as wrong, or to insist that this term has not yet established meaning, and therefore any use of term resolution is possible only after the definition within the same source.

dima 03:15, 1 February 2007 (UTC)

Link problem--at least when I checked--at http://fred.dsimprove.be/__offSite/pixel/ (see "Pixel & Resolution Tables").

--C-U RPCV 15:40, 5 April 2007 (UTC)

## Yukky image

I removed Landen99's image with comment Horrible gross ugly too-big image. He put it back with Image was put in thumb form. Image is high quality work which illustrates this section very well and intuitively. But more importantly, he replaced it with a white-background version instead of the horrible gross ugly gray-background one that had no usable luminance contrast, which took care of at least of a few of my negative characterizations. While I still can't agree that it's high quality work, or particularly useful, it's at least more nearly tolerable, so I'll leave it for others to comment on or decide on now. Dicklyon 06:05, 11 October 2007 (UTC)

On second look, it's not tolerable. The illustrated rectancles don't have the aspect ratios that they claim to have, so it does more harm than good. I'll take it out again. Dicklyon 06:07, 11 October 2007 (UTC)

You are right about the lack of scale on the subset image being intolerable, and now that issue has been easily corrected. Enjoy. ---landen99 "I am, therefore I think." Ayn Rand 17:31, 25 October 2007 (UTC) —Preceding unsigned comment added by Landen99 (talkcontribs)

Besides totally sucking, the image still has the wrong aspect ratio by a wide margin. And it's not really the most appropriate content for the topic of this article, which is trying to say that image resolution is not just a pixel count. Dicklyon 00:26, 26 October 2007 (UTC)

## Resolutions, comparisons, and non-square pixels

A discussion concerning the issue of comparing analog and digital resolutions and comparing non-square pixel resolutions with others and square-pixel resolutions has been started here as part of the Broadcast Engineering and Technology Taskforce of WikiProject Television. Please discuss these issues there so that as a community we can come to a consensus. --tonsofpcs (Talk) 18:04, 16 July 2008 (UTC)

## 520 > 480

Re: "720×480 (520 lines)"; how do you fit 520 lines into 480 pixels?

75.84.238.18 (talk) 11:23, 30 July 2009 (UTC)

## Lines: Resolution in various media - film

This section had the following statement: "Sequences from newer films are scanned at 2,000, 4,000 or even 8,000 columns (line measured the other directions),(...)" The statement in parenthesis makes no sense whatsoever. I'm not going to attempt to guess what the author meant, and neither are the readers. It starts off bad, "line measured," and gets uglier, "other directions." What other directions? I removed it. —Preceding unsigned comment added by DrSlony (talkcontribs) 13:49, 9 January 2010 (UTC)

## Film Resolution

• 35 mm original camera negative motion picture film can resolve up to 6,000 lines.

Where on earth did this figure come from? 35mm still images (~36mm x ~24mm) on average resolve around 2000 lines on the short side with color negative film. Motion picture formats are 24mm on the LONG side! There is no way that motion picture images on 35mm stock can resolve 6000 lines. There exists no lens that will put that down to an image size 24mm on the long side, let alone film that will resolve that. Check here http://www.normankoren.com/Tutorials/MTF1A.html for some stats on common films. According to these stats TMAX 100, (which is one of the highest res/finest grain films in current use) resolves around 125 lp/mm, so on a 35mm still image 24mm high that's 3000 lines, which is quite impressive. Panavision anamorphic is a little more than 17mm high, which gives us the potential of 2125 lines--hardly close to the 6000 stated in this wiki.

According to Arri, they've measured color film at 80 lp/mm: http://www.arri.de/fileadmin/media/arri.com/downloads/Camera/Tutorials/SystemsTechnologyBrochure.pdf which would give us a scant 1360 lines on anamorphic!

Scariest spec I can find is for the current Kodak 500T, around 25 lp/mm: http://motion.kodak.com/motion/uploadedFiles/TI2647.pdf ...which would give around 425 lines!

One better, Super 35 3-perf (popular these days) is only 13mm on the short side, which would give us only 325 lines if Kodak's specs are to be believed.

Films are scanned at up to 6k today for preservation, but this level of scanning resolves grains/dye packets which are beyond what could accurately be defined at "detail". This is done to minimize the appearance of the grain due to the aliasing it can cause, not because of the level of detail contained in the original negative. Higher resolution scans are later down sampled for archiving. That may be where the 6000 lines number comes from. But even then, that's 6000 columns, not lines, and a frame of which would have around 2500 lines/pixels of height at 2.39:1 aspect ratio.

I suspect 60-100 lp/mm is the most reasonable range for film resolution, with the low end representing common color film stock shot today, based on both academic research and personal experience. 2000 lines is a reasonable number to claim for negative film in anamorphic format and 1300 lines for Super 35. 6000 is unsupported and unless anybody has a source to backup that claim, I will edit the article appropriately. Probably by just removing the claim altogether. Micahmedia (talk) 10:13, 5 January 2011 (UTC)

:Rereading that Arri article, I see they measured 80 lp/mm at MTF25, not MTF50 which is the more accepted standard. At MTF50, the Kodak film tested 33 lp/mm. The pattern here is that figure of 6000 lines is way off. Micahmedia (talk) 11:29, 5 January 2011 (UTC)

Correction: I don't know what I was thinking--MTF50 is used for lenses, not capture media. However, the images show that 80 lp/mm is only just barely resolved. Micahmedia (talk) 11:37, 5 January 2011 (UTC)

## Pixel Resolution and Spatial resolution

"The measure of how closely lines can be resolved in an image is called spatial resolution"

"The spatial resolution of computer monitors is generally 72 to 100 lines per inch, corresponding to pixel resolutions of 72 to 100 ppi."

There is a fundamental error of interpretation here. To achieve spatial resolution, the system must see the gap between lines as well as the lines themselves. To resolve 72 lines per inch, the system must display 72 black lines and 72 intervening white lines. Thus to achieve a spatial resolution of 72 lines, 144 pixels are required. Seeing 72 black lines with 72 pixels means seeing solid black, so in other words, 72 lines are not resolved.96.54.42.226 (talk) 00:48, 13 October 2012 (UTC)

## Lines Of Res - CMOS and CCD Digital Camera

There is a part left out. It has to do with the resolution of CCD and CMOS devices used in digital cameras today. One has the following spec:

Resolution (TV Lines)     350 TV lines

It would be nice to put some practical context to things like this. - KitchM (talk) 15:55, 25 August 2013 (UTC)