Talk:Autofocus

This article is rated C-class on Wikipedia's content assessment scale. It is of interest to the following WikiProjects:

Photography Mid‑importance This article is within the scope of WikiProject Photography, a collaborative effort to improve the coverage of photography on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.PhotographyWikipedia:WikiProject PhotographyTemplate:WikiProject PhotographyPhotography articles Mid This article has been rated as Mid-importance on the project's importance scale.

Autofocus assist beam was nominated for deletion. The discussion was closed on 22 December 2009 with a consensus to merge. Its contents were merged into Autofocus. The original page is now a redirect to this page. For the contribution history and old versions of the redirected article, please see its history; for its talk page, see here.

I presume that the cameras (still and camcorders etc) which use infra-red light to autofocus must have an IR LED to provide the illumination? I ask because I am trying to estimate the market for such components. Given the number of DSCs and camcorders sold this must be a substantial market for IR LEDs.

Meanwhile, other types of LEDs are being used. For example:

Avago Technologies has introduced a high-brightness green auto focus auxiliary flash light emitting diode (LED) to assist the auto focusing function in digital still cameras when used in low ambient light conditions. The ASMT-FG10 flash LED offers a small footprint, measuring 4.8 x 4.8 x 5.33 mm.

Touted as a low-cost alternative to conventional infrared (IR) auxiliary auto focus solutions, the ASMT-FG10 compact green LED lamp is said to provide the highest brightness and smallest package in the industry. The narrow angle SMT lamp package is suitable for digital cameras that require long distance illumination and a narrow beam pattern. Digital still cameras have twice the sensitivity when responding to the color green than other colors, due to the availability of more photo diodes allocated for green in each pixel, said the company.

The ASMT-FG10 green (530nm) Indium Gallium Nitride (InGaN) dome LED lamp uses an un-tinted, non-diffused lens to provide high luminous intensity within a narrow radiation pattern. It incorporates an encapsulated LED chip on an axial leadframe to form a molded epoxy lamp package with six bended leads for surface mounting. The InGaN material has a very high luminous efficiency and is capable of producing high light output over a wide range of drive currents, said the company.

The easy to install ASMT-FG10 also simplifies the assembly process, which is said to result in a cost savings and better overall performance of the camera. It also features a six-degree viewing angle and is compatible with 2x solder reflow process. It meets IEC/EN 60825-1 Eye Safety Class 1 requirement. The device is RoHS compliant. Click the link for more information about the ASMT-FG10 LED.

The ASMT-FG10 SMT LEDs are priced at $0.60 in quantities of 10,000. They are supplied in 16-mm tape on 380-mm (15-inch) diameter reels for compatibility with automated manufacturing. Samples and production quantities are available now.

Avago Technologies, 1-800-235-0312, www.avagotech.com/led

Earlier this US company had released another type of visible LED: AlInGaP-based orange 612nm LED specifically designed to provide the illumination required to operate the automatic focus function of a digital still camera under low-light conditions. Most cameras use an infrared LED for this function, although some already use visible devices. http://www.ledsmagazine.com/news/2/11/13

RadioShack has a high power IR LED which it says is suitable for AF: http://www.radioshack.com/sm-high-output-infrared-led--pi-2062565.html

A high-performance infrared emitter whose peak wavelength is 940nm. These high-output infrared LEDs are a perfect light source for various kinds of security systems that need higher radiation. They also can be used in products ranging from TV remotes and auto-focus cameras to electronic gates.

81.86.144.210 11:34, 14 June 2007 (UTC)[reply]

Don't Sony's newer cameras have phase detection sensors on the main sensor as well as under the mirror? Should that be added to the phase detection section? — Preceding unsigned comment added by 24.3.214.1 (talk) 05:14, 31 December 2012 (UTC)[reply]

Most mirrorless cameras and many DSLR's in live mode use this technique.

It is genericly reffered to as Hybrid Autofocus. The Hybrid Autofocus sections could do with a little TLC to reflect the state of the art. — Preceding unsigned comment added by 80.94.115.69 (talk) 11:24, 18 July 2019 (UTC)[reply]

Canon EOS 3 has 45 AF points. 6 of them act as high precision cross-type AF sensor with f/2.8 or faster lens. Central point acts as high precision cross-type AF sensor with f/4 or faster lens.

Reference : http://www.bobatkins.com/photography/eosfaq/eos3af.html

I think that this means that Canon EOS 3 has

• 45 normal precision horizontal-line sensitive single-axis AF sensors (each AF sensor has 2 CMOS sensors)
• 7 high precision vertical-line sensitive single-axis AF sensors (each AF sensor has 2 CMOS sensors)

So, strictly speaking, we should say that "Canon EOS 3 has as many as 45 AF points" or "Canon EOS 3 has as many as 52 single-axis AF sensors".

What do you think? --shotgunlee 04:52, 9 December 2005 (UTC)[reply]

I think it has 45 Area-SIR points in total, arranged as follows:

• 38 are horizontal-only standard-precision sensors active for any lens faster than f/5.6
• 6 points come in cross-type configuration with additional vertical high-precision sensors active for f/2.8 lenses or faster, among the horizontal standard-presicion sensors
• the central point has a standard-precision horizontal sensor for lenses up to f/8, and features a higher-precision vertical sensor active for f/4 lenses or faster

However, it's only my guess and there could be additional horizontal high-precision CMOS sensors; it certainly does not indicate how many actual sensing devices each point employs. --Dmitry _{(talk •contibs )} 16:30, 31 August 2006 (UTC)[reply]

Three of the non-cited refs were added in 2005 at the end of a long string of edits, so presumably they are sources for some of what this guy added, here. One was added since then (by pvosta, who did no other edits, and which doesn't appear to be a likely source for anything in the article, so I took it out). We should try to cite them. Dicklyon (talk) 04:32, 26 December 2008 (UTC)[reply]

I have heard that the more autofocus points there are, the more accurate the autofocus is. Why is that? --The High Fin Sperm Whale (talk) 17:38, 23 September 2009 (UTC)[reply]

Things move. It is more likely that a moving object will be detected again at a different AF point, giving an opportunity to refocus. Maybe misleading to say "more accurate" but gives more sharp pics ! (I'm talking about Nikon's Dynamic AF-area system - others probably exist)

Plus you are less likely to have to recompose after focussing, although that's easier than selecting an AF point !

Digitutor D300

Select "Open Menu" then "Focus"

--195.137.93.171 (talk) 00:34, 16 May 2011 (UTC)[reply]

It appears unclear to the reader what the two-step method actually refers to. Please clarify in the section on active systems.

Polaroid SX-70 Sonar OneStep

I guess a single button-press did AE, AF, shutter !

An exception to the two-step approach is the mechanical autofocus provided in some enlargers, which adjust the lens directly.

Probably depends which make/model of enlarger - my only experience was a Leitz Focomat which had a wheel running on a linear cam parallel to the focussing-column, which moved the lens.

However, it might be best to delete this sentence, since it's unlikely to mean much to the youth of today ! Or tomorrow. --195.137.93.171 (talk) 02:15, 16 May 2011 (UTC)[reply]

"Phase detection" is something of a misnomer. Nikon show something that works exactly the same way as Leica's co-incident image rangefinder, except that the images are not co-incident, but have a 'correct' spacing when 'in focus'. --195.137.93.171 (talk) 02:36, 16 May 2011 (UTC)[reply]

whereas cameras like the EOS 40D and 7D have a custom function (III-1 and III-4 respectively) which can stop the camera trying to focus after it fails.

That doesn't match what is described in the rest of the trapfocus section, what is it talking about? --TiagoTiago (talk) 02:48, 27 October 2011 (UTC)[reply]

Not sure, will someone who knows this system confer? Should it be deleted or reworded?P0PP4B34R732 (talk) 03:00, 27 October 2011 (UTC)[reply]

IP 77.185.58.188 has recently added a misleading sentence into the introduction implying that hybrid AF systems always use more than one sensor, claiming (in edit summaries) that hybrid AF systems using a single sensor would not exist. I corrected this to say that hybrid AF systems are combining more than one AF method, but got reverted.

There are more examples of hybrid AF systems using a single sensor, but two examples should be enough to prove the point:

• A passive AF method like phase-detection AF becomes a hybrid system (active and passive) in low contrast situations, when it is assisted by a red AF illuminator LED. The same physical AF sensor is used in both modes of operation, although in active mode the AF algorithm's offset parameter gets adjusted to cope with the near-IR temperature of the light emitted by the AF illuminator.
• Various new cameras support both, contrast AF and on-sensor phase-detection AF off the same image sensor. In some sensor designs, the PD-AF pixels are separate from image pixels, whereas in other designs the PD-AF pixels are also utilized for the image itself. Both are hybrid AF implementations using a single sensor. The Canon EOS 650D is an example of a camera using a single sensor hybrid AF system.

So, while there are hybrid AF systems using two or more sensors, the property defining hybrid AF is not "multiple sensors", but the combination of "multiple methods". --Matthiaspaul (talk) 11:41, 11 March 2014 (UTC)[reply]

Totally wrong. (Update: Key messages in bold.)

• • An AF illuminator does NOT turn a passive autofocus into a hybrid (update: or active). Try to change all (update: passive AF) camera articles with AF illuminator into hybrid (update: or active) and see that you will be reverted by many. Read this article carefully to understand.
  • NO CAMERA with one sensor type is called hybrid autofocus. In detail: Canon 650D or others have a hybrid image sensor, or better: A device combining an image sensor and an autofocus sensor(s).[1][2]

And: The method(s) follows the single, related, specific sensor type. So:

• Dual or more methods on a single sensor type does NOT turn an autofocus to hybrid: See Face detection, Tracking, Trap.... etc. etc.. Too simple.

Probably your error as a programmer: This article is mainly about hardware and its function - not software. Write a software turning a single sensor type into a hybrid (for example your mentioned alleged SIMPLE image sensor with SOFTWARE phase detection AF) and get millions of $$$ or €€€. Canon 600D CHDK with phase AF????

All info given here, mostly already in the article and its links. If you read carefully it could be finally EOD. 77.186.115.191 (talk) 16:57, 30 March 2014 (UTC)[reply]

I agree with Matthiaspaul (talk · contribs). He is not saying that one can convert to a hybrid AF system if the needed hardware is not there. In his examples, one requires the AF-assist lamp (projecting a pattern usually, so it will allow focusing even on untextured surfaces in some cases), and other requires the special phase-sensitivity pixels that are built into some sensors to support hybrid AF. Hybrid AF systems predate digital cameras (see e.g. ELPH 370Z, or this '96 article on active/passive hybrid AF), so restricting the definition to the latest generation of phase + contrast hybrids is too narrowing. Dicklyon (talk) 19:05, 30 March 2014 (UTC)[reply]

@Dicklyon: Wrong: "restricting the definition to the latest generation of phase + contrast hybrids is too narrowing" : I said nothing about it. Read the Wiki article carefully.

Wrong again: You contradicted yourself: The given source said that an AF-illuminator does NOT change a passive AF into an hybrid. 77.12.86.115 (talk) 02:41, 31 March 2014 (UTC)[reply]

The article seems pretty clear, that the use of the assist lamp makes the system "active", and says, "The best is hybrid active/passive AF, in which one type backs up the other." Dicklyon (talk) 03:17, 31 March 2014 (UTC)[reply]

You are right that you are not "restricting the definition to the latest generation of phase + contrast hybrids." But you are restricting to two-sensor variants, and rejecting the active/passive hybrid that the 1996 article talks about. Dicklyon (talk) 03:21, 31 March 2014 (UTC)[reply]

Like the 1996 text the article uses the term "activates" for the illuminator, and both use quotation marks. NO text uses the term "hybrid" for that. A lamp is NO (update: additional or independent) AF-system.

And i am not restricting but clarifying: Your variant is USELESS for intelligent people reading the following index. I hope finally everything is clear. 77.186.2.16 (talk) 03:45, 31 March 2014 (UTC)[reply]

OK, we agree that a lamp is no AF system. But using a lamp can make an otherwise passive AF system active, which is what the article is about, if I read it correctly. How do you read it? Dicklyon (talk) 04:04, 31 March 2014 (UTC)[reply]

Ah, I see your point now. That article distinguishes the active beam from the assist lamp. It doesn't talk about different sensors, though. Why not leave it at different methods? Dicklyon (talk) 04:16, 31 March 2014 (UTC)[reply]

Finally.

Methods: Would be min. misunderstanding or wrong: Already answered: "Dual or more methods on a single sensor type does NOT turn an autofocus to hybrid".

All infos given. Thanks for careful reading and EOD ? 77.186.2.16 (talk) 08:14, 31 March 2014 (UTC)[reply]

I'm afraid your logic is flawed.

Your examples (face detection, tracking) do not belong, as these are software algorithms or "strategies" to determine what should be in focus, that is, they use the input from possibly multiple measurement methods and possibly multiple sensors to decide where the focus should be, they do not and cannot achieve focus by themselves. They work on top of whatever set of physical AF methods are implemented in a camera. Various methods exist, and they may use independent sensors or not.

The usage of multiple sensors does not necessarily make it a hybrid system. The defining property here is the number or type of methods, not the number or type of sensors. Just think about a traditional PD-AF system with dozens of sensors (of the same basic type) - the fact, that is uses more than one sensor, doesn't make it a hybrid system. However, a system combining contrast AF with PD-AF is called a hybrid system, regardless of the number of sensors used. And you are wrong, when you state, that hybrid systems must use more than one sensor and that there would be no hybrid systems based on a single sensor.

The red LED AF illuminator is an example where a passive and an active method use the same sensor. I specifically used the dark-red LED example, not the white linear flash example, because in the former a pattern is projected onto the subject making it possible to focus on subjects with untextured surfaces, and also because the algorithm's parameters are actually different in order to compensate for the temperature of the light (if you would use the same parameters, the focusing would be off). I agree, that this isn't often called hybrid (more actively assisted), but at least by the definitions for active and passive in the article, it is a combination of active and passive methods using the same sensor.

However, the image sensor example is a stronger one, as the term "hybrid" is in common use, when they combine contrast AF with phase-detection AF. As I already wrote, there are implementations using dedicated AF pixels, which do not contribute to the image, as well as implementations, where (modified) imaging pixels are utilized for PD-AF as well. While the former can be called (and actually is called) a hybrid system based on the fact that is uses contrast AF and phase-detection AF off the same image sensor die, the second type of PD-AF image sensors can even be called a hybrid system in a more microscopic view, as they use the same physical imaging pixels for contrast and phase-detection AF. If you are not aware of them, perhaps studying some patents helps.

--Matthiaspaul (talk) 11:11, 31 March 2014 (UTC)[reply]

Nothing new. read again. Stop senseless addition of text. 77.186.85.251 (talk) 06:40, 1 April 2014 (UTC)[reply]

PLEASE Think: You are supporting my view: Your messages above: "The usage of multiple sensors does not necessarily make it a hybrid system", "a system combining contrast AF with PD-AF is called a hybrid system" etc. : TYPE is the used word: see above.

hybrid image sensor: see above. 77.186.85.251 (talk) 07:03, 1 April 2014 (UTC)[reply]

The dynamic IP editor keeps reverting attempts by two of us to make the lead correct. Obviously, there's no consensus to say things like that AF type is named for the sensor type, since active and passive methods can use the same sensor, and do so in various combinations. His edits are getting disruptive, and I will ask for semi-protection if it continues. Talking here and trying to acheive consensus would be better. Dicklyon (talk) 15:16, 2 April 2014 (UTC)[reply]

Then stop reverting. Until consensus i set the intro back to the state which was accepted for a very long time.

"since active and passive methods can use the same sensor" What specific camera is called to have this? I doubt that you find one. 77.185.5.250 (talk) 21:18, 2 April 2014 (UTC)[reply]

I think with two of us explaining that "The methods are named by the used sensor: Active, passive and hybrid." is basically nonsense, plus the fact that it's not even good English and has wrong capitalization, is evidence that the lead has not received the attention it deserves. You had it right here, but then took a step in the wrong direction here. So we're slow to fix it, but it needs to be fixed. It's time to let go of your mistake. Dicklyon (talk) 23:20, 2 April 2014 (UTC)[reply]

@Dicklyon and Matthiaspaul: You both are irrational (unable to tell ONE real example) and unable to think and learn. EOD 77.185.5.250 (talk) 02:38, 3 April 2014 (UTC)[reply]

I'm pretty sure both Canon and Nikon now have sensors that do hybrid contrast and phase detection on the same main sensor chip; at least one is mentioned in the article, iirc. Looks like Sony has one, too (with "99 phase-detection AF points arrayed on the image sensor" ... "similar to the Hybrid AF technology we saw introduced in Nikon's 1 Series cameras, like the recently announced J2, which has 73 AF points on the main imaging sensor used for Phase Detect AF"). Read this page that contrasts Nikon J1's one-sensor true hybrid with other hybrid AF systems. And whether it exists or not in a camera, it's certainly not hard for a sensor to support both contrast and an active structured-light rangefinder; as in this microscope hybrid AF system. Dicklyon (talk) 03:28, 3 April 2014 (UTC)[reply]

Hello fellow Wikipedians,

I have just modified 2 external links on Autofocus. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:

• Added archive https://web.archive.org/web/20090621131329/http://www.lfi-online.com:80/ceemes/page/show/issue_4_09_3 to http://www.lfi-online.com/ceemes/page/show/issue_4_09_3
• Added archive https://web.archive.org/web/20100818143041/http://www.eosdoc.com:80/manuals/?q=ATrapFocus to http://eosdoc.com/manuals/?q=ATrapFocus

When you have finished reviewing my changes, please set the checked parameter below to true or failed to let others know (documentation at {{Sourcecheck}}).

This message was posted before February 2018. After February 2018, "External links modified" talk page sections are no longer generated or monitored by InternetArchiveBot. No special action is required regarding these talk page notices, other than regular verification using the archive tool instructions below. Editors have permission to delete these "External links modified" talk page sections if they want to de-clutter talk pages, but see the RfC before doing mass systematic removals. This message is updated dynamically through the template {{source check}} (last update: 5 June 2024).

• If you have discovered URLs which were erroneously considered dead by the bot, you can report them with this tool.
• If you found an error with any archives or the URLs themselves, you can fix them with this tool.

Cheers.—InternetArchiveBot (Report bug) 05:39, 22 October 2016 (UTC)[reply]

Now in article this:

"A newer approach included in some consumer electronic devices, like mobile phones, is based on the time-of-flight principle, which involves shining a laser or LED light to the subject and calculating the distance based on the time it takes for the light to travel to the subject and back. This technique is sometimes called laser autofocus, and is present in many mobile phone models from several vendors. "

but more relible sources wrote [3]:

"The laser focusing aid provides this light. The output from a low-power laser is scattered out across the scene, giving multiple points of light upon which the contrast detection system can hang. Focusing on a little point of light on a subject is just as good as focusing on the subject itself."

"The LG G3 focus aid uses a low-power Class I laser, whose output is effectively split into fragments in order to give multiple points of focus. An already low-power laser’s output will be tiny by the time it actually reaches your subject. " Warmagain (talk) 13:35, 5 August 2020 (UTC)[reply]