Jump to content

Talk:24p

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

DVDs: 24fps or not?

[edit]

This page claims that Hollywood authors movies on DVD at 24 frames per second. The DVD-Video page only mentions 25 and 29.97 frames per second as valid encodings. This seems like a direct factual conflict. Neither article has much in the way of references (but I haven’t tagged them as such quite yet). If anybody knows the answer and can cite sources, please fix the article(s) and/or add citations. MacMog (talk) 04:02, 30 March 2011 (UTC)[reply]

Information about Telecine must be moved to Telecine page

[edit]

About 90% of material in the 24p atricle is devoted to explaining Telecine methods. This is wrong as Telecine has a dedicated Wikipedia page. Useful information about telecine must be merged into the Telecine page, and then removed from the 24p page.

Rebuttal

[edit]

I appreciate your comments, but I still think that this article needs improvement and that the changes I made could be worked into the text rather than just obliterated.

About high motion

I don't see how it causes "problems" for home movies? Home movies were at one time shot on film (eg. 8mm) and surely this is not a problem for people? Associating interlaced video with home movies is based on an expectation, as with news. I'd argue that a documentary might use a high-motion format for the same reason, but not that it necessarily introduces any problems. What is the meaning of "problems" here?

Lack of high motion on Super 8 isn't fatal, but the cost and complexity of film meant that it tended to be used only by people willing to put more effort and knowledge into it. Film users wouldn't hold a shot for 30 minutes, or try to shoot handheld from a moving car, ect. High Motion can make such practices tolerable. (barely)Algr 20:32, 14 June 2006 (UTC)[reply]

oddball workflow

My change was basically the same as the 2 frames for 3 fields description. It's a matter of frames vs fields, not speed-up. Speed-up is the most common method, of course, but if you're going to mention the duplication method, I think it's better to simplify this as duplicating a frame. From what I have read, this is common pratice, as opposed to bothering with fields. Of course, this is only when speed-up is not used, which is rare.

If you duplicate frames the motion defect would be twice as severe. I don't know of any workflow that would result in this. You'd get field duplication (and some lag) if you ran an NTSC tape shot on film through a converter to PAL. In the 1990s you got this a lot, as lots of US shows were shot on film, but edited on video, and speeding up the result wouldn't work because editing disrupts the 3:2 pulldown pattern. Star Trek TNG, and the Fox Doctor Who movie are programs that were converted like this. Here is a link to a group that dealt with this issue: [1] Algr 20:32, 14 June 2006 (UTC)[reply]

What is the "most" video editors comment based on? There is wide support for 25 FPS if nothing else, so it is misleading to claim that "most video editors are designed for 30 frames per second".

Most editors are designed for NTSC and PAL, only high end ones are designed for any form of 24p. Algr 20:32, 14 June 2006 (UTC)[reply]

"To avoid this, American movies and 24p video are usually sped up to 25 hz for viewing - but this can create audio problems"

Speeding up the film does cause problems. Obviously, speeding up the video means that the audio no longer matches. I think it is important to elaborate on this point. Either the audio is sped up - making it a semitone too high - or it has to be compressed with pitch correction. The most common is probably speeding it up; a change that you don't have to be an audio expert to notice.

I also object to the use of Hz (the "H" should be capitalised) instead of FPS. This could be confusing for people and FPS is more easily understood. FPS is a specific term that applies here, while Hz is more global term.

FPS is ambiguous since it could refer to frames or fields. Also it doesn't indicate if a video frame has motion within it or not. (As it would in a normal video camera's output since the two fields are exposed separately. )Algr 20:32, 14 June 2006 (UTC)[reply]

(Ironically NTSC nations have no motion problems with either 24 hz or Europe's 25 hz.)

First off, the use of parentheses here is incorrect and this is at least the second instance of such incorrect usage. Secondly, I disagree that the situation is "ironic". Although I would tend to agree that 25-to-30 FPS conversion is relatively problem-free, I'm not sure that everyone will agree. I again object to the use of Hz instead of FPS and if 25 FPS is going to be called "Europe's" then 24 FPS ought to be called "film's". I understand that my solution was convoluted, but I think this version is unacceptable and should be re-done.

Lastly, if you noticed errors in the original text, have they been corrected?

Eradicator (Monday, April 24, 2006 at 23:43:06 UTC)

Revert

[edit]

Sorry about the revert Eradicator, but while some of it was good, it added as many problems as it fixed, particularly in the disadvantages section. I thought it might be better if we take the changes one at a time.

(I wrote some other things and then noticed that the errors were there before and you only moved them. I want to get this posted before you reply, so I'll go straight to the disadvantages section: )

In duplicating the look of film, 24p also inherits some of film's disadvantages. Like film, 24p video is incapable of high motion. This can hurt the credibility of newscasts by making news footage look too much like staged movie clips. Similarly, it also may be less desirable for things such as home movies, documentaries and sporting events, which can benefit from high motion.

This breaks an important cause-and-effect relationship. The lack of High Motion directly causes problems for home movies, documentaties ect. But the effect on news credibility is much more intangible and based on viewer expectations.

Incorrect user settings can cause problems with timing and result in a loss of quality.

What I described was not a timing problem, it was lost data. Timing problems in video are usually about syncing two signals together, (For example picture with sound, or subtitles.) 3:2 pulldown doesn't effect this directly, although other issues do.

requires one frame to be duplicated in each second of video.

Only a very oddball workflow would result in this, any pro would avoid it. 4% speedup is what is usually preferred, and field duplication is the second choice.

Converting 25 FPS or 24 FPS to NTSC's 30 FPS, however, is generally free of such problems.

This rephrasing is just more confusing, and obscures the point by reducing it to a list of numbers.

24p and PAL

[edit]

As far as I am aware, the "speeding up" process of converting from 24P to PAL-compliant video is very common and quite un-problematic. I live in a PAL country, and all of our movie titles, e.g. those on DVD are obviously converted from 24 fps film up to 25p/50i PAL video... with no noticeable problems.

The reason is that converting 24p to PAL's 50 hz scan rate would require 2 frames in each second of video to be held for 3 video fields instead of two. The viewer would thus see motion stutter twice per second.

This is a convoluted way of explaining it, because it assumes 3:2 pulldown is the standard - which may be true in NTSC countries, but is quite irrelevant in PAL countries because the speed-up method is preferred anyway.

That passage explains what would happen if you DIDN'T speed up the film. It isn't 3:2 pulldown, but 3:2:2:2:2:2:2:2:2:2:2:2:2:2:2 pulldown. (Not that it is usually described that way.) Before nonlinear editing became the standard, American programs that were shot on film but edited on video usually ended up with the two stutters per second. Speedup is usually preferred, but not always possible. Also, I understand that the 4% speedup bothers people with perfect pitch. It is also quite noticeable if you hear it both ways. (For example moving the DVD from your PC to a TV. ) I once downloaded a clip, and knew right away from the sound of the voices that it had been sped up to 25 hz. Algr 19:46, 22 June 2006 (UTC)[reply]

24p resolution

[edit]

I don't think it's entirely accurate to say that:
the resolution of 24P DV is no higher than regular video when viewed on a television screen-- a point of confusion for many film and video makers

I assume that regular video = 60i. Most 60i cameras will use rowpair summation, which decreases the vertical resolution to about 360 lines. http://digitalcontentproducer.com/shoot/video_progressive_need_know/

For a filmout, 24p is obviously superior (although PAL 25p is arguably better than NTSC 24p). For television, we may eventually hit a point where most displays are progressive and can interpret 24p signals and not try to de-interlace them.

---

Although resembling film look in color and motion, the resolution of 24p DV is no higher than regular video when viewed on a television screen - a point of confusion for many film and video makers. When viewed on a computer screen however, progressive scan retains twice the vertical resolution than the otherwise requisite de-interlaced video one would have if not shot as progressive scan in the first place.

I removed that info because the claims are unqualified; if you qualified them, then it would make more sense.

Interlaced video is not half resolution on a computer scheme if a decent de-interlacer is being used.Glennchan 03:47, 27 November 2006 (UTC)[reply]

Panasonic Evangelism?

[edit]

Is there perhaps too much panasonic evangelism in this article? I certainly agree that the DVX100 is a notable camera for being the first DV camera (and the first 'affordable' cameras) to do 24p. However, the mention of the HVX200 isn't really relevant in this article (and neither is the Canon). Glennchan 03:47, 27 November 2006 (UTC)[reply]

Is anyone else besides Panasonic doing this "Advanced Pulldown"? It seems to me that with HDV cams all over, this may already be obsolete. Algr 15:02, 18 October 2007 (UTC)[reply]

The Canon cameras (XL2, XLH1, XHA1/G1) do. But even if they didn't, there are still tens of thousands of Panasonic cameras out there, and they're used by more indie filmmakers (i.e., those most likely to care about 24p) than any other cameras. Advanced pulldown is as relevant now as ever. —Preceding unsigned comment added by 75.178.92.119 (talk) 21:19, 13 December 2007 (UTC)[reply]

NTSC and 23.976 fps

[edit]

I suggest pointing out more clearly the relationship between the 23.976fps and NTSC, a video standard used in some parts of the world, such as North America. PAL countries shoot video at 25fps, so the "film effect" of low frame rate is not necessary. And in PAL, converting from 24.000fps film to 25.000fps video is not through 3-2 pull down, but by speeding up the film or duplicating 1 frame every second. Stephen B Streater 22:50, 20 December 2006 (UTC)[reply]

What do you mean about film effect not being necessary in PAL countries? And both those things you suggest pointing out are in the article (24p to PAL, and 23.976fps) David 09:49, 21 December 2006 (UTC)[reply]
The PAL frame rate is similar to film, so you don't need to play back at film frame rates to get a film effect - the progressive scan is more important. The introduction paragraph doesn't mention NTSC - but this is the main use for 24p video. In fact shooting at 23.976fps only makes sense when converting to NTSC for TV broadcast. Stephen B Streater 11:14, 21 December 2006 (UTC)[reply]
PAL is not 25.000 fps. PAL is 50 fields per second. Each field is an interlaced half-frame taken at a discrete moment in time from every other field. This isn't the same as 25.000 fps progressive; the motion has an entirely different look, more like 60i in NTSC. --75.178.92.119 (talk) 18:59, 3 January 2008 (UTC)[reply]
Progressive NTSC (29.976fps) would still look film-like, IMO - slightly smoother than 24p but nowhere near as smooth and "live" as interlaced NTSC/PAL. Most "filmic" PAL productions get (though perhaps not for much longer) shot interlaced, then get deinterlaced. David 21:20, 22 December 2006 (UTC)[reply]

Hype over nothing?

[edit]

This article, same as the telecine and filmizing articles, does not really answer the question of what the difference between 24p and 25p (PAL) is for the average viewer, assuming both were recorded using digital cameras. There is a lot of hype over 24p, yet I could find no explanation of why video filmed at 24p is allegedly so different to video filmed at 25p. Overlooking viewing equipment framerate issues (or assuming the video is being watched on a computer which has no pulldown problems since it can play all framerates), can anyone really tell the difference between 24p and 1 frame per second more? In my opinion a 1 fps difference is indistiguishable, and even if an experienced eye can perhaps spot that the video being watched is 24p and not 25p, would the difference be great enough to cause this much hype? Please explain what all the 24p filmlook hype is about for the average viewer, and if there really is no noticeable difference, then it might be a good idea to mention that 24p in digital cameras is not much more than marketing hype, a gimmick, created by digita camera manufacturers.
(please note, I am fully aware of what properties recording on traditional film has, of the larger dynamic level of film, etc. My confusion is about what makes 24p THAT WAS RECORDED USING A DIGITAL CAMERA so different to eg. 25p also recorded using a digital camera.)
DrSlony 11:16, 28 September 2007 (UTC)[reply]

My presumption on the matter is that 24p really is more attractive for traditionally NTSC markets, since there is a more noticeable shift between 29.97 and 24. But 24p of course also creates a 1:1 equivalence with film frames, so this also makes video to film transfers that much easier, which benefits all international distribution markets. Girolamo Savonarola 14:28, 28 September 2007 (UTC)[reply]
All else being equal, a viewer watching an NTSC or 60hz HDTV set would see almost no difference between 24p and 25p. However, 25p material would be easier to convert to PAL, (or 50hz HDTV) and 24p material would be easier to print to film. In both cases, if you use the "wrong" speed, you would need to speed up the program for PAL, slow it down for film, or allow for some studder in motion. The worst case is 25p to film without slowdown. Since film has no fields, you would have to simply drop every 25th frame. Algr 14:24, 18 October 2007 (UTC)[reply]

Clarify differences and reasons for numerous frame rates

[edit]

Hello all... I am interested in Digital cinema, (i.e. 'video to movie film transfer) so HDV and 24p is of extreme interest. (Not hype, if camera specs are valid) Since Film frame rates were established at the beginning of sound as being 24 FPS (And not a higher rate like 30 or 60 due to expense and size of film stock) this was a form of ' Analog Compression' of the information. Early projection of film at 24 had a problem of causing headaches and vertigo, (thus they were called "Flickers") so to hide artifacts, multi-blade shutters for projection were used. Each frame in a cinema is flashed on the screen 3 times, then advanced to the next frame- this is above the 'normal' human perception. NTSC TV was developed to synch with the 60 hertz of the N.A. power grid. So even though the resolution of TV was LOWER than cinema, it looked Smoother because it was OVER sampling 'movement' every 60th of a second- also higher than 'normal' perception. (Interlacing was ingenious method to SAVE bandwidth ) This was early ELECTRONIC analog compression. To show a 24th of a second Higher image resolution sample on a "lower" resolution 60th of a second medium, they came up with the 2:3:3: pulldown where some frames of film are shown twice... Also when USA films were shown in England on 50 hertz Film projectors that projected at 25, it caused some slowdown of the image but most audiences just figured Americans talked and moved that way... ( I assume that now films are projected there at 24FPS) But since English film (Esp. for TV) IS shot at 25 FPS, they have fewer problems transferring film to PAL, since their grid operates at 50 hertz and 25 divides into that much easier. The many issues and problems of choosing between all these different sampling systems is just HD & HDV & 21st Century Cinema inheriting the dilemmas of past 'solutions'. I Like the 24p on my hdv cam for the reason that it keeps the total chip vertical resolution of 1080 pixels progressive and I take the hit of losing over HALF'Bold text' my 'Temporal resolution' ( 60 minus 36 = 24 images per second) ... which is smoothed over by shooting at a rigid 48th of a second shutter speed -letting any fast moving part of the image smear slightly! (Which is what FILM does!) This can then be transfered to either film, video, or web and still start out at an impressive resolution. If I have any errors in this please let me know. Thanks! 208.74.106.71 20:24, 7 October 2007 (UTC)(MattBD)[reply]

There's no such thing as 25 fps English projectors - all film projectors in all countries are designed specifically for 24 fps speed. (Some occasionally allow for other playing speeds, but no others are standard.) American films and English films have no difference in presentation. It's only when transferring to video that there are incompatible systems. When not shooting for theatrical release, however, many filmmakers will use the framerate equivalent to their local video standard. As for "analog compression" bc of it being at 24 fps, it can't be defined as a compression - it's a speed standard choice. It would be compressed as compared to what? There were no standard shooting speeds prior to sound. Girolamo Savonarola 00:05, 8 October 2007 (UTC)[reply]

Double the bandwidth for 48p -- just speculation?

[edit]

I notice this section: "48p has twice the (video-like) motion (temporal resolution) of 24p, but also requires twice the bandwidth and data storage."

I'm not sure where this information came from and my guess is it is speculation. Obviously for uncompressed footage one would expect double the bandwidth to be required, but in all honesty most of today's video editing world works with slightly compressed intermediaries. Consider Panasonic's HVX200 and HMC150, both of which can shoot in 24p, 30p, or 60p. In all practicalities there is very little file size difference, because the extra frames in 60p footage are so similar (visually) to the frames next to it. Compression algorithms have a pretty easy time incorporating those extra frames without making the file sizes significantly larger.

I haven't independently verified this fact but I've heard it in passing. Maybe there's an analysis somewhere. --Oreckel (talk) 23:46, 23 June 2009 (UTC)[reply]

On the flip side, the extra data raises the distance between similarities, making the compressor's job harder. Rising definition also raises the prominence of noise.
I suspect the effect is minor on intermediary (capture, edit) compression because it doe not look far anyways ; but much harder on end (broadcast, storage, distribution) compression because it is much more process-intensive and must account for decoder limitations, especially limited memory to hold simultaneous frames for look-up.
Cutting frame rates is not the best way to ease your CCTV bandwidth burden has some figures (albeit surveillance-oriented), there is probably better to find.--Musaran (talk) 08:49, 31 January 2020 (UTC)[reply]

Is 24p a digital or film format?

[edit]

That needs to be made clear in the very first sentence. Thanks. Softlavender (talk) 07:15, 30 October 2009 (UTC)[reply]

[edit]

During several automated bot runs the following external link was found to be unavailable. Please check if the link is in fact down and fix or remove it in that case!

  • http://spng.se/frame-rate-test/
    • In Frame rate on 2011-03-19 11:26:47, Socket Error: 'A connection attempt failed because the connected party did not properly respond after a period of time, or established connection failed because connected host has failed to respond'
    • In 24p on 2011-06-19 05:39:09, Socket Error: 'A connection attempt failed because the connected party did not properly respond after a period of time, or established connection failed because connected host has failed to respond'

--JeffGBot (talk) 05:39, 19 June 2011 (UTC)[reply]

24p on DVD

[edit]

"DVDs, however, are capable of storing the native 24p frames. Every Hollywood movie is laid to disc as a 24p (actually 23.976p – see below) stream. With a progressive-scan DVD player and a progressive display, such as an HDTV, only the progressive frames are displayed and there is no conversion to an interlaced format – eliminating the appearance of any interlace or de-interlacing artifacts. When displayed on a standard NTSC TV (which only displays 60i) the DVD player will add 3:2 pulldown to the signal."

This confused me. The wording seems to imply that pulldown would only be needed on a standard NTSC display, but isn't it true that pulldown is needed on any display, including progressive ones that only display at 30 fps?

http://en.wikipedia.org/w/index.php?title=Talk:24p&action=edit&section=new#

Htjunkie (talk) 14:35, 1 October 2011 (UTC)[reply]

They meant progressive 24.000 and 24/1.001 displays (both are supported, dumb windows says it is 23 fps for 24/1.001 by windows 10 new setting correct that). Valery Zapolodov (talk) 16:30, 17 August 2021 (UTC)[reply]

23.976

[edit]

In the Section titles “23.976p“Some … actually have a frame rate of 24000 ÷ 1001, or 23.976023 frames per second. … This is because the "30 frame/s" frame rate of NTSC is actually 30/100.1%, also referred to as 29.97 frame/s”

This unsourced statement is inaccurate. The frame rates are calculated as 0.1% slower than their non-fractional counterparts. Thus the frame rates discussed are 30fps & 29.97fps, and 24fps and 23.976fps. The listed math and the figure “23.976023 repeating” are not from any standards body or reliable source. 107.179.173.217 (talk) 15:51, 22 April 2023 (UTC)[reply]

You are right that in the 90 the timecode was in fact 29.97000 and people suffered because of it. Because some frames accumulated, see SMPTE timecode ("This causes timecode to skip 18 frames each ten minutes (18,000 frames @ 30 frame/s) and almost perfectly compensates for the difference in rate (but still accumulates 1 frame every 9 hours 15 minutes).") But we are not in the 90. Nowadays most movies are either 24/1.001 or 24.000. Now, some older movies may have 23.976000 but it is rare as hell. I know of two: Dolby Vision "The Sword in the Stone" from Disney+ and also the 2021 US Blu-ray the Good, the Bad and the Ugly. As for displays you are wrong. Modern HDMI and DisplayPort standards all specify the fractional rate alternate clock to be perfect /1.001 (even if that means you need 9 hours 15 minutes to see the difference), not all displays support the alternate clock though. Valery Zapolodov (talk) 21:41, 13 August 2023 (UTC)[reply]