Explain how the 29.97fps works exactly, please!

I didnt know if this was supposed to go in technology or not so I put it here. didnt think it was a viewfinder issue.

Anyways, I have a few question about the actual NTSC video frame speed.

Is there something called a "Skip Frame"? in which to make up for the 3/100th of a missing frame to be placed in at some point? does it just double a single frame every 3 seconds or so?

the reason Im asking is I saw this UFO thing on sci fi in which they can tell if something is composited into a video by a certain frame that a hoaxter would have a hard time faking or weather the object is just a video anomoly.

i hope this question makes sense. i heard the term skip frame before. Also if thats not what its called can you tell me the technical term or common term for it. and does it just slip a frame in somewhere to play catch up?
 
The original television system was Black and White and it used exactly 30 fps.  When color systems were developed, they were modeled after B&W, but the frame rate had to be changed ever so slightly, to exactly 29.97 fps, so that the color signal would synchronize properly with the sound signal.

Unfortunately, this has placed all the modern video hardware/software manufacturers in a difficult situation.  They could no longer report the elapsed time as before, where they used frames to run the clock  .  .  .  or maybe they could (we will get to that later).   The frame rate no longer fit nice and evenly on a per minute basis.  The number of frames per minute was no longer a whole number.

The SMTPE tackled the problem.  If they continued to run the clock from the frames and number them consecutively as before - then the first second of elapsed time, the frames would be numbered 1 through 30, and the timeline would report 1 second has elapsed.  But only 29.97 seconds has elapsed, therefore, the 30th frame would go a bit beyond the 1 second mark.  The reported time would lag behind the actual time.

For the periodic corrections, they needed to drop 18 frames for every 10 minutes of time.  Sounds easy - just drop 1.8 frames each minute.  But no - it must be an exact number of frames, since there is no such thing as a partial frame.  

To address this new frame rate (which is now 30 years old), the SMPTE came up with a standard known as Drop-Frame timecode.  Actually, they addressed four frame rates:  30, 29.97, 25 and 24 fps.  We will only talk about the 29.97 rate.

They defined both drop frame and non-drop frame formats.  Again, drop-frame timecode only skips frame numbers  -  no actual frames are dropped. Therefore with both drop-frame and non-drop-frame, the actual frames run along at 29.97 fps.  Drop-frame does not change the frame rate.  It's just a numbering trick that synchronizes the frame count.

They would use the 10 minute cycle, since 29.97 has an exact number of frames every 10 minutes.   They also stuck with 1-minute intervals for performing the corrections.  10 minutes of video at 30 fps contains 18000 frames.  With 29.97 fps they needed to drop 1/1000 of that, which is exactly 18 frames, or 1.8 frames a minute.  But again - we can't drop a fraction of a frame.

Exactly two frames are dropped each minute for the first 9 minutes, and no frames are dropped the 10th minute - repeat this continually, and you will drop 18 frames every 10 minutes.
 
sweet.. thanks for explaining.


so this is why when im ripping my home videos on vegas video the sound starts to go off track? that was another question I had. how do I fix that :( do i just manually delete 2 frames out of each minute of video for every 9 minutes of the frames then skip the 10th minute. then 11-19 and skip 20th and so on?

will that correct the problem?
 
Ok

im replacing this question. :D


in a video editor there is an option that says Every NTH frame" then an option to change the number 1.. does this have anything to do with drop frames? or is it this specific softwares jargan?

also when using divx it asks me something about "Keyframes every" then its set at 10 seconds not minutes.
 
I don't use that program, but it's probably a keyframe designation because all DVD is in MPEG2 format and requires keyframing. I've seen something similar in other MPEG compressor programs.
 
No. Drop-fame has to do with frame numbering. Keyframe has to do with encoding. See, most video is not encoded with the information of every single frame in order. Uncompressed video like that is very large and causes other problems, but is the best for editing because it preserves the highest picture quality. When encoding for distribution, there are more efficient ways of writing the information. The software that does this is called a codec. Some compress the information of each frame to reduce information and storage space. For instance a codec might take the binary sequence 10010000100100000100000000111 and encode it like 10014010015018011 - in computer terms. This type preserves all information and is called “lossless”. MPEG uses a completely different system where it designates 3 frame types. The Keyfame is a full frame with all the information. Then there is an Intraframe that records just the changes from the previous frame and the keyframe to that frame. So, anything that hasn’t changed is just copied from the keyframe and the previous frames. There is a Predictive frame. The codec says if a object is traveling from left to right at one pixel per frame, that at a certain number of frames from the keyframe the object will be at a certain point. It’s a predictive format that looks at future and past frames to determine the frame it is showing now. A sequence might look like K I I I P I I I P I I I K. The program is asking how often you want the keyframe. This makes it horrible for editing but great for distribution because it makes a relatively small file size. It’s a “lossy” format in that some picture information is lost, but the results are still acceptable.

Go to http://www.pixeltools.com/pixweb2.html for more information.
 
Oh OK. thanks for that information. I use alot of codecs so.

So when editing you dont want to use any codecs obviosly so as to be sure to get every frame you captured?

if i remember correctly standard video took up about 200 megs uncompressed per minute.
 
It’s not so much to make sure you get every frame you captured, but to capture at the highest possible quality. As imperfections are encountered, they become more and more noticeable with subsequent generations because every time it is recompressed, the imperfection is recompressed just like the rest of the signal. A compression artifact becomes very noticeable after 2 or 3 generations. The computer doesn’t know what’s an imperfection and picture, so it treats all information as true and correct.
 
I'm suddenly enlightened as to the 29.7fps myth!... cheers guys... but hey ;) I don't really have to worry about it because I live in the wonderful world of PAL! Pretty Awesome... erm... what could the L stand for?... er...
Anyway... 25fps!
 
Shot Renegade said:
I'm suddenly enlightened as to the 29.7fps myth!... cheers guys... but hey ;) I don't really have to worry about it because I live in the wonderful world of PAL! Pretty Awesome... erm... what could the L stand for?... er...
Anyway... 25fps!

Isnt PAL at a different resolution too?

It seems whenever I watch British shows its always soft and a little fuzzy. Not that it looked bad. but even the more modern shows before HDTV started taking over the shows still had a 1970s quality to the video. Like East Enders or AbFab.

do the british prefer a softer style look? and what about HDTV in Britian, is it using 25fps as well or will all video be universal speed in the future?
 
Zensteve said:
Regarding the Keyframes (and Div-x in particular), here is the official Div-x description of what keyframes are.

http://www.divx.com/support/guides/guide.php?gid=9

I don't understand the first half of the new Q...

in a video editor there is an option that says Every NTH frame" then an option to change the number 1

What is the mysterious video editor programme, and what were you doing when this option became available?

:P

Thanks Zen, now it makes sense. its like a bookmark of sort.

I knew MP3s (Which i guess is an audio layer of MPEG) uses the same method. where it deletes bits of sound that either cant be detected by the average human or the average human wont miss it. only a true audio editor or sound man could hear the lost of data.

so this should have been an obvious for me from the start. but thats a pretty cool concept to remove part of a frame thats just going to repeat itself.

and Its a pretty basic concept. I always thought it was far more Perplexing as to how it compressed.
 
ZEN -- So which do you prefer, Badger or Mushroom? I'm kind of partial to Mushroom today. Did you make that Flash, it is very disturbing, I love it!
 
Not mine, that's for sure. (Wish it was)

That's from the Weebl & Bob's site, the UK's most famous pair of eggs.

Here's a nice nod of his to Bergman... featuring that pie-loving pair.
 
Back
Top