# Fields per second vs. Frames per second

Discussion in 'Electronic Basics' started by Radium, Oct 21, 2006.

1. ### JamieGuest

i did, and it's correct.
currently my clock is saying 10:25 PM

2. ### Rich GriseGuest

Only by virtue of the fact that a video frame is delivered half at a time.

A film frame is exactly where they got the name for a video frame.

They're at slightly different rates, typically film is 24 FPS, and US
TV is 30 FPS, but other than that and the interlace, the defninition of
what a "frame" _is_ is practically identical.

You might be interested to know that 24FPS movie projectors actually
project each frame twice, so you get an effective flicker rate of 48
FPS.

Cheers!
Rich

3. ### Gene E. BlochGuest

Well, to be exact, the two fields of video are not quite the same as
the three frames (not two, as I recall it - but it probably varies from
projector to projector) of a film.

The two or three frames (or call it fields - why not?) of the film are
identical to each other, since they are just the identical piece of
film projected again. If there is any motion in the subject, the two
fields of the video are different from each other, since they are taken
at different times.

I'm just repeating Jukka Aho's point here for emphasis.

4. ### Jukka AhoGuest

No, in other respects, too. Didn't you read the text you quoted above?
US TV is 60 * 1000/1001 fields per second. The picture will update 60 *
1000/1001 (~59.94) times per second. If a moving object - say, a ball
rolling across the screen from left to right - is shot with an NTSC
video camera, its motion (location on the screen) updates 60 * 1000/1001
(~59.94) times per second. (That is, if we're still talking about
material that was shot with a regular tv/video camera running in
interlaced mode.) The location of the rolling ball differs from field to
field, not only from frame to frame.
It's not. Fields are shot at different times. When two adjacent fields
are paired and combined into a single video frame, the end result will
contain material from two different instants of time. A film frame only
contains material from a single instant of time. (This is the reason why
motion recorded with a video camera looks a lot smoother than motion
shot on 24 fps film. Video _is_ smoother, by a factor of (about) 2.5.)
It was interesting when I first heard about it. But that was a long time
ago.

5. ### EeyoreGuest

What does it matter to you ?

You'll just misunderstand it because you're so damn stupid.

Graham

6. ### Michael A. TerrellGuest

As usual, the greaseball is wrong. A TV projector in a film chain has
a special shutter that runs the 24 FPS film to match the 30 FPS video
rate by showing every 4th frame twice:

Frames:

Video Film

01 01
02 02
03 03
04 04
05 04

06 05
07 06
08 07
09 08
10 09

11 09
12 10
13 11
14 12
15 12

16 13
17 14
18 15
19 16
20 16

21 17
22 18
23 19
24 20
25 20

26 21
27 22
28 23
29 24
30 24

This was done on the RCA TP-66 film chain projectors used by TV

--
Service to my country? Been there, Done that, and I've got my DD214 to
prove it.
Member of DAV #85.

Michael A. Terrell
Central Florida

8. ### Bob MyersGuest

[example deleted]
Hate to have to disagree with you, Michael, but that's no longer the
way it's done, the TP-66 method notwithstanding.

In modern practice, 24 FPS film is actually run at ~23.97 FPS
(to enable a match to the 59.94+ Hz "NTSC" field rate), and then
converted to video using a method generally known as "3:2 pulldown,"
in which frame films are alternately captured as either 3 or 2 video
fields (i.e., one frame of film winds up as 1.5 frames of video, while
the next winds up as 1.0 frames). An example: if we have four
successive film frames, A, B, C, and D, the resulting pattern of
video FIELDS (not frames) would then contain the following:

A A A B B C C C D D ...

and so forth. While this still introduces errors (motion artifacts)
into the resulting video stream, the errors are in general more
acceptable than those which resulted from the earlier frame-
doubling process.

The original point in all this, though, I believe has been missed.
TV "frames" are in almost all cases almost a fiction from the standpoint
of image content; we speak of them only because, with the 2:1 interlaced
scanning format, it's the only way to justify talking about the effective
vertical resolution of the system (at least for still images). The two
fields, however, DO represent different sample points in time, and
therefore cannot in reality be assembled to produce a complete
frame of the resolution (or line count) that one would expect, if there
is any motion in the scene. (Moving objects, of course, will appear
in slightly different positions between the two fields.) This is one of
the factors that reduces the effective as-delivered resolution of an
interlaced system.

Bob M.

9. ### Richard CrowleyGuest

Perhaps that is why Mr Terrell used the past-tense "was"?
OTOH, his "greaseball" remark raises his "plonkability"
score on my end.

10. ### Jukka AhoGuest

Michael A. Terrell wrote:

[awkward display of personal animosity snipped]
The OP's question was "PAL video system uses 50 fields per second and 25
frames per second. Whats the difference between 'field' and 'frame'?" He
did not further specify the source or acquisition method, much less
describe any practical situation from where it could be derived that
film was the likely source medium. Keeping that in mind (and observing
that the question was posted to "rec.video.desktop" which mostly deals
with home and hobbyist video shot with home and prosumer video cameras),
a reasonable assumption is that we're talking about "video" in its
purest and most original form: interlaced video shot with an interlacing
video camera.

My comments in this thread have been based on that basic assumption
(interlaced video shot with an interlacing video camera), and I have
also explicitly _restated_ that assumption in a couple of my posts,
precisely because there are some other acquisition and post-processing
methods that can deprive video of its natural, smooth 50 Hz or ~60 Hz
temporal resolution, or underutilize that capability.

(Followups set to "rec.video.desktop" once again.)

11. ### Bob MyersGuest

And I'm thinking at this point that we still haven't answered that
particular question very well, so I'm going to take a stab at it.

In the most general sense I can think of, the word "frame" refers
to the smallest unit in a motion image stream (a clumsy phrase,
perhaps, but I'm trying to cover both film and video usage
here) which includes ALL of the information for a single,
complete image (i.e., it contains all the color, luminance, etc.,
information for one image at the full resolution of the system).
Ideally, it represents one temporal sample - i.e., it is an image
which is captured at one particular point in time, in a series of
such images which together are used to give the illusion of
motion when displayed. This last point is where the notion of
a "frame" in video starts to break down, since clearly in a
raster-scanned system the entire image is not captured at the
same time. But we still use the term nonetheless.

A "field," on the other hand, is some defined sub-part of a
frame. In the most common usage of this term - interlaced
video, which is what we're talking about here - two "fields"
are produced, which are supposed to correspond to (or
be capable of being combined into) one complete frame.
Again, it's not quite that simple, since in interlaced video the
two fields are generally produced as such, separately, by the
camera or telecine, as opposed to actually being the result
of separating the odd and even lines of an original complete
frame. This isn't the only possible use of the term; for
instance, in "field-sequential" color systems (as in the case
of the original CBS color system approved by the FCC in
the early 1950s), the full-color frame is separated not into
odd- and even-line fields but rather into red, green, and blue
fields. These, when shown in rapid succession, restore the
appearance of a full-color image.
Hope you don't mind my adding sci.electronics.basics back
in, as many of us (myself included) are following the thread
there.

Bob M.

12. ### Michael A. TerrellGuest

Bob, since the video camera scans both fields from the same frame of
film, there is no "Difference in Time". The TP 66 could also run the
color scan rates, and did in most installations. The pair that I
maintained and ran happened to be used at a B&W station, but I've seen
them in use at color stations.

As far as slight differences, the persistence of the human eye tends
to average out the minor differences. Most people can't see a single
frame that doesn't match the film content, yet I could see the single
"Insert Commercial here" frames in the AFRTS 16 mm films we ran. No one
at the station believed me, so they opened a case of films that had just
arrived, and loaded one. I hit stop when I saw it, and a few frames from
where the shutter stopped, they saw it. I did it on six reels before
they believed me.

You may not like the "Frame" concept, but I find "Progressive Scan"
to be stupid. Progressive? they went back to the earliest video scan
method, and have the nerve to call it "Progressive"?

The NTSC developers were not concerned about minor differences
between fields, and unless a scene has a lot of motion, there may be no
difference other than the offset spacing. Some talking head is on the
screen, with nothing but their lips moving, and an occasional blink. Who
cares about minor timing errors you can't see?

--
Service to my country? Been there, Done that, and I've got my DD214 to
prove it.
Member of DAV #85.

Michael A. Terrell
Central Florida

13. ### Michael A. TerrellGuest

I am tired of all his rants and preaching alchol and nicotine
addiction and pushing a religious cult. I have 18 different screen names
of his kill filtered, and still have to add new ones a couple times a
year.

If you want to "Plonk" me, go ahead.

--
Service to my country? Been there, Done that, and I've got my DD214 to
prove it.
Member of DAV #85.

Michael A. Terrell
Central Florida

14. ### Jukka AhoGuest

That was a fairly good explanation, assuming the OP didn't know what a
"frame" was in the first place.
Originally - with tube cameras - none of the image was captured at the
same time. Instead, it was a continuous process where the tube camera
scanned the scene line by line, sweeping across the length of each
scanline in lock-step with the tv sets painting the same picture on
screen at homes. Essentially, each point in the picture was captured at
a different time. There is a good illustration of this near the bottom
Simple":

<http://lurkertech.com/lg/fields/fields.html#reality>

Telecine processes and modern CCD-based video cameras of course do not
produce images this way, but it is still useful to keep in mind that
video cameras originally did - as late as in the 1980s. The original
reality of "video" - and its "temporal dimension", so to speak - was
fundamentally different from that of any frame-based system.

Over the time, video has _become_ more like film, though - first by
ditching the "scanning" cameras in favor of CCD-based ones, and now by
introducing "progressive-scan" [1] HD formats, where video frames really
_are_ frames, in the film sense.
I don't mind if that's OK with the rest of the people in s.e.b. I just
thought this subject might be a bit too off-topic there, and better
suited for r.v.d only.

_____

[1] "Progressive scan" is becoming more and more of a misnomer in
itself. These days, usually neither the camera producing the images -
and, increasingly, not even the display device - "scans" any longer.

15. ### Jukka AhoGuest

The "greaseball" remark was (seemingly) posted in reference to me in
your post <Given the context in
which that was remark was uttered, it's quite hard to interpret it in
any other way.

From the sound of the segment quoted above, it however now appears that
you have me confused with the OP, who has displayed some trollish
behavior in some of the other recent threads, and whom you specifically
claimed to be a troll earlier in this thread (with much of the same kind
of accusations about his earlier behavior.) We actually agreed on his
trollishness, if you can still remember.

I have never used anything but my own name on Usenet. You can check out
my posting history from the Google Groups Usenet archives:

CQBxjOrkrNYwvJW2bfk89k>

I have also no recollection of ever discussing alcohol, religious cults,
or smoking on Usenet at any great length, other than what could be made
in passing reference during a normal discussion. (I'm not much of a
drinker, I don't smoke, and if anything, I would preach against
religious cults rather than for them.)

(I hope the above-mentioned URL is not temporary, but in case it is,
look up any message by me from the Google Groups archives and click on

16. ### Gene E. BlochGuest

However, the "greaseball" remark was in a direct reply to my post in
this part of the thread, so I assumed he meant me. Maybe he meant you,
maybe he meant Radium, I don't know. Also consider that the post to
which I replied was from someone name Rich Grise; maybe Terrell thinks
his name is pronounced like 'grease'.

If he meant me, it is in truth the most complimentary remark I've

As for rest of your remarks below, it happens that I am in complete
compliance & agreement (except for the URL, of course).

It also happens that my remarks about movie projector shutters were in
the context of projecting for view, not in the context of film chains
for converting to video.

18. ### Gene E. BlochGuest

Malleum mortuorum.

19. ### Jukka AhoGuest

Hmm, yes, I can see now how there are other possible ways to interpret
it. But, my name was mentioned in the quoted segment just above that
comment, and the comment seemed to be in some sort of disagreement with
what I had said. Or what you said I had said, and seemed to agree with,
if you know what I mean. (This is the sort of confusion we get when
people quote entire messages without trimming their quotes to the
point.)
Perhaps he meant himself. There are certain archetypical categories of
people who are said to refer to themselves in third person, and
self-hate is not a too uncommon disorder. It could be taken as some sort
of self-irony as well.

20. ### Michael A. TerrellGuest

It wasn't amied at you, but the person you replied to. (grice) He is
kill filed under 18 aliases, and keeps adding new ones.

--
Service to my country? Been there, Done that, and I've got my DD214 to
prove it.
Member of DAV #85.

Michael A. Terrell
Central Florida