Film to Video
M Turner
is taken care of in the UK and USA.
As I see it the situation is:
A cinema film is shot at 24 frames per second.
UK tv requires a frame rate of 25 fps.
Therefore when a film is shown on UK tv at 25 fps it is 4% faster.
In the US they require 30fps on their tv, if they showed a cinema film at
this speed it would be 20% faster. Is this so or do they do something else?
The second part of my question is about tv programmes made in the US and
shot on film.
What frame rate do they use for this?
Nick Jeffery
> This question is about the transfer of film to video and how the frame rate
> is taken care of in the UK and USA.
> As I see it the situation is:
> A cinema film is shot at 24 frames per second.
> UK tv requires a frame rate of 25 fps.
> Therefore when a film is shown on UK tv at 25 fps it is 4% faster.
Yes.
> In the US they require 30fps on their tv, if they showed a cinema film at
> this speed it would be 20% faster. Is this so or do they do something else?
http://www.zerocut.com/tech/pulldown.html
--
Nick Jeffery.
Stephen Neal
> This question is about the transfer of film to video and how the
> frame rate is taken care of in the UK and USA.
> As I see it the situation is:
> A cinema film is shot at 24 frames per second.
> UK tv requires a frame rate of 25 fps.
Yes - though it is worth remembering that UK TV is actually made up of 50
fields per second (each frame is made of two interlaced fields sent in
succession) Therefore each film frame is shown in 2 fields in the UK
> Therefore when a film is shown on UK tv at 25 fps it is 4% faster.
> In the US they require 30fps on their tv, if they showed a cinema
> film at this speed it would be 20% faster. Is this so or do they do
> something else? The second part of my question is about tv programmes
> made in the US and shot on film.
> What frame rate do they use for this?
The US TV system runs at 60 fields per second - again two fields per frame.
The US film transfer system runs the film at 24 frames/second (no speed
change), but shows one frame for 3 fields, the next for only 2 fields. (This
is known as 3:2 pulldown) This means that successive frames of film are
shown for different periods of time, causing motion judder in some cases.
It is also the reason that poor conversions of US TV series shot on film and
transferred to US video, then converted to UK video, can be poor. (Some
converters can detect the 3rd repeated field and remove it - others just get
confused by the artificial motion judder)
In some cases US TV series are shot on 30fps film - which can be transferred
to US TV more easily. However this is less common (well it costs 6 more
frames per second to shoot and process!)
Steve
Martin Underwood
news:mfssbb...@wschwanke.de...
> "M Turner" <m.tu...@btinternet.com> wrote in
> news:bbsjfa$n6r$1...@titan.btinternet.com:
>
> > The second part of my question is about tv programmes
> > made in the US and shot on film.
> > What frame rate do they use for this?
>
Presumably only two frame rates make any sense for them to use: 24 fps with
3:2 pulldown and 30 fps shown 2 fields/film frame (as we do for made-for-TV
film, except we use 25 fps). Why would any frame rate except 24 or 30 be
used? (I'm assuming that we're talking about natural-speed motion, not
slow-motion or speeded-up!)
Stephen Neal
Some BBC / US Co-productions have been shot on HD at 25fps - this is shown
at 50fps in UK territories and at 60fps using 3:2 pulldown but with the
25fps material slowed down to 24fps beforehand. No idea if this is because
the UK "won" in the fps contest, or if 25fps HD kit was more easily
available than 24fps kit at the time. (HDTV can be shot at 24,25,30 fps
progressive, as well as 50/60 fields/sec interlaced, and 50/60fps
progressive)
Steve
BillJ
of 4%, in semitones (or commas)?
Thanks,
BillJ, Edinburgh
Stephen Neal
> In article <vts4ev88kgardr8es...@4ax.com>,
> BillJ wrote:
>> As a matter of interest (musically speaking), what's a speed increase
>> of 4%, in semitones (or commas)?
>
I believe that some 24->25fps transfers also utilise a harmoniser or similar
device - so that the speed change does not cause the soundtrack to change
pitch - just speed.
Steve
charles.hope
BillJ <g...@8apx.freeserve.co.uk> wrote:
> As a matter of interest (musically speaking), what's a speed increase
> of 4%, in semitones (or commas)?
Obviously enough to upset people with "perfect pitch". I remember taking
a call (in the '60s) from an irate viewer who was upset becasue the music
in the film was nearly a semi-tone sharp.
SpamTrapSeeSig
<REMOVE.use...@wschwanke.de> writes
>> rate is taken care of in the UK and USA.
>> As I see it the situation is:
>> A cinema film is shot at 24 frames per second.
>> UK tv requires a frame rate of 25 fps.
>> Therefore when a film is shown on UK tv at 25 fps it is 4% faster.
>
>
>When dealing with different frame rates, the preferred method is frame rate
>conversion. The resulting material then has the exact speed and time length
>as the original, which is after all what you want. The process works by
>doubling or dropping frames, which introduces artefacts (jumpy motion).
Not necessarily.
A polygonal prism telecine was the preferred method for many years. It
'detaches' the film speed from the scanning frame rate (works at any
film speed).
They're very hard to implement well (optically complex), and all the
ones I've seen have been pretty soft, but they *do* work. BBC Bristol
had a 'broadcast-quality' one used for regional TX in the 1970s and
1980s, and I think there were quite a few around the UK. We also had a
non-broadcast one in the dubbing theatre (but that was ghastly!).
I've been told the polyprism was the preferred approach before 3:2
pulldown in the states, and the nasty conversion problem was why we got
flying-spot before they did, but I've no idea if it's true.
Regards,
Simonm.
--
simonm|at|muircom|dot|demon|.|c|oh|dot|u|kay
SIMON MUIR, UK INDEPENDENCE PARTY, BRISTOL www.ukip.org
EUROPEANS AGAINST THE EU www.members.aol.com/eurofaq
GT250A'76 R80/RT '86 www.kc3ltd.co.uk/profile/eurofollie/
M. J. Powell
<no-...@nospam.demon.co.uk> writes
They were commonplace when I started In TV in 1958. Most made by Cintel.
Mike
--
M.J.Powell
-
> Some BBC / US Co-productions have been shot on HD at 25fps - this is shown
> at 50fps in UK territories and at 60fps using 3:2 pulldown but with the
> 25fps material slowed down to 24fps beforehand.
How is this slowdown from 25 to 24 fps done? Surely you still have 25
source frames every second to be converted to 30.
or is some other more complicated pulldown ratio used?
Martin Underwood
news:crYNpdDS...@muircom.demon.co.uk...
> In article <mfssbb...@wschwanke.de>, Wolfgang Schwanke
> <REMOVE.use...@wschwanke.de> writes
> >"M Turner" <m.tu...@btinternet.com> wrote in
> >news:bbsjfa$n6r$1...@titan.btinternet.com:
> >
> >> This question is about the transfer of film to video and how the frame
> >> rate is taken care of in the UK and USA.
> >> As I see it the situation is:
> >> A cinema film is shot at 24 frames per second.
> >> UK tv requires a frame rate of 25 fps.
> >> Therefore when a film is shown on UK tv at 25 fps it is 4% faster.
> >When dealing with different frame rates, the preferred method is frame
rate
> >conversion. The resulting material then has the exact speed and time
length
> >as the original, which is after all what you want. The process works by
> >doubling or dropping frames, which introduces artefacts (jumpy motion).
>
> Not necessarily.
>
> A polygonal prism telecine was the preferred method for many years. It
> 'detaches' the film speed from the scanning frame rate (works at any
> film speed).
I've heard of this type of device but I've always wondered why it didn't
cause visible "beating" at the difference frequency between projection speed
and scanning speed (TV frame rate) - eg at 1 Hz for 24 fps film and 25 fps
TV. I presume it leads to some fields which are a mixture of the bottom of
one film frame and the top of the next, which is unlike all other telecine
material where the odd and even fields are always taken from the same film
frame.
What is the current technology that's used in the UK? Do they project at 24
fps as you could for the polygonal prism or do they project at 25 fps?
Steve Roberts
>"Stephen Neal" <stephe...@NOSPAM.as-directed.com> wrote in message news:<bbt4oc$lkt$1$830f...@news.demon.co.uk>...
>> Some BBC / US Co-productions have been shot on HD at 25fps - this is shown
>> at 50fps in UK territories and at 60fps using 3:2 pulldown but with the
>> 25fps material slowed down to 24fps beforehand.
>
>How is this slowdown from 25 to 24 fps done? Surely you still have 25
>source frames every second to be converted to 30.
You use a VTR to play out the 25 fps tape slower than normal, so it's
producing 24fps. Then you use an external box to add 3:2 pull down to
generate 60 fields / per second.
Steve
The Doctor Who Restoration Team Website
http://www.restoration-team.co.uk
Steve Roberts
<martin.u...@virgin.NO-SPAM.net> wrote:
[Polygonal prism telecine]
>I've heard of this type of device but I've always wondered why it didn't
>cause visible "beating" at the difference frequency between projection speed
>and scanning speed (TV frame rate) - eg at 1 Hz for 24 fps film and 25 fps
>TV. I presume it leads to some fields which are a mixture of the bottom of
>one film frame and the top of the next, which is unlike all other telecine
>material where the odd and even fields are always taken from the same film
>frame.
It doesn't work like that. The prism optically mixes variable amounts
of light from the adjacent frames, so you get a mix between frames,
rather than the top of one frame with the bottom of another etc.
This leads to seveal unpleasant artefacts, as I've found out
remastering various Doctor Who episodes! Even running at locked speed,
you find one sharp field followed by one mixed field, so the motion
takes on an almost stands-converted look. If the telecine was locked,
you can get round this by basically filmising the material to double
up the sharp field and throw away the soft - generally the pictures
are so soft anyway that you don't lose any resolution.
The real problems occur when the telecine is running unlocked, so that
sharp field drifts over every so often to the other field and then
back again cyclically! And of course it's never running steady enough
to make the cycle length predictable!
We invented a software solution we called the 'auto-depolygoniser',
running on Inferno to help us with this. It looks at field pairs and
uses a cunning method to work out which is the sharper field and makes
up a new frame from that field. Works pretty well!
Martin Underwood
"Steve Roberts" <stever...@compuserve.com> wrote in message
news:3ee469cf...@news.reith.bbc.co.uk...
> On 9 Jun 2003 02:05:40 -0700, cbd...@my-deja.com (-) wrote:
>
> >"Stephen Neal" <stephe...@NOSPAM.as-directed.com> wrote in message
news:<bbt4oc$lkt$1$830f...@news.demon.co.uk>...
> >> Some BBC / US Co-productions have been shot on HD at 25fps - this is
shown
> >> at 50fps in UK territories and at 60fps using 3:2 pulldown but with the
> >> 25fps material slowed down to 24fps beforehand.
> >
> >How is this slowdown from 25 to 24 fps done? Surely you still have 25
> >source frames every second to be converted to 30.
>
> You use a VTR to play out the 25 fps tape slower than normal, so it's
> producing 24fps. Then you use an external box to add 3:2 pull down to
> generate 60 fields / per second.
Blimey! That would result in severely non-625/25 signals because the line
frequency would also be reduced a bit as well. But this may not matter
because it would be relatively minor compared with the need to convert 625
lines to 525 lines, as long as the interpolation algorithm can cope with
slightly non-standard signals.
Steve Roberts
Well, as we're talking about Hi-Def, 525/625 isn't a problem here! :)
It does result in completely non-standard frequencies, although it's
all done in digits anyway.
However, the same basic system is used for standard def as well, using
a modified Digibeta or indeed the various Panasonic systems. Panasonic
do a nice format converter for this application:-
http://www.panasonic.com/PBDS/subcat/Products/formatconver/f_aj-ufc1800.html
Tony Quinn
Underwood <martin.u...@virgin.NO-SPAM.net> writes
>> A polygonal prism telecine was the preferred method for many years. It
>> 'detaches' the film speed from the scanning frame rate (works at any
>> film speed).
>
>I've heard of this type of device but I've always wondered why it didn't
>cause visible "beating" at the difference frequency between projection speed
>and scanning speed (TV frame rate) - eg at 1 Hz for 24 fps film and 25 fps
>TV. I presume it leads to some fields which are a mixture of the bottom of
>one film frame and the top of the next, which is unlike all other telecine
>material where the odd and even fields are always taken from the same film
>frame.
it doesn't work quite like that, the prism effectively mixes between
frames (I've got my Evesham notes somewhere) when I was in London we had
2 16 mm and 2 35mm units - one of the 35mm suites was Nitrate capable
(at least when I last worked at the Grove in 83).
------------------------------------------------------------------------
Tony Quinn --- The Voice Of Insanity
replies to to...@sixpints.demon.co.uk
------------------------------------------------------------------------
SpamTrapSeeSig
<to...@sixpints.demon.co.uk> writes
>one of the 35mm suites was Nitrate capable (at least when I last worked
>at the Grove in 83).
In summer '84, when I was there for a brief period (the 'Starship' and
Theatre B as an assistant dubbing mixer), the nitrate vaults were still
up on the roof - great spot for sunbathing...
Chris Packman
Tony Quinn <to...@sixpints.demon.co.uk> wrote:
> In article <QyZEa.2319$M56.4...@newsfep2-win.server.ntli.net>, Martin
> Underwood <martin.u...@virgin.NO-SPAM.net> writes
> >> A polygonal prism telecine was the preferred method for many years. It
> >> 'detaches' the film speed from the scanning frame rate (works at any
> >> film speed).
> >
> >I've heard of this type of device but I've always wondered why it didn't
> >cause visible "beating" at the difference frequency between projection speed
> >and scanning speed (TV frame rate) - eg at 1 Hz for 24 fps film and 25 fps
> >TV. I presume it leads to some fields which are a mixture of the bottom of
> >one film frame and the top of the next, which is unlike all other telecine
> >material where the odd and even fields are always taken from the same film
> >frame.
>
> it doesn't work quite like that, the prism effectively mixes between
> frames (I've got my Evesham notes somewhere) when I was in London we had
> 2 16 mm and 2 35mm units - one of the 35mm suites was Nitrate capable
> (at least when I last worked at the Grove in 83).
The 35mm machines at the Grove survived until the end of Nationwide, two
of the four Lime Grove 16mm twin lens machines survived for about a
year into London Plus. AFAIK all six machines were scrapped about a
year after Nationwide ended. The nitrate facility was retained by moving
it to MK 3 machines at the Centre, first in the basement and then into
stage 5. Safety regulations changed and we dropped the facility in the
early 90s.
The two 16mm polygon machines at the Centre survived until about 1986
and 1988. The first one withdrawn was sold I think to the Imperial
War Museeum. The second removed from service was scrapped as it was
in an area hevily contaminated with asbestos.
Polygons were great fun. On a good day they could produce good pictures
(in the case of the 16mm machines in both 525 and 625). Sadly on an
off day they were crap ! (all soft and wobbly).
Chris Packman
--
Ariel Radio Group Website http://www.xdinet.demon.co.uk/ARG/
Railway Pictures http://www.xdinet.demon.co.uk/rail/
Stephen Neal
611952a3.03060...@posting.google.com, "-" <cbd...@my-deja.com>
wrote:
Not really - the slow down from 25fps to 24fps means you have 24 frames per
second to convert to 30 (in an identical manner to film), so 3:2 pulldown
can be employed. The slowdown just means that the programme runs for a
longer duration.
Steve
Stephen Neal
Um_Ea.2499$M56.5...@newsfep2-win.server.ntli.net, "Martin Underwood"
<martin.u...@virgin.NO-SPAM.net> wrote:
Yep - though this is all happening in the digital domain - whether SD or HD
- so not such a problem changing the line and field frequencies as would be
the case with composite subcarrier based formats. Not sure how the SD
signal is carried over SDI...
I believe that in the case of SD material, a 625/25 (50i) DigiBeta runs
slower, outputting a 625/24 (48i) signal. This is then downconverted to
525/24 (48i) and 3:2 pulldown applied to create a 625/60i signal. I believe
that this processing is done externally in an external box. (In fact the
DigiBeta VTR may run at 23.97ish frames/sec to cope with 525/60 actually
running at 525/59.94)
In HDTV it is easier, as the vertical resolution remains at 1080 lines, so
all that is required is for 1080/25 material to be converted to 1080/24 by
slower playback, then 3:2 being added.
In both HD and SD cases this can easily be applied using non-linear
hard-disc based editing. Quantel have been espousing shooting film at
24fps, telecineing at 625/25 (with a speed up), then treating this as
625/24fps for non-linear editing purposes, then playing this out at 625/25
for 50Hz regions, and 525/60 with 3:2 pull-down for 60Hz regions. The same
edit master can be used for both regions, and the 50Hz regions benefit from
625 resolution.
Steve
Martin Underwood
news:BB0B70F0.9FD7%stephe...@as-directed.com...
> On 9/6/03 12:40 pm, in article
> Um_Ea.2499$M56.5...@newsfep2-win.server.ntli.net, "Martin Underwood"
> <martin.u...@virgin.NO-SPAM.net> wrote:
>
> In HDTV it is easier, as the vertical resolution remains at 1080 lines, so
> all that is required is for 1080/25 material to be converted to 1080/24 by
> slower playback, then 3:2 being added.
Oh, do both the European and US versions of HDTV use the *same* number of
lines? I hadn't realised that: I'd assumed that European HDTV was 1250 (2x
625) and US HDTV was 1050 (2x 525) so that HDTVs could display SDTV by
simply dividing the relevant frequency of the HDTV line oscillator by two
and broadcasters could derive an SDTV signal from an HDTV source by simple
averaging of two adjacent lines.
How many visible lines does the 1080 system have? I presume as technology
has improved, the need for such long line and field blanking periods has
been removed and shorter blanking times are now possible. I suppose it's too
much expect that the digital versions of it would have 1024x768 pixels (to
match the computer screen resolution). For widescreen, does HDTV have square
pixels or is it the same as for SDTV where the horizontal resolution is much
less than the vertical and the picture is stretched laterally to fill the
screen?
When (if ever) will we get HDTV in Britain? And will it ever be available by
terrestrial aerial (as opposed to satellite) for those poor sods like me who
aren't allowed to have a satellite dish and who haven't got cable?
Jim Guthrie
wrote:
Bill,
>As a matter of interest (musically speaking), what's a speed increase
>of 4%, in semitones (or commas)?
According to my SADiE when I do 24<>25 resamplings, 0.71 of a
semitone.
Jim.
Stephen Neal
SWkFa.2631$Cm5.6...@newsfep2-win.server.ntli.net, "Martin Underwood"
<martin.u...@virgin.NO-SPAM.net> wrote:
> "Stephen Neal" <stephe...@as-directed.com> wrote in message
> news:BB0B70F0.9FD7%stephe...@as-directed.com...
>> On 9/6/03 12:40 pm, in article
>> Um_Ea.2499$M56.5...@newsfep2-win.server.ntli.net, "Martin Underwood"
>> <martin.u...@virgin.NO-SPAM.net> wrote:
>>
>> In HDTV it is easier, as the vertical resolution remains at 1080 lines, so
>> all that is required is for 1080/25 material to be converted to 1080/24 by
>> slower playback, then 3:2 being added.
>
> Oh, do both the European and US versions of HDTV use the *same* number of
> lines? I hadn't realised that: I'd assumed that European HDTV was 1250 (2x
> 625) and US HDTV was 1050 (2x 525) so that HDTVs could display SDTV by
> simply dividing the relevant frequency of the HDTV line oscillator by two
> and broadcasters could derive an SDTV signal from an HDTV source by simple
> averaging of two adjacent lines.
The current flavours of HD include standards for both 50 and 60Hz regions
with the same source image resolution.
There are 1080 line 24,25,30,50 and 60 frame/field standards. These are
based on 1920 horizontal x 1080 vertical images. (There are also 1440 x 1080
variants I believe where a lower horizontal resolution is acceptable)
The 1080 refers to the number of active lines - not the total number of scan
lines including blanking. This is because digital transmission systems
don't have to transmit the blanking lines.
There were 1050, 1125 and 1250 line systems - the first two were 60Hz
interlaced, the third one a 50Hz standard. These number of lines refer to
the total number of lines, not the active lines. ISTR that 1125 was the
most common format (it was the Japanese HD format that was once broadcast on
an analogue system) and the bulk of HD production equipment was built to
this. The 1125 format supported 1035 or 1050 active lines I think - so was
quite close to the 1080 formats now in favour.
The 1050 and 1250 (total not active) line standards were espoused when
analogue transmission was still thought likely - the European analogue HDMAC
system was based on 1250 lines (it was broadcast as an analogue 625 line
signal, with digital assistance data that allowed separate blocks to be
effectively interlaced over differing numbers of fields. Static pictures
info was interlaced over 4 rather than 2 fields to create a 1250 line image)
The US HDTV transmission systems also support 720line standards, but these
can run at 60 FRAMES/second progressive.
>
> How many visible lines does the 1080 system have?
1080 - the spec is based on active not total lines.
(In the same way that 625 material is now commonly referred to as 576 - the
number of active lines in the digital system)
> I presume as technology
> has improved, the need for such long line and field blanking periods has
> been removed and shorter blanking times are now possible. I suppose it's too
> much expect that the digital versions of it would have 1024x768 pixels (to
> match the computer screen resolution). For widescreen, does HDTV have square
> pixels or is it the same as for SDTV where the horizontal resolution is much
> less than the vertical and the picture is stretched laterally to fill the
> screen?
>
I believe that the 1920x1080 image structure is based on sqaure pixels - to
give a 16:9 image.
> When (if ever) will we get HDTV in Britain? And will it ever be available by
> terrestrial aerial (as opposed to satellite) for those poor sods like me who
> aren't allowed to have a satellite dish and who haven't got cable?
>
Well European HDTV services are being tested at the moment. Alfacam - a
European facilities company that have a number of HD trucks - are I believe
working with a company to provide a Europe-wide satellite HDTV service. No
idea whether it will take off. I imagine that unless US HD receivers can be
used it will be quite an expensive hardware purchase?
Not sure if the Alfacam broadcasts will be 1080/50 or 1080/60 - if the
latter then I guess US kit may be usable?
Steve
BillJ
BillJ (Edinburgh)
On Sun, 8 Jun 2003 11:34:42 +0100, "Stephen Neal"
<stephe...@NOSPAM.as-directed.com> wrote:
Nick J
> On Sun, 8 Jun 2003 11:34:42 +0100, "Stephen Neal"
> <stephe...@NOSPAM.as-directed.com> wrote:
>
>>Paul Martin wrote:
>
>>I believe that some 24->25fps transfers also utilise a harmoniser or similar
>>device - so that the speed change does not cause the soundtrack to change
>>pitch - just speed.
>
> Hey, neat...any clues how that works?
My audio editing package of choice has a facility to raise and lower by
n semitones without affecting the speed. It's quite a quick operation
to perform, so I can't imagine the processing required is too difficult.
I used it recently on a voice-over to make it sound more lively and
upbeat. Or something. :)
--
Nick J.
Laurence Taylor
On Wednesday, in article
<5ppcev8ainvh1bhoj...@4ax.com>
g...@8apx.freeserve.co.uk "BillJ" wrote:
B> On Sun, 8 Jun 2003 11:34:42 +0100, "Stephen Neal"
B> <stephe...@NOSPAM.as-directed.com> wrote:
B>
B> >Paul Martin wrote:
B>
B> >
B> >I believe that some 24->25fps transfers also utilise a harmoniser or similar
B> >device - so that the speed change does not cause the soundtrack to change
B> >pitch - just speed.
B> Hey, neat...any clues how that works?
With analogue, it can be easily done by heterodyning the input with a
controlled oscillator. I remember many years ago a JVC VCR did this,
to give correct pitch sound on double speed playback.
I should think a digital circuit could be made to do it with no problem.
rgds
LAurence
-------------------------------------------------------------------
This is an Honour Virus. Please pass it on, then format your drive.
===================================================================
->> This message produced entirely in DOS <<-
Laurence Taylor
<slrnbej6...@laptop.nowster.org.uk> p...@zetnet.net
"Paul Martin" wrote:
PM> > With analogue, it can be easily done by heterodyning the input with a
PM> > controlled oscillator. I remember many years ago a JVC VCR did this,
PM> > to give correct pitch sound on double speed playback.
PM>
PM> I wouldn't have thought that would work.
PM>
PM> A heterodyne shifts frequency by a constant offset, whereas a speed
PM> change is a constant multiplier.
It could be done ... take the original frequency, multiply it by some
suitable amount, devide it by a different suitable amount to get what
you want. I thought I'd worked it out before I wrote that, but I can't
remember what I though of now! The incoming would have to be used to
control the local oscillator, anyway.
How did JVC do it? The model was a HR7700/3V23, I think. Huge beast,
weighed half a ton.
rgds
LAurence
-------------------------------------------------------------------
This message has been ROT-13 encrypted twice for extra security