Questions on Digital Motion Pictures
Karl McBurnett
Lucas has been having his team of gnomes investigate the possibility of
filming one of his future films in a digital format. While the idea
intrigues me from a purely "gee-whiz" point of view, I find it hard to
believe that some gussied-up camcorder could begin to approach the quality
of plain old-fashioned film. So, I've got a few questions about what a
digital motion picture would have to achieve to reach 35mm standards:
1) First off, if you know of any web-sites that covers this subject in
detail, that would be helpful.
2) What kind of resolution (number of points/pixels) would a digital film
need to reach to equal a frame of 35mm film.
3) How many unique colors would a digital camera need to be able to discern
to equal 35mm film.
4) I guess most important would be how would a digital film keep from
looking like all other videos - overly crisp, without the "warmth" of film.
I'll just thank anyone with answers ahead of time. Thanks :)
Karl McBurnett
joe
digital camera is something we are all working on... It's just a matter of
data rate and storage. Film scanners, digital cameras and the like... film
is really on its way out.
companies with $$$ will shoot first and then the tech will flow downwards...
Karl McBurnett wrote in message <#muM9LCW#GA.132@upnetnews03>...
>I read an article a little over two months ago talking about how George
>Lucas has been having his team of gnomes investigate the possibility of
>filming one of his future films in a digital format. While the idea
>intrigues me from a purely "gee-whiz" point of view, I find it hard to
>believe that some gussied-up camcorder could begin to approach the quality
>of plain old-fashioned film. So, I've got a few questions about what a
>digital motion picture would have to achieve to reach 35mm standards:
>
>1) First off, if you know of any web-sites that covers this subject in
>detail, that would be helpful.
>
>2) What kind of resolution (number of points/pixels) would a digital film
>need to reach to equal a frame of 35mm film.
typical size is 2048x1556 for 35... probably start with at least that rez...
>
>3) How many unique colors would a digital camera need to be able to discern
>to equal 35mm film.
16 bits per color.
>
>4) I guess most important would be how would a digital film keep from
>looking like all other videos - overly crisp, without the "warmth" of film.
that crisp warmth of film will just be a thing of the past....and replicated
and changed and re-invented with plugins and processes.....
>
>I'll just thank anyone with answers ahead of time. Thanks :)
you're welcome! and remember... there's no turning back. film is dead.
>
>
>Karl McBurnett
>
>
Jesus M. Medina
>
I would not dismiss film just yet. Not to sound like a smart ass, but for a
medium that is dead, it still looks pretty good, and drives the industry like a
train.
Just my $.02
Regards.
jon springer
>
>
> >
> >
> >Karl McBurnett
> >
> >
>
> Digital motion pictures (especially for theatrical release) won't become
viable until a mass audience can accept the attributes of a medium that
will always be different and never equal to film--because it is a different
medium. It's hard to breed 100 years of film viewing out of the asthetic
subconscience of human beings, and believe me George Lucas is not going
to do it with a souped-up video camera in the year 2005!
No matter how much resolution or color rendering is achieved, it will
always be different because it is an electronic medium. It is lifeless
where as film is organic. No matter how good digital acquisition is
it still looks like the six o'clock news--it's still video.
--Jon Springer
joe
film is dead.
Jesus M. Medina wrote in message <36C770D1...@americanisp.com>...
>> and remember... there's no turning back. film is dead.
>>
Dan Ayers
is not dead. HDTV is more expensive than renting Panavision, has four
times the equipment, you can't get an NTSC signal for steadicam, need
all new monitors, can't transmit a signal, and has more potiential for
problems than I've ever seen. Sony dosen't seem to care about any of
the issues, believing that this is it. It's still easier and faster to
shoot on film, but I think that video shows, espically sports who are
already experimenting with HDTV, will go to this format because the
cables and electronic considerations will not be an issue. I still
believe that film aquisition to HDTV post will be the future.
Dan
joe
best of luck to ya. i hope all the gear works well and if it doesn't I do
hope Sony takes your input seriously.
j
Dan Ayers wrote in message <36C855...@pacificnet.net>...
Chris Rasmussen
> viable until a mass audience can accept the attributes of a medium that
> will always be different and never equal to film--because it is a different
> medium. It's hard to breed 100 years of film viewing out of the asthetic
> --Jon Springer
I really do hope so, Jon. But I've stalked out of a number of theaters,
cringing in disbelief that general audiences put up, uncomplaining,
with uncorrected: out-of-focus, out-of-frame, garbled-sound,
repetitively-shutdown, presentations, not to mention the many
emulsion-gouged prints.....by the second performance on opening
night!
I've witnessed innumerable instances of audience apathy, but have yet to
see any sign that general film-goers will complain about ANYTHING as
long as the concession lines are short, and the seats marginally
comfortable.
An even more blatant example would be the peculiar popularity of the VHS
format: urine-poor, even for NTSC. I'm afraid most western viewers
will accept whatever is offered them - as long as it's cheap.
Cheers in spite of all!
Chris : )
(Sorry 'bout the spout of negativity there, you caught me with a cold
today.)
Frazer Churchill
thats for an open gate frame, you can cut digital pictures of this rez into
original neg and not see the difference.
A colour bit depth of 10 bits (logarithmic) or 16 bit (linear) is necessary
to maintain good detail in the blacks and whites.
Some FX companies work 8 bit but you can often see the difference.
For wide panoramas, landscapes, sunsets etc it is often worth working at a
higher rez
(4096x3112) anything above that is bigger than the physical size of the film
(35mm).
All that aside, film will be around for a long time to come, all the
technology around digital film
is based around scanning 35mm neg and recording back onto it. Even if the
whole movie was shot
with a digital camera, and all the imagery was computer generated it would
still be recorded back
onto 35mm film because of the global infrastructure of cinemas, they are all
equipped to show
35mm film.
However there are some interesting plans to store entire movies in a digital
format, and "broadcast" them
by satellite to specially equipped cinema's (none exist to my knowledge at the
moment). In much the same way as TV
commercials are 'played out' to the TV stations from Post Production houses.
This would effectively cut out the physical distribution of prints; although
it's highly likely that these films
will be shot on 35mm and scanned. Digital cameras are crap.
The only people who have got anything to worry about are the film
distributors; which isn't a bad thing.
I think film as a means of image aquisition is unsurpassed.
-------------------------------------------
Frazer Churchill
Digital VFX Artist
Double Negative
77 Shaftsbury Avenue
London W1
>
SalayC
be >different because it is an electronic medium. >It is lifeless where as
film is organic. No >matter how good digital acquisition is it still >looks
like the six o'clock news--it's still video.
I remember Audiophiles saying that about CDs vs Records. The truth is Analog
records can produce sound better that 44Khz sampling buy who can tell? The
biggest difference now is resolution and video is not there, but it will be...
Kevin Bjorke
use the Star Wars franchise to cpature it. There has been talk of releasing
parts of the trilogy as digital ONLY -- or at least digital first. When that
happens, every major city will have (one assumes) at least one theatre
converted to digital projection. Since Lucas controlls StarWars, Lucas can
specify the exact standard for projection, much like their current THX
specification. The source for such projection gear would be.... Lucasfilm! See
a pattern?
Once the theatres exist, a de facto industry standard would be set, and
Lucasfilm could reap the profits for decades.
If this is truly the plan, we should all be glad -- I'd *much* rather have
Lucas determining projection standards than, say, the NTSC :)
kb
Sotirios Papakonstantinou
After having read all the replies and that original post on newsgroup
alt.movies.cinematography I decide to tell my opinion. Please excuse me
if I make any error, since I am only an amateur.
I read some days ago, in a website about Super 8mm film, that this film
has maximum resolution (grain pattern, whatever you like) of 1500 lines.
Having in mind that it is a 4:3 medium, then with easy calculations we
can find that it is capable of 2000x1500 pixels. Taking into account
that a 35mm film has about six time the grains of a Super 8mm, due to
larger area and finest grains, then we can easily say that a good 35mm
negative can reach as much as 12000x9000 resolution. However, this
performance degrades due to the many stages of copying and editing, so
the final theatrical release copy would be of less resolution, but still
have a reasonable amount of detail. If the quality of the final print is
80% (say, I suppose) of the original in each dimension, then the
resolution would still be as much as 9600x7200. If digital video where
to reach that resolution with 24 fps (say for easiest case), then we
would need a vast amount of data. Taken we should use 24-bit color
("truecolor", 16777216 colors) for best reproduction, this translates to
use 3 bytes (each byte has 8 bits, 2 in the order of 8 is 256, 256 in
the order of three is 16777216) for each pixel. Some calculations...
9600*7200*3*24 = 4976640000 bytes/sec (about 4.64 TeraBytes/sec!). I do
not think a computer today, even the fastest workstation could reach
this data transfer rate, let alone process them in real-time!
One would think, "why do not use some kind of compression"? OK smart
guys. The only method of compression (as far as I know) used to digitize
video, that is LOSSLESS (required feature, if we are going to represent
film electronically), is that used in DV cameras, ie DCT (Discrete
Cosine Transformation), which only can do five times compression. Say
that we refine the method and be able to do 10 times LOSSLESS
compression, or we devise another algorithm capable of that compression
or even 100 times. Then we would need 4976640000/10 = 497664000
bytes/sec (about 474.6 MB/sec) with the first option, or 49766400
bytes/sec (about 47.46 MB/sec) with the second option. While for the
first option, modifying the DCT algorithm or find a new one to compress
ten times LOSSLESSLY can be relatively easy, it will not work, since it
is extremely difficult (if not impossible) to achieve a 474.6 MB/sec
video data rate with today's technology (in my opinion). The second
method does things easier since we have a lower data rate, which could
be achieved, but it is very difficult to make 100 times LOSSLESS
compression. So 35mm film will not be replaced in the near future (maybe
after 20 years or so, but there is nothing to worry about today!). It
would be more reasonable to try and make a digital camera having
resolution comparable to small gauge films such as regular 8mm and Super
8mm. I think HDTV in its highest resolution (1920x1080) may reach or
surpass regular 8mm (this is a good sign), but certainly has not reached
yet the "humble" (as many film buffs say) resolution 2000x1500 of Super
8mm!
Regarding the number of colors needed to match the quality of 35mm,
then without second thought at least 24-bit (16777216, "truecolor").
16-bit (65536, hi-color) are NOT enough, and this is evident when a
trained eye sees a "truecolor" photo on a computer with display set at
16-bit. I admit that the artifacts of dithering are not seen easily, but
they exist and WILL BE seen on larger displays as that of a film
theatre! And if one is to replace 35mm, there should be NO artifacts or
flaws! Another thing one should notice: Color encoding should be EQUAL
for both luminance and color components, 8:8:8 used in professional
digital editing being the ideal, or at least 4:4:4. NOT 4:2:2 or
anything less! The ideal would be to encode each primary color
separately in R,G,B components, each one given at least one byte (8
bits), so for true color we have 24-bit (16777216, "truecolor"). I would
not recommend the encoding in Y, U,V components. RGB gives the best
results, since in YUV there might be errors due to approximations and
other factors!
In the far future, when digital video-cameras will be able of such
FLAWLESS performance, with these sort of resolutions and color
reproduction, or better, then be sure that all theatres will be equipped
of projecting both 35mm film and digital video. As for the "warthm" of
the film, I think it will be the same reaction as with the advent of TV.
When TV was invented, I am sure there were debates about the quality of
the image and how good it represents the original film. But all these
have been overcomed (by improving the technology involved) and all say
"how good the film looks in DVD", whereas it is true that it does not
look the same as film, even with a good video-projector! When digital
cameras and videos will reach the quality of 35mm, then will again be
debate for the digital projections, but this will be overcomed soon and
everybody will enjoy the new medium.
I hope I helped. Sorry for the lengthy post! Excuse me for all these
details, but I am a computer/technology freak and I would like to be
presice when talking about these.
Sotirios Papakonstantinou
Karl McBurnett wrote:
>
> I read an article a little over two months ago talking about how George
> Lucas has been having his team of gnomes investigate the possibility of
> filming one of his future films in a digital format. While the idea
> intrigues me from a purely "gee-whiz" point of view, I find it hard to
> believe that some gussied-up camcorder could begin to approach the quality
> of plain old-fashioned film. So, I've got a few questions about what a
> digital motion picture would have to achieve to reach 35mm standards:
>
> 1) First off, if you know of any web-sites that covers this subject in
> detail, that would be helpful.
>
> 2) What kind of resolution (number of points/pixels) would a digital film
> need to reach to equal a frame of 35mm film.
>
> 3) How many unique colors would a digital camera need to be able to discern
> to equal 35mm film.
>
> 4) I guess most important would be how would a digital film keep from
> looking like all other videos - overly crisp, without the "warmth" of film.
>
> I'll just thank anyone with answers ahead of time. Thanks :)
>
> Karl McBurnett
S. Pincus
The Lucasfilm format isn't video based.
David Mullen
S. Pincus wrote in message <36CD3F...@idirect.com>...
>I think we must remember this isn't digital video we're talking about.
>The Lucasfilm format isn't video based.
What's the difference? I thought Lucas wanted Sony to build a camera with 24
fps progressive scan camera with more than 2K resolution (I hope he shoots
for 4K) using a CCD in the 2.35 format. But anyway, if you create a digital
video camera with 4K resolution - why is that not video?
Doesn't video just mean "electronic image capture" as opposed to film, which
is "photo-chemical image capture"? What's different between a digital
camera and a video camera? Or are you simply saying that video is
interlaced and non-video is progressive scan?
David Mullen
David Mullen
your eyes) can tell you that Super8 does not equal HDTV in terms of
resolution. I saw some clips from a TV adventure show about Jules Verne
being shot in Canada in HDTV and it's really hard to tell if it wasn't being
shot on 35mm. If it had been shot on Super8, I could guarantee I would be
able to see the difference.
HDTV is closer to the resolution of Super16, but the lack of grain can make
it seem like it has more resolution (Super16 transferred to HDTV looks sharp
but grainy.) HDTV transferred to 35mm looks about as soft as Super16
transferred to 35mm, but again, with less grain.
I love Super8 but I'm not going to fool myself about its resolution and
grain. In fact, its charm is in its lack of resolution and its grainy
texture.
David Mullen
MGamlin
>> has maximum resolution (grain pattern, whatever you like) of 1500 lines.
Hahahahahahahahahahahaha..
I read it on the net so it must be true.
Hahahahahahhahahahahaha.........
Nice set of numbers in your post there.
Are you the guy they based "PI" on?
Karl McBurnett
>The Lucasfilm format isn't video based.
Do you have some information about what this format would be and how it
would differ and/or improve upon digital video?
christoph manz
---+ hi there -+
> I hope I helped. Sorry for the lengthy post! Excuse me for all these
> details, but I am a computer/technology freak and I would like to be
> presice when talking about these.
---+ well, i like it if somebody gives me lots of technical
details, however i wish sometimes that people would believe
their eyes rather than their mind...
---+ please have a look at the the new film by vinterberg,
"festen" or the new lars von trier movie coming out soon
[i've just seen a preview] "the idiots"... these films were
both shot completely with cheap sony DV cameras and then
printed to 35mm...
i'd love to say that nothing beats the quality of super8,
but these two films told me otherwise:
it's hard to notice that those movies were shot on video...
with no additional lighting... no super8 stock could
do that... and as i said, it wasn't even Digital Beta or
DVC-Pro, just cheap DV.
super8 is great, but don't have the illusion digital
cameras will never be able to produce neat pictures...
just my two cents
+- christoph -+
jon springer
You're right, of course. But people like me make statements like that
to secure a place in the "I told you so!" archives. What are the perceived
advantages of digital acquistion in the first place? It's certainly not
cost: as someone else has already stated, it is cheaper to shoot 35mm.
It's certainly not camera mobility: most serious HD and Digital
camera configurations are just as cumbersome or even more cumbersome.
It's not image quality: it is lifeless, flat, inorganic, electronic,
textureless, too exact, too precise, too "good", to fake, too
uninteresting! Your CD comparision is very insightful and even prophetic.
And Digital acquisition will have the same legacy. There will be a
backlash and a re-discovery of the wonderful tactility of film. And I
will have told you so!
--Jon Springer
jon springer
> S. Pincus wrote in message <36CD3F...@idirect.com>...
> >I think we must remember this isn't digital video we're talking about.
> >The Lucasfilm format isn't video based.
>
>
> fps progressive scan camera with more than 2K resolution (I hope he shoots
> for 4K) using a CCD in the 2.35 format. But anyway, if you create a digital
> video camera with 4K resolution - why is that not video?
>
> Doesn't video just mean "electronic image capture" as opposed to film, which
> is "photo-chemical image capture"? What's different between a digital
> camera and a video camera? Or are you simply saying that video is
> interlaced and non-video is progressive scan?
>
> David Mullen
>
>
That is what this argument boils down to: replacing a photo-chemical
process with an electronic one. Why? Because people like, and
especially Lucas want to re-invent the wheel--he's used to being a
pioneer and is trying to force his legacy into being. But why is
electronic acquistion necessarily the right tool for the job? We are
not trying to stifle progress--we're just pointing out that people
like Lucas are barking up the wrong tree.
--Jon Springer
--Jon Springer
George Hupka
jon springer <spri...@pclink.com> wrote in message
news:7alcbs$c...@enews2.newsguy.com...
>What are the perceived
>advantages of digital acquistion in the first place? It's certainly not
>cost: as someone else has already stated, it is cheaper to shoot 35mm.
>It's certainly not camera mobility: most serious HD and Digital
>camera configurations are just as cumbersome or even more cumbersome.
Huh? An HDCAM is about 25 pounds including lens.... When was the last time
you tried to lift an Arri 535 or a Panaflex???
George Hupka
Downstream Pictures, Saskatoon
downs...@saskatoon.com
George Selinsky
> I'm not sure where you get your numbers, but practical experience (using
> your eyes) can tell you that Super8 does not equal HDTV in terms of
> resolution. I saw some clips from a TV adventure show about Jules Verne
> being shot in Canada in HDTV and it's really hard to tell if it wasn't being
> shot on 35mm. If it had been shot on Super8, I could guarantee I would be
> able to see the difference.
Yea, there's a website about Super 8 that claims it can achieve 1200 horizontal
lines, and I cannot see that happening at all. When you look at Kodak's spec
sheets, the resolution per milimeter for it's black and white reversal stocks
goes up to 125 lines for a 1:1000 contrast target (and for 16mm, since the
frame is about 10mm high, this works out to 1250 horizontal lines). For Super
8, it's about half, so figure that's 600-700 lines. It only gets worse from
there.
I wrote to that website, but they were stubborn and didn't want to correct
their mistake (they wanted to start a film to file scanning service with a
rank). I pointed out to them that the Nyquist frequency for getting optimal S8
resolution probably requires scanning it at 1200 lines per frame (just like you
scan 35mm film at 4K, which in itself can only go up to 2K), but the format
itself is blessed if it can get a solid 600 lines with a sharp subject in black
and white, with good lab service and all those other things.
Super 8, as you pointed out, has it's charm in it's low rez film look, but
anyone with eyes will tell you that the HDTV formats have surpassed it already,
as you also implied.
However, comparing DV to S8, one can say that even though DV has the same
resolution that S8 has, the color performance/contrast range of even the S8
reversal stocks makes DV look less film-like when it's transferred to film. Of
course, it's a matter of time when HDTV formats reach the consumer level (price
wise), and at that point it will be more easy to simulate S8 on tape to film
transfers.
Personally, I see a future in Super 8 if people start digitizing it, editing it
that way, then outputting the digital file to 35mm. This would give an
interesting look to the film that would be worth experimenting with for some
projects, whilst avoiding the entanglements with tape splices, sound, and
optical stages. I might try this approach with an upcoming project.
- G.
jon springer
hollywood motion pictures and the more substantial digital cameras that
Lucas and others are considering (the "Panacam" is one example).
--Jon Springer
A.J. Clarke
> ---+ please have a look at the the new film by vinterberg,
> "festen" or the new lars von trier movie coming out soon
> [i've just seen a preview] "the idiots"... these films were
> both shot completely with cheap sony DV cameras and then
> printed to 35mm...
(snip)
> super8 is great, but don't have the illusion digital
> cameras will never be able to produce neat pictures...
I've not seen "The Idiots" (it's just about to be released here in
the UK) but my experience of watching Lars Von Trier films is that he's
pretty well indifferent to "conventional" ideas of cinematography, so
shooting on DV and then transferring to film sounds his typical
cheapskate style.
Many of his films look like they're done on 16mm film, with very
grungey sound and optical effects, and The Kingdom (TV series, but got a
cinema release here, to cultist who can sit still for five hours!) looks
like it was shot on out-of-date stock and developed in pickling vinegar! I
think even the S8 underground film-making community would think twice
about adopting him as a role-model... (though his ideas are sometimes
amazing).
Tony Clarke
gabs
I think we are talking on a different basis. When I say Super 8mm is capable of
2000x1500 resolution, I mean that IF the grains were arranged in order (and not
randomly) and one wished to digitize a Super 8mm frame with one-to-one
correspondence (ie one pixel per grain) then he would need that resolution. Put
it in other words, if Super 8mm was a digital format, then its resolution would
be 2000x1500 pixels (3 megapixel). I think when you say Super 8mm can resolve L
lines you mean L separate black lines on a white background. This is different!
One should take account of the white background as well, so as to make a total
resolution of 2000x1500.
I hope now I am clear.
Sotirios Papakonstantinou.
David Mullen wrote:
> I'm not sure where you get your numbers, but practical experience (using
> your eyes) can tell you that Super8 does not equal HDTV in terms of
> resolution. I saw some clips from a TV adventure show about Jules Verne
> being shot in Canada in HDTV and it's really hard to tell if it wasn't being
> shot on 35mm. If it had been shot on Super8, I could guarantee I would be
> able to see the difference.
>
gabs
It depends of the projection size. If you project on typical home movie
sizes, then I think, one cannot make any difference if the material is
Super 8mm, HDTV, 16mm or 35mm film. If however, one projects to large
sizes comparable to the screens of movie theaters, then obviously any
differences will show up.
Bye.
Sotirios Papakonstantinou
christoph manz wrote:
> ---+ hi there -+
>
> > I hope I helped. Sorry for the lengthy post! Excuse me for all these
> > details, but I am a computer/technology freak and I would like to be
> > presice when talking about these.
>
> ---+ well, i like it if somebody gives me lots of technical
> details, however i wish sometimes that people would believe
> their eyes rather than their mind...
>
> ---+ please have a look at the the new film by vinterberg,
> "festen" or the new lars von trier movie coming out soon
> [i've just seen a preview] "the idiots"... these films were
> both shot completely with cheap sony DV cameras and then
> printed to 35mm...
> i'd love to say that nothing beats the quality of super8,
> but these two films told me otherwise:
>
> it's hard to notice that those movies were shot on video...
> with no additional lighting... no super8 stock could
> do that... and as i said, it wasn't even Digital Beta or
> DVC-Pro, just cheap DV.
>
> super8 is great, but don't have the illusion digital
> cameras will never be able to produce neat pictures...
>
>
gabs
Maybe we talk on different basis. When I say Super 8mm has 2000x1500 resolution,
then I effectively compare it to video. IF the grains were arranged in order as the
pixels on a PC monitor, and one wished to digitize a Super 8mm frame with
one-to-one correspondence (one pixel per grain), then Super 8mm would have the
resolution of 2000x1500. When you say about resolving lines, you probably mean
separating black lines on a white background. This is totally different! One should
take account of the white background as well, in order to make the complete
resolution of 2000x1500 pixels (grains, dots, whatever).
Also, I think, if one is projecting in typical home movie sizes, then it is hard to
tell whether is a Super 8mm, 16mm, 35mm film or HDTV without knowing. If however,
one projects to large sizes comparable to the screen of a movie theater, then
obviously any differences will show up.
I hopw I am clear now.
Sotirios Papakonstantinou
George Selinsky wrote:
> David Mullen wrote:
>
> > I'm not sure where you get your numbers, but practical experience (using
> > your eyes) can tell you that Super8 does not equal HDTV in terms of
> > resolution. I saw some clips from a TV adventure show about Jules Verne
> > being shot in Canada in HDTV and it's really hard to tell if it wasn't being
> > shot on 35mm. If it had been shot on Super8, I could guarantee I would be
> > able to see the difference.
>
gabs
However, when I say Super 8mm is capable of 2000x1500 resolution, I effectively
compare it to video. IF its grains were arranged in order, as the pixels on a
PC monitor, and one wished to digitize a Super 8mm frame with one-to-one
one says about the resolving power of Super 8mm, then he says how many black
lines it can separate in a white background. This is totally different! One
should take into account the white background as well in order to make the
complete resolution!
If you still think I am making a mistake, OK "know-all" guy. This is a question
for you. Can you tell me what the actually resolution of Super 8mm is (based in
the pixel structure, as explained above)? Can you tell me where can I find this
information on the Internet? I would like also to tell me how did you derive
your result.
Thank you in advance, and forgive me; I am only an amateur video-grapher who
wants to compare film with video. And not I am not the guy they based "PI" on.
Are you relative with Einstein?
Bye.
Sotirios Papakonstantinou
MGamlin wrote:
> >> I read some days ago, in a website about Super 8mm film, that this film
> >> has maximum resolution (grain pattern, whatever you like) of 1500 lines.
>
Scott Norwood
>
>Hi!
>It depends of the projection size. If you project on typical home movie
>sizes, then I think, one cannot make any difference if the material is
>Super 8mm, HDTV, 16mm or 35mm film.
I have 16mm and 35mm projectors in my basement, and have run s8 there,
too. There is a BIG difference among the formats, although some of that
can be attributed to factors other than film gauge. Running 35mm on a
Simplex projector with a xenon light source and good lenses isn't a fair
comparison to s8 on a plastic home-movie projector with cheap lenses and
a tungsten light source. But there is nonetheless better sharpness and
less grain with the larger formats. 16mm (with OK lenses and tungsten
light source on a classroom projector) looks much better than s8 with
a similar quality setup.
The screen size is about 6' wide at the moment...wider for CinemaScope.
Obviously, differences would be more visible on a larger screen (but
then brightness would be a problem with s8).
--
Scott Norwood: snor...@nyx.net, snor...@redballoon.net, sen...@mail.wm.edu
Cool Home Page: http://www.redballoon.net/
Lame Quote: Penguins? In Snack Canyon?
gabs
Since I am not at the Computer Services of our University at the moment, but at a Lab,
please reply to my own address
at...@city.ac.uk
The other address ag...@city.ac.uk is not mine! It is the last student's who
logged in before me. I just didn't change the settings.
Thanks
Sotirios Papakonstantinou
gabs wrote:
> Hi!
> Maybe we talk on different basis. When I say Super 8mm has 2000x1500 resolution,
> then I effectively compare it to video. IF the grains were arranged in order as the
> pixels on a PC monitor, and one wished to digitize a Super 8mm frame with
gabs
answer the folowing simple question: If we forget the disadvantages of each medium for the
moment (talking about color or grain etc), then which would be the sharpest? A scene shot
on Super 8mm with enough lights and projected to a good projector with good lenses, or the
same scene shot on a HD camera at 1920x1080 and projected on a relevant HDTV set?
Obviously Super 8mm would be sharper, since it is film. But HDTV reaches 1920x1080
already, so if Super 8mm is sharper, then it MUST have better resolution. This justifies
me saying that Super 8mm has resolution of 2000x1500 and not significantly lower than
that. Of course larger formats such as 16mm and 35mm, obviously will have even greater
resolution. So, how one can say that 35mm film is capable of 2K (2000 lines)? It should be
far better than that (even if the average human eye cannon judge the difference with high
accuracy).
Bye.
Sotirios Papakonstantinou
PS: My true e-mail address is at...@city.ac.uk as shown above. The other is of another
student of our University. I simply did not change the settings!
David Mullen
You can see with your eyes that its resolution is not as good as HDTV (if
you are familiar with HDTV). Ultimately, what matters is how things look,
not what the numbers are.
Sony has made transfers to 35mm of HDTV and shown side-by-side comparisons
with 35mm photography. Most who have seen the test say it is a close match,
even if the 35mm is still better. This suggests that Super8 blown-up to
35mm would look much worse than HDTV-to-35mm.
The numbers you quote suggest that anyone watching Super8 transferred to
tape and shown on an NTSC monitor could not distinguish it from 16mm or 35mm
or HDTV because the resolution exceeds what NTSC could show. Yet Super8
footage is always easy to spot in such a comparison. There is a lot of
debate currently if even Super16 is high-quality enough for future HDTV
broadcast - if Super16 is debateable, then Super8 isn't even in the running.
Have you looked at much current HDTV photography on an HDTV monitor?
You should make choices based on what your eyes tell you, not what the
technical numbers tell you. Super8 can look beautiful, plus it is a film
image not a video image, but it's basically a soft & grainy format and
that's why so many people love the way it looks. It's not the format to
choose if you are concerned about resolution.
David Mullen
George Selinsky
>
> You should make choices based on what your eyes tell you, not what the
> technical numbers tell you. Super8 can look beautiful, plus it is a film
> image not a video image, but it's basically a soft & grainy format and
> that's why so many people love the way it looks. It's not the format to
> choose if you are concerned about resolution.
>
> David Mullen
A-men!
This is not to say that Super 8 is a bad format, but if it was so sharp and
competitive to HDTV video, then everyone would be shooting it!!!
- G.
MGamlin
I don't need to be polite, this is America. We pride ourselves on both free
speech and whipping others into shape. My derision can strengthen you if you
have the guts to wake up. Otherwise, hahahahahaha.
>However, when I say Super 8mm is capable of 2000x1500 resolution, I
>effectively
>compare it to video. IF its grains were arranged in order,
AH-HA... "IF"...but they AREN"T. Further, this is why film looks BETTER
( Apparently harper/ less grainy) in motion than in still- because of the
randomness of grain pattern.
as the pixels on a
>PC monitor, and one wished to digitize a Super 8mm frame with one-to-one
>correspondence (one pixel per grain),
one can wish for anything but one may not get it...
then he would need 2000x1500 pixels.
>When
>one says about the resolving power of Super 8mm, then he says how many black
>lines it can separate in a white background.
What about the lens...?
This is totally different! One
>should take into account the white background as well in order to make the
>complete resolution!
Eh?
>
>If you still think I am making a mistake, OK "know-all" guy. This is a
>question
>for you. Can you tell me what the actually resolution of Super 8mm is (based
>in
>the pixel structure, as explained above)?
As I recall, best case scenario, Super 8 is around 500. That does not negate
the granularity which sets it appart from a soft 16mm image of 500 lines... and
we've never toushed on contrast or dynamrange. You want to talk Kodachrome,
Ektachrome, negative, what? I 've been there, likely before you were born.
Can you tell me where can I find
>this
>information on the Internet?
You can find whatever you want on the internet.
I would like also to tell me how did you derive
>your result.
>
Ever heard of BOOKS?
Try Mikolas & Hoos, Lipton, etc....the SUper 8 era which is long dead had
plenty of state-the- art BS about how great super 8 was...
Find a library.
>Thank you in advance,
No charge
and forgive me; I am only an amateur video-grapher who
>wants to compare film with video.
Whatever floats your boat.
And not I am not the guy they based "PI"
>on.
Too bad. You could get a royalty check.
>Are you relative with Einstein?
As a matter of fact, yes, Al was my dad. Don't believe me? It says so, right
here on the internet...
Or myabe my point, or sarcam, won't translate to Greek....
Sotirios Papakonstantinou
I case that you do not know, Sony has transferred HDTV to 35mm using its
Electronic Beam Recorder (EBR) method. This method (if I understood
well), uses a high grain 35mm negative in place of a CRT and slowly
exposes the film as the electron beam pass over it. All in complete dark
conditions. It does not record all the colors at once, but it uses
sequential color exposing and combines the negatives to give the color
replica (a similar methor to Technicolor). More info at
http://www.spe.sony.com/Pictures/Hidef/ebr01.htm
That is why the transfer is very good. They do not do it by shooting a
HDTV monitor or whatever which of course would produce "bad" results.
The comparison with Super 8mm blown-up to 35mm would be fair, if they
were transfering video to 35mm the "traditional" way. However, I have
not seen HDTV yet, since in Greece Digital TV is yet to come, and in UK
I should pay a lot for devices and subscriptions etc, and in addition
there is no HDTV yet either. You may be right and HDTV look good, but
film is film, and one's eyes can fool him sometimes. So I look for some
data to confirm or reject the impression I will be given when I see both
formats. We will need some time to see film replaced by video.
Bye.
Sotirios Papakonstantinou
PS: Just for curiosity's sake. If the grains of a Super 8mm frame wre
arranged in order, like the pixels on a Digital TV say, then how much
resolution we would need if we were to digitize a frame with one-to-one
correspondence (one pixel per grain/dot/whatever)? If you don't know the
answer, can you tell me where I can find this information on the
Internet? I would like responsible answer, based on the film's chemical
attributes, not assumptions please. I would like to compare Super 8mm to
video (Hi8 or DV say) and see how much better (sharper etc) it is.
David Mullen wrote:
>
> Have you ever sat in good telecine suite and transferred Super8 to video?
> You can see with your eyes that its resolution is not as good as HDTV (if
> you are familiar with HDTV). Ultimately, what matters is how things look,
> not what the numbers are.
>
> Sony has made transfers to 35mm of HDTV and shown side-by-side comparisons
> with 35mm photography. Most who have seen the test say it is a close match,
> even if the 35mm is still better. This suggests that Super8 blown-up to
> 35mm would look much worse than HDTV-to-35mm.
>
> The numbers you quote suggest that anyone watching Super8 transferred to
> tape and shown on an NTSC monitor could not distinguish it from 16mm or 35mm
> or HDTV because the resolution exceeds what NTSC could show. Yet Super8
> footage is always easy to spot in such a comparison. There is a lot of
> debate currently if even Super16 is high-quality enough for future HDTV
> broadcast - if Super16 is debateable, then Super8 isn't even in the running.
>
Sotirios Papakonstantinou
>
> > Have you looked at much current HDTV photography on an HDTV monitor?
> >
> > You should make choices based on what your eyes tell you, not what the
> > technical numbers tell you. Super8 can look beautiful, plus it is a film
> > image not a video image, but it's basically a soft & grainy format and
> > that's why so many people love the way it looks. It's not the format to
> > choose if you are concerned about resolution.
> >
> > David Mullen
>
>
> This is not to say that Super 8 is a bad format, but if it was so sharp and
> competitive to HDTV video, then everyone would be shooting it!!!
>
> - G.
Yes, everyone would be shooting it, if there were not other constraints
such as the sound (which would need double system and difficult
sunchronization later, since Kodak has discontinued the sound films),
the difficulty in tranfering to video, as opposed with larger formats.
Another difficulty is the limited shooting duration of a Super 8mm
cartidge. One would have to change cartidge every a while and maybe he
won't be fast enough and miss something when shooting live action. The
cost. OK HDTV may be much expensive, but (I think so), if one had to
shoot the same duration on Super 8mm, as he would shoot on a HD video
cassette, the cost of the cartidges would be higher than of the
cassette. In addition, as a film (and of limited ASA ratings at the
moment) Super 8mm would need plenty of light for good results, as
opposed to HDTV.
If believe that if there were not all these constraints, many TV
broadcasters would employ Super 8mm (or 16mm if their budget was
higher), at least for the less "high" definition broadcasts of 720p
lines, 1080i being the maximum at the moment.
Bye.
Sotirios Papakonstantinou.
A.J. Clarke
> Hi!
> Maybe we talk on different basis. When I say Super 8mm has 2000x1500 resolution,
> then I effectively compare it to video. IF the grains were arranged in order as the
> pixels on a PC monitor, and one wished to digitize a Super 8mm frame with
> one-to-one correspondence (one pixel per grain), then Super 8mm would have the
> resolution of 2000x1500.
I can't quote the maths to back this, but you won't get accurate
rendition with one pixel = one grain. To resolve film grain - which
follows a semi-random (Brownian?) distribution in size and position on the
emulsion - you will need to OVERSAMPLE it, in order to resolve its density
and size to any accuracy.
Certainly in order to lose pixel boundaries when projected on a
screen, you need colossal density, in that effectively you need a pixel
size equivalent to the eye's resolution at the screen distance. HDTV is
getting there at TV viewing distance, but the front 20 rows of any cinema
are still not going to accept the quality of a HDTV projected image.
Digital is a conveniently manageable (now) way to manipulate image storage
but it's still a compromise on the limitless resolution of analogue
reality, as any book on chaos theory will tell you.
FWIW I'm certain that the resolving power of the lens (expressed
as the "circle of confusion" criterion in a depth of field table)
represents the limiting factor in optical resolution. A sharp lens on S8
will yield a better quality picture than a rubbish lens on 35mm, and film
cameras, even cheap S8, are still on the whole using glassware an order of
magnitude sharper than the compromises fitted to camcorders. Try a
"surveillance camera" lens or one of those "wide-angle converters" sold
for camcorders, on your movie camera and watch the picture turn to cheap
mush.
Tony Clarke
David Mullen
>HDTV is
>getting there at TV viewing distance, but the front 20 rows of any cinema
>are still not going to accept the quality of a HDTV projected image.
Nor will they accept the quality of a Super8 image, either projected in
Super8, or blown-up to 35mm and projected.
David Mullen
Frazer Churchill
between 2K and 4K outputs.
2048x1556 is sufficient resolution on 35mm to faithfully reproduce original negative.
Putting a picture size of 2000x1500 onto super 8 would be massive overkill, totally
unecessary.
I would suggest 512x389 would be more than adequate to fill a super 8 frame.
A.J. Clarke
It's been done, though as a chichi "textured" effect such as in
Oliver Stone's JFK (where he creates "matching" footage for the Zapruder
film of Kennedy's assasination), or as in Derek Jarman's The Last Of
England, where it's a deliberately grungey apocalyptic look he's after.
The random grain and deeper colour density of film, even S8, is
still more visually acceptable than the thin pastel tones, image bleed and
visible raster lines of the video blowups I've seen to date. I've not yet
seen HDTV demonstrated, nor movies transferred to DVD (word is out over
here that DVD is due to wipe VHS off the face of the earth for home movie
watching in real quality - though we still suffer from that Zone 1 / Zone
2 encoding scam), and await a demonstration on real material - not
synthesized SFX from video games - to prove that it is the real thing.
Tony Clarke
George Selinsky
>
> If you still think I am making a mistake, OK "know-all" guy. This is a question
> for you. Can you tell me what the actually resolution of Super 8mm is (based in
> information on the Internet? I would like also to tell me how did you derive
> your result.
I have answered your question on rec.movies.production under the DV column, please
seek your answer there. I made a great effort to elucidate this to you.
If you are an amateur, in all honesty you do speak like you know it all in some
posts. This has tried the temper of those who indeed are professinals like myself
and have used this medium. We obviously know what we're talking about.
- G.
Regnirps
>Yea, there's a website about Super 8 that claims it can achieve 1200
horizontal
>lines, and I cannot see that happening at all. When you look at Kodak's
>spec sheets, the resolution per milimeter for it's black and white reversal
>stocks goes up to 125 lines for a 1:1000 contrast target (and for 16mm, since
>the frame is about 10mm high, this works out to 1250 horizontal lines). For
>Super 8, it's about half, so figure that's 600-700 lines. It only gets worse
from
>there.
>
>I wrote to that website, but they were stubborn and didn't want to correct
>their mistake (they wanted to start a film to file scanning service with
>a rank). I pointed out to them that the Nyquist frequency for getting optimal
>S8 resolution probably requires scanning it at 1200 lines per frame (just like
>you scan 35mm film at 4K, which in itself can only go up to 2K), but the
format
>itself is blessed if it can get a solid 600 lines with a sharp subject in
>black and white, with good lab service and all those other things.
I think there is a difference here. You are applying Nyquist to Kodak's
resolution figures, but to capture ALL the information in a piece fo film you
will have to have pixels that are less than half the size of a film grain.
Anyone have figures on grain size for typical films? I need to know for some of
my own work and may resort to the microscope. (I'm trying to digitize to the
extent that the decay of the old film won't be a concern anymore).
Charlie Springer
Regnirps
>Date: Mon, Feb 22, 1999 8:02 AM
>Message-id: <36D17FAC...@city.ac.uk>
>
>Hi!
>Maybe we talk on different basis. When I say Super 8mm has 2000x1500
resolution,
>then I effectively compare it to video. IF the grains were arranged in order
>as the
>pixels on a PC monitor, and one wished to digitize a Super 8mm frame with
>one-to-one correspondence (one pixel per grain), then Super 8mm would have
>the
>mean
>One should
>take account of the white background as well, in order to make the complete
>resolution of 2000x1500 pixels (grains, dots, whatever).
Yes indeed, except you must capture at greater than twice the highest spacial
frequency or in this case two pixels per grain. To truly capture the image to
the point where you can re-create the film correctly you will need 4000x3000!
This is the big problem for archivists who want to transfer old film to
digital. Mostly today they wind up making a new print as the only cost
effective solution with high enough resolution.
Charlie Springer
Regnirps
snip ignorant nonsense,
>Or myabe my point, or sarcam, won't translate to Greek....
No, but it translates pretty well to "redneck shit-for-brains". It says so
right here on the internet!
Charlie Springer
Regnirps
>Have you ever sat in good telecine suite and transferred Super8 to video?
>You can see with your eyes that its resolution is not as good as HDTV (if
>you are familiar with HDTV). Ultimately, what matters is how things look,
>not what the numbers are.
This hardly seems fair unless the telecine you use is one of the few with an
HDTV camera in it? I find 16mm to D1 is way insufficient to apply good
deblurring algorithms to detailed images on a frame by frame basis. I need to
go to 1300x1024 at a minimum and would like to do twice that.
In fact, I am looking for a telecine tranport mechanism for just this purpose.
I have found a 10 bit 12 fps camera with high enough resolution to dot he
restoration of military footage I am working on. Anyone know where to look for
such a piece of equipment? I'm carving up an old Bolex H16 at the moment but
not too happy with what I have to do to control the drive, not to mention
cutting a hole in the backplate and puting in a prism, etc.
Charlie Springer
David Mullen
>>You can see with your eyes that its resolution is not as good as HDTV (if
>>you are familiar with HDTV). Ultimately, what matters is how things look,
>>not what the numbers are.
>
>This hardly seems fair unless the telecine you use is one of the few with
an
>HDTV camera in it? I find 16mm to D1 is way insufficient to apply good
>deblurring algorithms to detailed images on a frame by frame basis. I need
to
>go to 1300x1024 at a minimum and would like to do twice that.
My point was and still is that one can quote all the figures in the world,
but Super8 does not have more resolution than HDTV no matter in what form
you see either (on print, on a video monitor, in telecine, etc.) So this
notion that "shoot Super8 because NTSC can't see the difference" because
NTSC has less resolution doesn't take into account that one CAN see the
difference between Super8, 16mm, and 35mm with one's own eyes on a TV
monitor.
My point is that you choose a format because how it looks to your eyes, not
what a chart tells you. Someone who says that Super8 is technically better
than HDTV because of what the numbers tell him is not using his eyes. I had
just come from a display on HDTV monitors and samples of current shooting in
HDTV at a seminar at Sundance - after just having been in a telecine suite
transferring Super8 film to digi-beta. So I have some recent experience
looking at both.
And my final point, again, is that there is nothing wrong with Super8 - I've
shot much of it in my lifetime - but that one chooses it for its unique
look - which can almost describe as "anti-HDTV". It's grainy and soft but
screams "FILM" in a wonderful way. I'm not really interested in how many
pixels of resolution it has... One doesn't have to shout its praises by
claiming that "it is superior to HDTV in resolution"; it doesn't need to
compete with ANY format - it is what it is.
David Mullen
Sotirios Papakonstantinou
LARGER resolution than the 1920x1080 of HDTV. It may be small gauge, but
it is FILM, and no video format is capable of true film
resolution/detail yet! Maybe HDTV has matched or surpassed regular 8mm
(the "worst" film format), but certainly NOT Super 8mm yet! If one used
an appropriate video-projector to project HDTV to the same dimensions as
Super 8mm, I bet Super 8mm would look better, whereas in HTDV one might
see pixelation (the pixels that make the image become visible due to
size).
Thank you very much!
Sotirios Papakonstantinou
Sotirios Papakonstantinou
Thank you again for your support. The guy should be more polite with
amateurs! It is not sensible to expect an amateur having the same
knowledge and experience with a professional! One is likely to make a
mistake when speaking about something he is not "professional" at.
Afterall he insulted me and my country Greece. Don't worry however. I
have send him a couple of weeks ago, a reply as he deserved. If he has
the guts to insult me again, I will not post to the newsgroup (so I'll
avoid to be banned), but I'll reply directly to him and be as rudest as
possible for my English.
Thanks again.
Sotirios Papakonstantinou.
Sotirios Papakonstantinou
You may be right with all these comments. Afterall you are a
professional (I guess) and you know better from me that I am amateur.
However, if one had an appropriate video-projector and projected HDTV at
the same dimensions as Super 8mm, I think that he would experience the
"pixelation" effect (the pixels making the image, show up as large
squares because of the projection size). I am not sure if Super 8mm
shows up grain at the same size, but certainly this grain is not as
annoying as the "pixelation" effect, since the grains are random on the
film. This is another "proof" that Super 8mm film is (or at least looks)
better than HDTV. And you cannot say about numbers this time. One could
notice the effect "using his eyes".
I do not want to tell Super 8mm is superior from any video format, but
this is what I realize, based on my little amateur experience and some
arbitrary assumptions.
Sotirios Papakonstantinou
Regnirps
>My point was and still is that one can quote all the figures in the world,
>but Super8 does not have more resolution than HDTV no matter in what form
>you see either (on print, on a video monitor, in telecine, etc.) So this
>notion that "shoot Super8 because NTSC can't see the difference" because
>NTSC has less resolution doesn't take into account that one CAN see the
>difference between Super8, 16mm, and 35mm with one's own eyes on a TV
>monitor.
>
>My point is that you choose a format because how it looks to your eyes, not
>what a chart tells you. Someone who says that Super8 is technically better
>than HDTV because of what the numbers tell him is not using his eyes. I had
>just come from a display on HDTV monitors and samples of current shooting in
>HDTV at a seminar at Sundance - after just having been in a telecine suite
>transferring Super8 film to digi-beta. So I have some recent experience
>looking at both.
>
>And my final point, again, is that there is nothing wrong with Super8 - I've
>shot much of it in my lifetime - but that one chooses it for its unique
>look - which can almost describe as "anti-HDTV". It's grainy and soft but
>screams "FILM" in a wonderful way. I'm not really interested in how many
>pixels of resolution it has... One doesn't have to shout its praises by
>claiming that "it is superior to HDTV in resolution"; it doesn't need to
>compete with ANY format - it is what it is.
I understand. I'm talking about what you have to do to capture all the
information in an image and what kind of resolution you need for good image
processing. I used to use dual 8 in a Bolex dual 8 that looks just like a H16.
It was the cheepest way to learn with something like 15 minutes from a 100 foot
16mm roll.
However, regarding super 8, I have never seen it projected from top grade
optics. We had one of those goofy Bolex Super 8 cameras for a while for
time-lapse studies but I can't swear to the projector quality.
Charlie Springer
Sotirios Papakonstantinou
OK. I'll attempt to clarify what I mean when I say "Super 8mm has
2000x1500 resolution" and end this debate. I hope that after my
explanations. anyone will understand what I want to say and will be able
to confirm or reject my assumptions without guesses for the meaning.
Please note that I am an amateur and I may make a mistake. OK, here it
goes:
All of us that have a computer know that the larger the resolution, the
greater the detail we can see on the screen (in the SAME dimensions),
and the better looks an image. Say we digitize a photo at 320x200
pixels. Suppose we always use 16777216 colors (24-bit, "truecolor") for
the digitization AND for the display, so the color and image
reproduction is as good as the dimensions (resolution) can allow. A
photo at this low resolution of 320x200, may look good (depending on the
amount of detail it contains), but it certainly looks better when we
digitize it at the higher resolution of 640x480 say. Of course, 800x600
resolution allows for more detail and better image, 1024x768 even
better, and so on.
Now the question is: Suppose we use the maximum amount of colors
possible for faithfull color reproduction (usually 16777216 colors, or
24-bit, or "truecolor" for a PC or MAC computer). So the color
reproduction is not an issue. Suppose further that we have enough memory
on a graphics card and an appropriate monitor, so we can have as high a
resolution we want, up to 4000x3000 or something. Now, if we wanted an
image to show the same amount of detail as a frame of Super 8mm film can
show, then what would be the appropriate resolution? Suppose that we
have plenty of light and a high contrast, so we reach the limits of
Super 8mm in terms of quality. What would be the resolution on a PC
monitor in order to see the same amount of detail? I think in this case
we would need 2000x1500 resolution. Of course, I am refering to the
equivalent resolution, and not the Nyquist sampling resolution which
should be at least double for faithful reproduction of a Super 8mm
frame. This why I say that Super 8mm is capable of (has the equivalent
resolution of) 2000x1500 pixels. Because, I think, this is the
resolution we would need on a PC monitor, in order to see or digitize
the same amount of detail that a Super 8mm frame (in good shooting
conditions, as described above) is capable of.
I hope now you clearly understood what I mean. Can anyone confirm or
reject this assumption? About the 1500 lines figure, I found it from an
Internet site about Super 8mm that said that Super 8mm can show about
1500 lines. Since Super 8mm is 4:3 medium, one can easily find the
number of columns as 1500*4/3=2000, so the complete resolution is
2000x1500 (2000 columns times 1500 lines in pixels, if we compare it to
a PC display).
I hope I am clear now. If you are not sure what I mean, please do not
hesitate to ask me. Please confirm or reject this assumption, but I
would like to somewhat PROVE your opinion, based on some data or
comparison with another medium of known resolution, such as HDTV of
1920x1080 resolution. I have not seen any HD video at 1920x1080, but I
think the amount of detail displayed is somewhat less of what Super 8mm
film can show. This format, in my opinion, may have reached the
resolution (amount of detail) of regular 8mm, or it may have surpassed
it, but certainly it has not reached yet Super 8mm. Can you confirm or
reject that?
Thank you in advance.
Sotirios Papakonstantinou
David Mullen
looks best projected from the original, while HDTV is best seen on an HDTV
monitor, one possible method of comparing each of their resolutions would be
the blow them both up to 35mm.
I think if you saw such a blow-up, you'd find that the HDTV image would beat
the Super8 image in terms of overall quality, especially in terms of
resolution. Sony Hi-Def Center has demos comparing 35mm prints made from
35mm negative to 35mm prints made from HDTV - and most professional DP's who
have seen these tests tell me that HDTV is almost as good as 35mm in the
final pojected image. So assuming that Sony is showing HDTV in its best
light and that HDTV is actually closer to Super16 than 35mm in terms of
resolution, that still makes it better than Super8. Believe me, you
couldn't blow-up Super8 to 35mm and show it to people hoping that their
reaction would be "that almost looks like 35mm!"
But hey, if you had a chance to shoot a feature to be blown up to 35mm and
had a choice between Super8 and HDTV and told your producers that your math
has proven that the Super8 format has more resolution than HDTV so that it
will produce a sharper blow-up - then go ahead... Don't listen to anyone,
learn it for yourself.
David Mullen
David Mullen
If 4000x3000 is considered necessary to render a 35mm Academy frame, which
is .864" x .630" (.544 sq. inches). The 16mm frame is .404" x .295" (.119
sq. inches) - meaning that the 16mm frame is only 22% of the size of the
35mm frame in total surface area.
Assuming that the Super8 frame is slightly larger than 22% of the 16mm frame
(let's round up tp 30%), that still makes the Super8 frame only about 6% of
the total area of a 35mm frame. Let's round up some more and say that it is
10% of the 35mm frame.
So how does a format that only takes up about 10% of the 35mm frame going to
equal 2000x1500, or half of the 35mm's 4000x3000?
David Mullen
Alex Reckie
>
> will produce a sharper blow-up - then go ahead... Don't listen to anyone,
> learn it for yourself.
My advice to this guy who keeps claimin' that Super 8 has very high resolution
is to go out there, find a script, and a willing crew, shoot a feature film in
Super 8, get it blown up to 35mm, and watch it projected on a large screen.
This is what he needs to do, he loves the format, he thinks it's the greatest
thing on earth. Let him use it as much as possible.
Either that, or go taking road trips and smashing HDTV cameras, sending
Unabomber styled threats to those who dare shoot on that damned electronic
media - perhaps even kidnap the president of Kodak and demand the increased
production of Super 8 film, with a 500% increase in laboratories as ransom.
As far as making a stand in the technical world, even despite tons of
professionals explaining to him all the details of resolution and
determination, he just won't *$@! listen to anyone. He's apparently more
interested in proving the superiority of Super 8.
Well, my advice to him is to stop talking and go out there and make a goddamn
movie in Super 8, and prove it to the world, not torture the brains of others
who are trying to prove something to this man in vain. Put your money where
your mouth is, so to speak.
If he's smart, he'll discover that it's not the car, but the driver that wins
the race.
ALEK Reckie.
Regnirps
>Date: Tue, Mar 2, 1999 8:23 AM
>Message-id: <7bh37t$m8h$1...@fir.prod.itd.earthlink.net>
I think there is a natural confusion over resolution when we mix the fields of
image processing, which deals with information theory and maximum sampling
required to capture ALL the information in a signal to avoid aliasing (spatial
equivalent of the time phenomena of wagon wheels going backwards in old films -
too slow a sampling rate for the speed of the wheel) and the field of
photography. In photography the resolution is usually defined in terms of the
optical system's modulation transfer function of MTF and there is a convention
of resolving line pairs to characterize the MTF.
On computers, like in painting or viewing a photograph, the only optical system
involved is the eye. There is no camera/projector pair to worry about. The
question becomes, with great optics, what is the equivalent compter resolution
for working with Super8 with no loss of image information? The answer is you
have to work with pixels near the grain size, or idealy less than half the
grain size to avoid artifacts die to aliasing. This applies to any film UNLESS
the MTF of the camera used to make the film shows the optical resolution to be
much lower and there is therfore no usefull information below some threshold of
resolution. A line pair test can give you a good approximation.
Note however that new methods from astronomy, like 'maximum entropy
deconvolution" allow you to determine the point spread function of a lens and
"undo" its bluring effect to some extent. So in order to be sure you will meet
the needs of the future, you must digitize at the Nyquist level (more than two
samples per grain) for archival or preservation purposes.
Charlie Springer
Sotirios Papakonstantinou
Yes, I get your point. Do not forget however, that 4000x3000 may be the
resolution of the final theatrical release. The original resolution of
the negative should be much greater, but due to the many generations of
copies and vast editing etc etc, it has degraded to 4000x3000. What is
the resolution of an original negative, as it goes out of the camera
without any copying or editing? I think it must be much higher than
400x300, about a figure of 12000x9000, if we use analogy of the size of
35mm frame in comparison to Super 8mm (supposing Super 8mm is capable of
2000x1500).
What is your opinion now? Can you confirm or reject that Super 8mm is
capable of such a resolution (amount of detail) as 2000x1500?
Thank you for your answer anyway.
Sotirios Papakonstantinou
David Mullen wrote:
>
> I still don't understand your math or logic.
>
> If 4000x3000 is considered necessary to render a 35mm Academy frame, which
> is .864" x .630" (.544 sq. inches). The 16mm frame is .404" x .295" (.119
> sq. inches) - meaning that the 16mm frame is only 22% of the size of the
> 35mm frame in total surface area.
>
> Assuming that the Super8 frame is slightly larger than 22% of the 16mm frame
> (let's round up tp 30%), that still makes the Super8 frame only about 6% of
> the total area of a 35mm frame. Let's round up some more and say that it is
> 10% of the 35mm frame.
>
> So how does a format that only takes up about 10% of the 35mm frame going to
> equal 2000x1500, or half of the 35mm's 4000x3000?
>
> David Mullen
Sotirios Papakonstantinou
I reply again, since this is a separate portion of your reply.
OK. I think I had posted it before, but here it is again: Sony uses a
special mehtod to transfer HDTV footage to 35mm film, which is called
"Electron Beam Recorder" (EBR). They replace a typical HDTV monitor with
a 35mm negative, and they focus an electron beam on its surface. All
this in complete dark conditions. The beam goes over the film's
emulsion, as it would go on a TV screen, with varying strenght,
according to the luminance. For this reason, they cannot record the
color information at once, but they do sequential color recording (I
think the method is similar to Glorious Technicolor). You can have the
full details of the method at the site below...
http://www.spe.sony.com/Pictures/Hidef/ebr01.htm
That is why Sony's transfers of HDTV footage to 35mm film look so good.
I think, that if they used the "traditional" method (optical printer?),
the results would be comparable to Super 8mm blown up to 35mm, or may be
worse than that. The image may look better from Super 8mm, since HDTV
can produce vibrant colors easier than Super 8mm can (with worse
shooting conditions than Super 8mm). So, even if the resolution were a
bit lower than Super 8mm's, the overall image would be better. Finally,
I believe that if they used the "traditional" method to transfer HDTV to
35mm (not the EBR method!), then in as large projection sizes as in
movie theaters, the "pixelation" effect would be very noticeable and
very annoying. I would prefer Super 8mm's graining than pixelation, even
if the colors were not so vibrant. Colors can be corrected, but to
correct pixelation is very difficult.
Can you confirm or reject my assumptions above?
Thank you again.
Sotirios Papakonstantinou
> will produce a sharper blow-up - then go ahead... Don't listen to anyone,
> learn it for yourself.
>
> David Mullen
Sotirios Papakonstantinou
Sorry, if I tortured your brain. Your post had a rather humorous style,
I liked that.
OK! I don't try to prove that Super 8mm is superior to HDTV. I just try
to confirm or reject that. Telling me that Super 8mm is very grainy,
does not help much. I would like some figures/calculations/whatever to
back this up. OK! I believe you that Super 8mm is worse than HDTV;
afterall you are the professionals, you have used both formats and you
know better than me. The whole purpose of me is to determine how much
amount of detail can Super 8mm show (in good shooting conditions, with
enough light and high contrast). I am a student of Electrical &
Electronic Engineering in City University, London, and I am curious to
see if it is feasible to make a digital video-camera capable of
resolution (amount of detail) equal to Super 8mm's, with existing
technology. If HDTV is indeed better that Super 8mm, then it has be done
already. The next step is to digitally emulate 16mm or 35mm. If Super
8mm can show more detail than HDTV (at 1920x1080 resolution), then there
is some improvement to be done. That is what I want to know. Of course,
if I want to find out whether making that camera is feasible, I will
need some data, such as the actual resolution of Super 8mm. That is why
I insist that Super 8mm has 2000x1500. I want someone to confirm or
reject that (preferable telling me the actual resolution). Do you know
where I could find such information on the Internet? I would like some
comparisons of films (including Super 8mm) with video formats, digital
displays etc. Do you know where I can find such information?
I wish I could make a feature length film on Super 8mm, as you suggest,
but I cannot afford it. Of course, the purpose would be for my pleasure,
not to prove that Super 8mm is the best. Besides, I am only an amateur
(incidentally shooting on Hi8 video, not Super 8mm film...) Regarding
the smashing of "damned HDTV equipment", I would do it (if I would not
have any trouble), because I don't like seeing others using gear I
cannot afford. I feel a bit jealous...
Thank you, and excuse me if I "tortured" you.
Sotirios Papakonstantinou.
Alex Reckie wrote:
>
> >
> > will produce a sharper blow-up - then go ahead... Don't listen to anyone,
> > learn it for yourself.
>
David Mullen
>On computers, like in painting or viewing a photograph, the only optical
system
>involved is the eye. There is no camera/projector pair to worry about. The
>question becomes, with great optics, what is the equivalent compter
resolution
>for working with Super8 with no loss of image information? The answer is
you
>have to work with pixels near the grain size, or idealy less than half the
>grain size to avoid artifacts die to aliasing.
So what you're saying is that one must digitize Super8 at a higher
resolution in order to capture the grain structure? That I can understand,
but that doesn't mean that the Super8 format actually HAS that higher
resolution just because you need to scan it at that level - as you said, the
optics for one thing will reduce the visible resolution.
My point - again - is that from a PRACTICAL and VISIBLE point-of-view,
Super8 does not equal HDTV in terms of resolution. And if it isn't
practical, then what's the point? I would assume the purpose of this
discussion is to determine a practical shooting format, with the end result
either being a 35mm release print or a release on home video or a broadcast
on TV. In all three scenarios, HDTV is more practical than Super8, with
technically better, sharper results ("broadcast quality"). And definitely
finer-grained results.
The main advantage of Super8 over HDTV is that it is a film image, not an
electronic one.
If the end result is a Super8 projection, then shoot Super8. If you want a
nice, grainy and soft image for a blow-up to a 35mm print, or to be seen on
TV or tape, shoot Super8. If you like the way Super8 looks, then shoot
Super8. If you like film over video, shoot Super8. There's LOTS of reasons
you find to shoot in Super8 - resolution is just not one of them.
Really, I'm not against the format - I love it. I've shot Super8 since I
was 14 years old - that's 22 years of Super8 shooting! Now I mostly shoot
35mm and some betacam, but I just used some Super8 for a feature last
summer.
Since this discussion has gone in circles, I'm going to stop posting about
it now.
David Mullen
David Mullen
>Finally,
>I believe that if they used the "traditional" method to transfer HDTV to
>35mm (not the EBR method!), then in as large projection sizes as in
>movie theaters, the "pixelation" effect would be very noticeable and
>very annoying. I would prefer Super 8mm's graining than pixelation, even
>if the colors were not so vibrant.
I said I would stop posting, but I just wanted to point out that the
documentary "Visions of Light" used old analog HDTV interviews with the
cinematographers (shot with lots of net diffusion at times) and transferred
it to 35mm film for a small theatrical release. I saw this movie on fairly
large screens and there was no pixellation or lines or grain/noise in the
HDTV image. It was a little pastel, soft (partly due to diffusion) and with
weak blacks, but not pixellated. It looked more like Super16-to-35mm.
David Mullen
Sotirios Papakonstantinou
OK. I get your point. I do not want to sample or digitize a Super
8mm frame, nothing to do with such processes. My question was specific.
While increasing the display resoltuion of a computer monitor (from
640x480 to 800x600, from 800x600 to 1024x768, etc etc), then what is the
appropriate resolution, if one wants to see the same amount of detail as
a typical Super 8mm frame is capable of (in good shooting conditions,
with plenty of light and high contrast)?
Or see it that way: One uses a scanner to digitize a large image. If
he does 100% zoom (after the initial step of "preview"), then the image
will be very large, say 4000x3000 pixels or over. In order to save hard
disk space, one lowers the resolution of the image, reducing the amount
of detail at the same time. What should be the image's resolution (size
in pixels) if we wanted it to show the same amount of information, as if
it were a Super 8mm frame (shooted in good conditions, as described
above)? I think, it would be 2000x1500. Can you confirm or reject that?
I would prefer stating the actual resolution (image size in pixels), if
you know what it is. Do you know where I can find Super 8mm's
specifications? I tried Kodak's site, but I didn't find this info. Can
you point me to a specific site (without having to click other links, so
I won't get lost again).
I hope my points are clear now. Thank you in advance.
Sotirios Papakonstantinou
Regnirps wrote:
>
> >"David Mullen" <dav...@earthlink.net>
> >Date: Tue, Mar 2, 1999 8:23 AM
> >Message-id: <7bh37t$m8h$1...@fir.prod.itd.earthlink.net>
> >
> >I still don't understand your math or logic.
> >
> >If 4000x3000 is considered necessary to render a 35mm Academy frame, which
> >is .864" x .630" (.544 sq. inches). The 16mm frame is .404" x .295" (.119
> >sq. inches) - meaning that the 16mm frame is only 22% of the size of the
> >35mm frame in total surface area.
> >
> >Assuming that the Super8 frame is slightly larger than 22% of the 16mm frame
> >(let's round up tp 30%), that still makes the Super8 frame only about 6%
> >of
> >the total area of a 35mm frame. Let's round up some more and say that it
> >is
> >10% of the 35mm frame.
> >
> >So how does a format that only takes up about 10% of the 35mm frame going
> >to
> >equal 2000x1500, or half of the 35mm's 4000x3000?
>
> I think there is a natural confusion over resolution when we mix the fields of
> image processing, which deals with information theory and maximum sampling
> required to capture ALL the information in a signal to avoid aliasing (spatial
> equivalent of the time phenomena of wagon wheels going backwards in old films -
> too slow a sampling rate for the speed of the wheel) and the field of
> photography. In photography the resolution is usually defined in terms of the
> optical system's modulation transfer function of MTF and there is a convention
> of resolving line pairs to characterize the MTF.
>
> On computers, like in painting or viewing a photograph, the only optical system
> involved is the eye. There is no camera/projector pair to worry about. The
> question becomes, with great optics, what is the equivalent compter resolution
> for working with Super8 with no loss of image information? The answer is you
> have to work with pixels near the grain size, or idealy less than half the
Regnirps
I have a Sony TRS-900 three chip camera. I notice that the video looks a LOT
better than a single chip camera. I can think of several reasons, but the one
that might account for the better detail is that I assume the three chips are
not EXACTLY lined up pixel edge to pixel edge so that there is a form of
ani-aliasing or jaggies removal caused by a slight overlap of pixels from each
chip.
Is this the case? Also, when you use these cameras in B&W mode do they average
the three chips? This would reduce image noise by a factor of the square root
of 3 or about 1.7 .
BTW, FireWire into my Mac G3 produces beutifull stills for printing and web
stuff. I wish the local telecine machines had FireWire outputs so that while a
transfer was being made I could take a dub straight onto my iDV tape.
Charlie Springer
Sotirios Papakonstantinou
One more time I am sure you didn't understand what I ask and what I
want. OK. I do not want to scan Super 8mm, I do not want to digitize it
or work with Super 8mm. My question is rather specific. I am just not
familiar with photo terms (like resolving power etc) to express it that
way. I am only familiar with computer terms, so I try to express it in
terms of pixels etc. That may be the reason why you do not understand
what I want to know. Think about computers and digital displays. Nothing
to do with photographic aspects.
Here it goes again: As anyone that has a computer knows, the greater
the resolution (ie number of pixels in the horizontal and vertical
dimension composing the image), the greater the detail he can see on a
monitor, or scan from a photograph or print with a printer. (I hope you
get the point now...). One can increase the available resolution
(provided his graphics card has enough memory to cope with) from 640x480
to 800x600 and have finer detail. Similarly an increase from 800x600 to
1024x768 allows for more data (more detail) to be displayed on the same
screen area. Now my question (be sure you understand this time): If one
increases the resolution (amount of detail), where he should stop so
that the amount of detail visible on screen equals the amount of detail
a Super 8mm frame can capture (with the best possible shooting
conditions)? I believe the required resolution to see the same detail as
a Super 8mm frame would provide, is 2000x1500. Can anyone confirm or
reject that? I would like some kind of proof, like stating the actual
resolution of Super 8mm film (if it were a digital display). If you do
not know the answer, then do you know were I could find Super 8mm film's
specifications? I tried Kodak's site but I didn't manage to find the
correct link (if any). Can you point me to the correct address?
Certainly my purpose is not to tell Super 8mm is better from HDTV. It
may be, it may not. I do not want to advertise Super 8mm as a good
format. All I want to do, is to examine whether one can build a video
format (probably a digital camera), capable to capture the same amount
of detail that a typical Super 8mm film can capture. Of course, I will
need some data to assess this, including the actual resolution (data
capacity, amount of detail, you name it) of Super 8mm film.
Can you help now (I hope you understood this time, at last!)
Thank you in advance...
Sotirios Papakonstantinou
David Mullen wrote:
>
> >On computers, like in painting or viewing a photograph, the only optical
> system
> >involved is the eye. There is no camera/projector pair to worry about. The
> >question becomes, with great optics, what is the equivalent compter
> resolution
> >for working with Super8 with no loss of image information? The answer is
> you
> >have to work with pixels near the grain size, or idealy less than half the
> >grain size to avoid artifacts die to aliasing.
>