i got confused

[Stefano Sarfati Nahmad]

i've been browsing through this great thread but i got confused.

what does footage mean ?
what is 30p or 25p ?
what is progressive scan ?


I have a Canon XL1s (PAL system)
so far i've been working in normal mode and i read here that if I use frame mode (plus shallow dof, etc) i get a film look. But what am i losing and what am i getting, exactly?

I shoot mostly interviews, can anyone suggest good books on this subject ?

Yes, i know, i lack the basics, but thanks to any kind soul who will help.

[Rob Lohman]

You do lack the basics indeed, so this will probably stay confusing
for you, but I'll try to do my best to explain it to you.

1) Footage is the stuff you shoot. So if you record a women walking down a street this is your footage

2) 30p or 25p is the framerate in progressive scan. Ie, 30 full frames per second (NTSC) or 25 full frames per second (PAL)

3) Normal TV (normal mode on the XL1S) is interlaced. Interlaced means that instead of one full frame of information you have two fields (half frames) of information which are captured 1/25th (PAL) second apart from eachother. So one half of the pictures is recorded a bit earlier then the other half. This fields are interlaced with eachother so you get 1 line from the first field, then one line from the second field, then one line from the first field again, then one line from second field again etc.

This will cause a video-ish look and will result in interlacing
jaggies when the footage is watched on the computer in full
resolution because the computer uses a progressive scan
monitor which the TV is not.

The Canon XL1S does not have true progressive scan CCD's but
uses a technology to convert interlacing to progressive scan.
The catch is that you loose a bit of color resolution (in other words
the image gets a bit softer) in the process.

I hope this explained it for you.

[Rob Belics]

Just to expand a little:

1) The lower case p means progressive. An 'i' would indicate interlaced

2) Footage was originally a film term for the number of feet of film shot.

[Stefano Sarfati Nahmad]

Thank you very much Rob L., that was quick !
(and thank you Rob B)
Your explenation on what interlaced is was excellent.
I still don't quite understand what progressive is.
And also: what happens when I shoot frame mode? Do I have full frames?
would that be preferable?
i hope that is not too much asking...

[Rob Lohman]

Frame mode is sort of like progressive scan (with lower color
resolution) and does result in full frames indeed.

Whether that is preferred depends on what you want to do and
how you want your footage to look. I highly like the look it gives
so I shoot with that.

But if you are shooting for a client or for TV broadcast then
shooting in normal interlaced mode is probably more common.

[Jean-Philippe Archibald]

in Progressive scan (30p or 25p), both two fields of a frame are recorded at the same time. You get one full frame of information. Interlaced footage (60i or 50i) have more "continuity" in the movement, like in the news on TV. Progressive scan and Frame mode look more like film because the frame rate is closer to the frame rate used in film (24p)

what happens when I shoot frame mode? Do I have full frames? would that be preferable?

Which mode is best for your work is really subjective. Try both normal and frame mode, and choose the one you prefer for the type of work you have to do.

[Ram Nagarajan]

Just to add a small correction to Rob L's note on interlacing:
In PAL, each frame is composed of two fields captured 1/50th of a second apart (rather than 1/25th). So two complementary fields are captured every (1/50 + 1/50 second) = 1/25th second. Ergo, 50 fields or 25 frames per second are captured in PAL video. A similar logic applies to NTSC video, where fields are captured 1/60th of a second apart, iving one frame every 1/30th of a second (or 30 frames per second). (More accurately, though, NTSC video is captured at 29.97 fps).
My two bits...
Best,
Ram

[Stefano Sarfati Nahmad]

Hello Jean-Philippe,
thanx for your answer.
you say in progressive scan both fields are taken at the same moment, what is the point in interlacing two same fields, what is the difference btw progressive and frame mode (both seem to have a full frame of a single moment in time)? And (question to Rob L) why should the XL1s want to convert interlacing to progressive scan, when it has already a frame mode (and frame mode seem to be similar to progressive scan)?

When i watch a film on dvd on a computer monitor it looks great. What am i seeing under a "frame point of view"?

---
the more you learn the more you want to learn ...

[Frank Granovski]

What am i seeing under a "frame point of view"?

You see picture after picture, 30 of them per second or 25 of them per second if you're using a PAL cam. "Normal video" consists of 1/2 frames, one odd, then an even, then odd..., etc.

Progessive scan uses the CCD to capture 1 full frame at a time; frame mode uses 2 rapid shutters at a time to cature an odd and even field, which is then interpolated by the cam. Both capture 1 full frame at a time.

As I understand it. :-))

[Jean-Philippe Archibald]

Stefano,

Frame mode is a technic of interpolation used by some compagnies to simulate progressive scan with CCDs made for interlaced video. During the process of interpolation, a bit of resolution is lost. But in the end, you have 30p footages, like progressive scan.

I own an XL1 (not s) myself and shoot 90% of the time in frame mode, especially for narrative work and interviews in a controlled environnement. I like the look. But I prefer the normal mode when I have to shoot live events or handheld interviews on the street for example.

This is really a question of personnal preferences and the look you want to acheive.

[Rob Lohman]

As I explained earlier, the difference between frame mode and
true progressive is (color) resolution. A true progressive scan
CCD will output the same resolution as an interlaced model.
But those chips are much more expensive.

Canon's frame mode EMULATES progressive scan by using a
special form of "de-interlacing". This results in progressive footage,
but with a slight resolution loss (thus softening the image slightly)

Interlacing was designed for TV's back in the old days. Since
we are still basically using this "old" system we are still being
confronted with it. If I'm not mistaken this system was invented
to increase resolution in those days and this was the only way
to do it back then (due to compatibilty or processing power or
something, forgot what the exact reason was).

[Ram Nagarajan]

Actually, interlacing was the only way to utlilize existing data bandwidth properly in the early days of televsion: The amount of picture data that could be passed through the circuit - and they were tube-and-valve systems, mind you, not solid state printed circuit boards - was extremely limited compared to the bandwidth we're getting today. (Today's consumer DV camcorders are transmitting 25 MBps of data; they were lucky to get a fraction of that in the early days.)
The problem that TV engineers back then ran into was: How do we carry all the picture information we're generating given the data bandwidth bottleneck in the system?
The solution was simple: Break each frame of information into two halves, and transmit them successively. Since the whole process of creating the moving image revolves around persistence of vision in the human cognitive process (and 18 frames a second is enough to take care of that), breaking the image into two halves made no difference to the continuity of the information IF the two halves followed each other FAST enough, and the net frame rate that resulted was faster than the minimum required for the persistence of vision effect.
Breaking the image image into halves was achieved by making the scanning electron beam cover all the odd numbered horizontal lines in the frame, then the even numbered horizontal lines, successively. Together, this ended up covering the entire frame.
The NTSC standard (30 fps/60 halves or fields) and the PAL standard (25 fps/50 halves or fields) emerged due to the practicalities of the prevailing AC current cycles in different parts of the world: At 110 V/60 Hz, it made sense to synchronize the scanning beam cycle to 60 cycles/60 fields. In PAL-land (220 V/50 Hz), it made sense to synchronize the scanning beam cycle to 50 cycles/50 fields.
The difference in horizontal scanning line resolution between PAL (625 lines) and NTSC (525 lines) was solely because NTSC predates PAL as a format. Technology had changed considerably by the time PAL emerged, allowing greater amount of data, and finer control of the scanning beam.
Todaya's data bandwidths on digital equipmentmedia allow the recording/transmission of a FULL frame of COLOUR image at hi-def resolution at the SAME TIME.... a far cry from the early days when interlaced B/W images were fuzzy as hell. So the rules and the concepts are altering drastically...and so are the results we can get.
...phew! Sorry for carrying on like that...hope this helped in some way.
Ram

[Frank Granovski]

N. Ramakrishnan, thanks for the history lesson and information. I have a question: what were the specs for the old BBC broadcast system?

[Stefano Sarfati Nahmad]

Thank you Rob, Jean-Philippe, Ram, Frank
So now I gather that (in a PAL system):
TVs and dvcams work @ 25 frames per second (30 frames in NTSC)
When shooting, according to the CCD technology of your camcorder, you can produce different types of frames:
Interlaced, progressive, progressive emulated (frame mode in the XL1)
- Interlaced: every 1/50 of sec half frame (a field) is recorded, which is then interlaced with the following one (1 line from the first field, then one line from the second field, etc.) to produce one full frame every 1/25 of second (i.e. 25fps). You call this “25i”
- Progressive: every 1/25 of sec one full frame is recorded. Don’t know if it does some interlacing or just plain frames, one after the other. You call this “25p”.
- Progressive emulated (frame mode in the Canon XL1): same as interlaced but going through an interpolated (i.e. via software) process of de-interlacing after being recorded. It crates full frames but loses color resolution.
Films are 24 full frames per second (but what a color quality…)
Is that right?

And on the output side?
Ram, when i watch a film dvd on a computer monitor looks great.
but film is only 24p ... is all the difference in there or is it something else?

What is the bottleneck when i watch what i recorded with my XL1 (interlaced or progressive emulated)? Or in other words: what is the best way to watch a miniDV video?

[Rob Lohman]

That is correct Stefano. The best way to watch miniDV is usually
on the TV, especially if it was shot interlaced. If it is progressive
then the computer might be a valuable option as well.

Just to note: most software DVD players can perform realtime
de-interlacing (with varying results) on interlaced DVD/mpeg2
footage.